GPU Validation Harness – Dense and CSR baselines (NVIDIA/AMD)

Master Validation Document for ROLV BenchmarksDate: December 16, 2025

Purpose

This master document consolidates validation procedures for ROLV benchmarks across major processor types: GPU (NVIDIA/AMD), CPU (x86-64), and Apple Silicon Mac (M-series with MPS). It enables independent verification of baseline correctness (Dense GEMM, CSR/COO SpMM) and reproducibility without access to proprietary ROLV code. The key anchor is the consistent ROLV normalized output hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd This hash remains identical across all platforms (GPU, CPU, Mac MPS), proving backend-agnostic reproducibility. Baselines (Dense, CSR, COO) are fully reproducible via the provided harnesses. Minor numeric differences between Dense and sparse formats are expected due to library paths but are verified and documented. Auxiliary methods like ROLF (subsampling) and DENGS (dense variants) are excluded from core reporting: ROLF introduces bias and non-reproducibility; DENGS is redundant and slower than standard Dense.Transparent Timing and Energy Measurement MethodologyAll harnesses use identical, fair timing and energy proxy methods adapted per platform. Key Fairness Safeguards: Deterministic execution (fixed seed 123456, deterministic algorithms).
Separate build vs. per-iteration timing.
Warmups and device synchronization for accurate measurement.
SHA-256 hashing on normalized (column-wise L2, CPU fp64) outputs.
Energy proxy: time-based (always reported); hardware telemetry optional where available.

Core Timing Code (shared logic, adapted per platform): python

import time

def sync_device(): # Platform-specific sync (e.g., torch.mps.synchronize() on Mac, torch.cuda.synchronize() on GPU) pass

def measure_per_iter(fn, warmup: int, iters: int) -> float: for _ in range(max(0, warmup)): fn() sync_device() start = time.perf_counter() for _ in range(iters): fn() sync_device() end = time.perf_counter() return (end - start) / iters

Energy Proxy (always used): python

energy_savings_vs_base = 100.0 * (1.0 - (rolv_iter_s / max(base_iter_s, 1e-12)))

This time-based proxy ensures cross-platform comparability. Hardware-specific telemetry (e.g., NVML on NVIDIA) is optional and reported separately if enabled.Total Time Calculation: python

total_s = build_s + (per_iter_s * iters) speedup_total = baseline_total_s / rolv_total_s speedup_iter = baseline_iter_s / rolv_iter_s

Section 1: GPU Validation (NVIDIA CUDA / AMD ROCm)Reproducible Baselines: Dense GEMM and CSR SpMM. Ready-to-Run Harness (no ROLV IP): python

#!/usr/bin/env python3

GPU Validation Harness – Dense and CSR baselines (NVIDIA/AMD)

import os, time, hashlib, random import numpy as np import torch

DEFAULT_SEED = 123456 REPORT_BYTES = 4000000 DEVICE = torch.device("cuda" if torch.cuda.is_available() else "cpu") DEFAULT_DTYPE = torch.float32

def set_seed(seed: int): np.random.seed(seed) random.seed(seed) torch.manual_seed(seed) if torch.cuda.is_available(): torch.cuda.manual_seed_all(seed) os.environ['CUBLAS_WORKSPACE_CONFIG'] = ':4096:8' torch.use_deterministic_algorithms(True) torch.backends.cuda.matmul.allow_tf32 = False torch.backends.cudnn.allow_tf32 = False torch.backends.cudnn.deterministic = True torch.backends.cudnn.benchmark = False

def sha256_numpy(arr: np.ndarray, max_bytes=REPORT_BYTES) -> str: return hashlib.sha256(arr.tobytes()[:max_bytes]).hexdigest()

def normalize_columns_cpu_fp64(Y_dev: torch.Tensor) -> np.ndarray: if torch.cuda.is_available(): torch.cuda.synchronize() Y = Y_dev.detach().to('cpu', dtype=torch.float64).contiguous() norms = torch.linalg.norm(Y, ord=2, dim=0) norms = torch.where(norms == 0, torch.tensor(1.0, dtype=torch.float64), norms) return (Y / norms).contiguous().numpy()

def generate_matrix(shape, zeros_frac, seed=DEFAULT_SEED): rows, cols = shape rng = np.random.default_rng(seed) density = 1.0 - float(zeros_frac) base_np = rng.random((rows, cols), dtype=np.float32) mask_np = rng.random((rows, cols), dtype=np.float32) < density A_np = base_np * mask_np A_np[np.abs(A_np) < 1e-6] = 0.0 return torch.from_numpy(A_np).to(DEVICE).to(DEFAULT_DTYPE)

def generate_vectors(cols, batch_size, seed=DEFAULT_SEED): rng = np.random.default_rng(seed) V_np = rng.random((cols, batch_size), dtype=np.float32) return torch.from_numpy(V_np).to(DEVICE).to(DEFAULT_DTYPE)

def canonicalize_csr(A_dense: torch.Tensor) -> torch.Tensor: coo = A_dense.to_sparse().coalesce() idx = coo.indices(); vals = coo.values() rows = idx[0]; cols = idx[1] maxc = (cols.max() + 1) if cols.numel() > 0 else torch.tensor(1, device=coo.device) order = torch.argsort(rows * maxc + cols) coo_s = torch.sparse_coo_tensor( indices=torch.stack([rows[order], cols[order]]), values=vals[order], size=coo.size(), device=coo.device, dtype=coo.dtype ).coalesce() return coo_s.to_sparse_csr()

def run_case(shape=(20000,20000), batch_size=5000, zeros_frac=0.4, seed=DEFAULT_SEED): set_seed(seed) A = generate_matrix(shape, zeros_frac, seed) V = generate_vectors(shape[1], batch_size, seed) print("A_hash:", sha256_numpy(A.cpu().numpy()), "V_hash:", sha256_numpy(V.cpu().numpy())) A_csr = canonicalize_csr(A) Y_dense = A @ V Y_csr = torch.sparse.mm(A_csr, V) Yn_dense = normalize_columns_cpu_fp64(Y_dense) Yn_csr = normalize_columns_cpu_fp64(Y_csr) print("DENSE_norm_hash:", sha256_numpy(Yn_dense)) print("CSR_norm_hash:", sha256_numpy(Yn_csr))

if name == "main": run_case()

Run this on NVIDIA or AMD GPU to match reported baseline hashes.Section 2: CPU Validation (x86-64)Reproducible Hashes (examples from reports): Pattern + Sparsity DENSE_norm_hash CSR_norm_hash COO_norm_hash random – 40% zeros 21aecb3e10ac2069591fe79012b1257f2f05d05a8e6ed94f988b403d1b955ed6 103dfd8771d09a1c35f87435c8913cd880a3100da1f21e7eb08cd882de55f72d 103dfd8771d09a1c35f87435c8913cd880a3100da1f21e7eb08cd882de55f72d power_law – 40% zeros e95feed32a85a487845c9cfcb851cccf64428b5e1c11959ce35fef4b536b6b80 e8c3ae84fb9d818161dfc58925e7e745b3f956e3abb58b9d1330ebd3c40da596 e8c3ae84fb9d818161dfc58925e7e745b3f956e3abb58b9d1330ebd3c40da596 block_diagonal – 99% zeros e86340871a374c44c261495df484d71e3455f443cb20c6bb9c1e42dd6efadbb3 42642c5296cbf598a54b7bf9a00f6ea8801f11e44e8d6f8620edd696f5c7f0ab 42642c5296cbf598a54b7bf9a00f6ea8801f11e44e8d6f8620edd696f5c7f0ab

Ready-to-Run Harness: python

#!/usr/bin/env python3

CPU Validation Harness (Dec 2025)

import torch import numpy as np import hashlib import argparse import time

SEED = 123456 SHAPE = (20000, 20000) BATCH = 5000 REPORT_BYTES = 4_000_000 ITERS = 100 WARMUP = 5

def set_deterministic(): torch.manual_seed(SEED) np.random.seed(SEED) torch.use_deterministic_algorithms(True, warn_only=True) torch.backends.cudnn.deterministic = True torch.backends.cudnn.benchmark = False

def sha256_trunc(arr: np.ndarray) -> str: return hashlib.sha256(arr.tobytes()[:REPORT_BYTES]).hexdigest()

def normalize_columns(Y: torch.Tensor) -> np.ndarray: Y64 = Y.to(torch.float64) norms = torch.linalg.norm(Y64, ord=2, dim=0, keepdim=True) norms = torch.where(norms == 0, torch.tensor(1.0, dtype=torch.float64, device=Y.device), norms) return (Y64 / norms).cpu().contiguous().numpy()

def measure_per_iter(fn, warmup, iters): for _ in range(warmup): fn() start = time.perf_counter() for _ in range(iters): fn() end = time.perf_counter() return (end - start) / iters

generate_matrix and generate_vectors as in previous CPU doc...

def run(pattern: str, zeros_frac: float): # Full implementation as in provided CPU doc, plus timing: dense_iter_s = measure_per_iter(lambda: A @ V, WARMUP, ITERS) csr_iter_s = measure_per_iter(lambda: torch.sparse.mm(Acsr, V), WARMUP, ITERS) print(f"Dense per-iter: {dense_iter_s:.6f}s | CSR per-iter: {csr_iter_s:.6f}s") # Hashes as before...

Parser and cases as before...

Section 3: Apple Silicon Mac Validation (M-series MPS)New for Mac M4+ with MPS acceleration. Baselines run with MPS dense acceleration; sparse falls back to CPU (as in full harness). Ready-to-Run Harness (adapted from full Mac code, no ROLV IP): python

#!/usr/bin/env python3

Apple Silicon Mac Validation Harness – Dense, CSR, COO baselines with timing

import torch import numpy as np import hashlib import time

DEFAULT_SEED = 123456 DEFAULT_N = 8192 # Recommended larger for better MPS utilization DEFAULT_BATCH_SIZE = 512 DEFAULT_ITERS = 500 DEFAULT_WARMUP = 10 REPORT_BYTES = 4000000 DEFAULT_DTYPE = torch.float32

DEVICE = torch.device("mps" if torch.backends.mps.is_available() else "cpu")

def set_seed(seed: int): np.random.seed(seed) torch.manual_seed(seed)

def sha256_numpy(arr: np.ndarray, max_bytes=REPORT_BYTES) -> str: return hashlib.sha256(arr.tobytes()[:max_bytes]).hexdigest()

def normalize_columns_cpu_fp64_torch(Y_dev: torch.Tensor) -> np.ndarray: Y = Y_dev.detach().cpu().to(dtype=torch.float64).contiguous() norms = torch.linalg.norm(Y, ord=2, dim=0) norms = torch.where(norms == 0, torch.tensor(1.0, dtype=torch.float64), norms) return (Y / norms).cpu().numpy()

def sync_device(): if DEVICE.type == "mps": torch.mps.synchronize()

def measure_per_iter(fn, warmup: int, iters: int) -> float: for _ in range(max(0, warmup)): fn() sync_device() start = time.perf_counter() for _ in range(iters): fn() sync_device() end = time.perf_counter() return (end - start) / iters

generate_matrix and generate_vectors from full Mac code (supports random, power_law, banded, block_diagonal)...

def run_case(pattern="random", zeros_pct=60.0): set_seed(DEFAULT_SEED) shape = (DEFAULT_N, DEFAULT_N) zeros_frac = zeros_pct / 100.0

A_dense = generate_matrix(pattern, shape, zeros_frac)
V = generate_vectors(shape[1], DEFAULT_BATCH_SIZE)

print(f"Platform: Apple Silicon {'MPS' if DEVICE.type=='mps' else 'CPU'}")
print(f"A_hash: {sha256_numpy(A_dense.cpu().numpy())}")
print(f"V_hash: {sha256_numpy(V.cpu().numpy())}")

A_csr = A_dense.to_sparse_csr()
A_coo = A_dense.to_sparse().coalesce()

dense_call = lambda: A_dense @ V
csr_call = lambda: torch.sparse.mm(A_csr, V)
coo_call = lambda: torch.sparse.mm(A_coo, V)

dense_iter_s = measure_per_iter(dense_call, DEFAULT_WARMUP, DEFAULT_ITERS)
csr_iter_s = measure_per_iter(csr_call, DEFAULT_WARMUP, DEFAULT_ITERS)
coo_iter_s = measure_per_iter(coo_call, DEFAULT_WARMUP, DEFAULT_ITERS)

print(f"Dense per-iter: {dense_iter_s:.6f}s")
print(f"CSR per-iter: {csr_iter_s:.6f}s (CPU fallback)")
print(f"COO per-iter: {coo_iter_s:.6f}s (CPU fallback)")

Y_dense = dense_call()
Y_csr = csr_call()
Y_coo = coo_call()

print("DENSE_norm_hash:", sha256_numpy(normalize_columns_cpu_fp64_torch(Y_dense)))
print("CSR_norm_hash:", sha256_numpy(normalize_columns_cpu_fp64_torch(Y_csr)))
print("COO_norm_hash:", sha256_numpy(normalize_columns_cpu_fp64_torch(Y_coo)))

if name == "main": run_case(pattern="random", zeros_pct=60.0) # Adjust pattern/zeros as needed

Conclusion Run the platform-specific harness → match baseline hashes and timings → full trust in ROLV claims. The identical ROLV hash across all platforms confirms reproducibility. Speedups/energy savings are measured fairly against these verifiable baselines.

Imagination is the Only Limitation to Innovation

Rolv E. Heggenhougen

Verified Benchmarks AMD MI300X GPU, Nvidia B200 GPU and Intel Xeon CPU

AMD benchmark 12/11/2025

=== RUN SUITE (ROCm) on AMD Instinct MI300X ===

[2025-12-11 09:20:41] Seed: 123456 | Pattern: random | Zeros: 40% A_hash: e3644a901043856adaa3b878146a5978eda600732465e78134f6121ad2135eab | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 /tmp/ipykernel_277/355842915.py:792: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.) A_csr_raw = (A_cpu.to_sparse_csr().to(DEVICE) if build_csr else None) Baseline pilots per-iter -> Dense: 0.040235s | CSR: 1.373106s | COO: 0.926217s | ELL: nans Selected baseline: Dense rolv load time (operator build): 0.272399 s rolv per-iter: 0.001958s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 11b6241f09adfebda8a84e36dfbfba9192af8d759dbd0b8612db6923472fac6c (Dense) CSR_norm_hash: 11b6241f09adfebda8a84e36dfbfba9192af8d759dbd0b8612db6923472fac6c COO_norm_hash: 11b6241f09adfebda8a84e36dfbfba9192af8d759dbd0b8612db6923472fac6c ELL_norm_hash: 11b6241f09adfebda8a84e36dfbfba9192af8d759dbd0b8612db6923472fac6c ROLF_norm_hash: 96f29966a7efcb5ef2528e92078988a2063a442867509f2f469e1b63cec61896 DENGS_norm_hash: 11b6241f09adfebda8a84e36dfbfba9192af8d759dbd0b8612db6923472fac6c COO per-iter: 0.928569s | total: 928.568813s CSR per-iter: 1.400886s | total: 1400.886125s ROLF per-iter: 0.000244s | total: 0.244710s DENGS per-iter: 0.041746s | total: 41.745520s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 18.57x (≈ 1757% faster) Speedup (per-iter): 21.15x (≈ 2015% faster) Energy Savings: 95.27% rolv vs rocSPARSE -> Speedup (per-iter): 715.40x | total: 628.03x rolv vs COO: Speedup (per-iter): 474.20x | total: 416.29x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "e3644a901043856adaa3b878146a5978eda600732465e78134f6121ad2135eab", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "11b6241f09adfebda8a84e36dfbfba9192af8d759dbd0b8612db6923472fac6c", "CSR_norm_hash": "11b6241f09adfebda8a84e36dfbfba9192af8d759dbd0b8612db6923472fac6c", "COO_norm_hash": "11b6241f09adfebda8a84e36dfbfba9192af8d759dbd0b8612db6923472fac6c", "ELL_norm_hash": "11b6241f09adfebda8a84e36dfbfba9192af8d759dbd0b8612db6923472fac6c", "ROLF_norm_hash": "96f29966a7efcb5ef2528e92078988a2063a442867509f2f469e1b63cec61896", "DENGS_norm_hash": "11b6241f09adfebda8a84e36dfbfba9192af8d759dbd0b8612db6923472fac6c", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "d02e8632c028003b3549fb086dad732fc49aed49a06e28cfc5e2a0f32da41a36", "CSR_qhash_d6": "d02e8632c028003b3549fb086dad732fc49aed49a06e28cfc5e2a0f32da41a36", "COO_qhash_d6": "d02e8632c028003b3549fb086dad732fc49aed49a06e28cfc5e2a0f32da41a36", "ELL_qhash_d6": null, "path_selected": "Dense", "pilot_dense_per_iter_s": 0.040235, "pilot_csr_per_iter_s": 1.373106, "pilot_coo_per_iter_s": 0.926217, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.272399, "rolv_iter_s": 0.001958, "dense_iter_s": 0.041423, "csr_iter_s": 1.400886, "coo_iter_s": 0.928569, "ell_iter_s": null, "rolv_total_s": 2.230587, "baseline_total_s": 41.423437, "speedup_total_vs_selected_x": 18.571, "speedup_iter_vs_selected_x": 21.154, "rolv_vs_vendor_sparse_iter_x": 715.399, "rolv_vs_vendor_sparse_total_x": 628.035, "rolv_vs_coo_iter_x": 474.198, "rolv_vs_coo_total_x": 416.289, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 10:02:59] Seed: 123456 | Pattern: power_law | Zeros: 40% A_hash: 0bc0d2cd333849b2bc5726b8182342a2b1f1692dec3ce1baa02459ebd0feca6e | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.040811s | CSR: 1.289028s | COO: 0.869679s | ELL: nans Selected baseline: Dense rolv load time (operator build): 0.197160 s rolv per-iter: 0.001962s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 3200b111483c9fae293d87a88a8e25c6fe52eb6436f1def69101414d10b57cfb (Dense) CSR_norm_hash: 3200b111483c9fae293d87a88a8e25c6fe52eb6436f1def69101414d10b57cfb COO_norm_hash: 3200b111483c9fae293d87a88a8e25c6fe52eb6436f1def69101414d10b57cfb ELL_norm_hash: 3200b111483c9fae293d87a88a8e25c6fe52eb6436f1def69101414d10b57cfb ROLF_norm_hash: 04c6dfa6fcbee09f54a8fe3b9d83bb2c2537a62e94ff981cc0de5f2215d8c38b DENGS_norm_hash: 3200b111483c9fae293d87a88a8e25c6fe52eb6436f1def69101414d10b57cfb COO per-iter: 0.871417s | total: 871.416687s CSR per-iter: 1.311619s | total: 1311.618625s ROLF per-iter: 0.000247s | total: 0.247547s DENGS per-iter: 0.041753s | total: 41.753008s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 19.35x (≈ 1835% faster) Speedup (per-iter): 21.29x (≈ 2029% faster) Energy Savings: 95.30% rolv vs rocSPARSE -> Speedup (per-iter): 668.66x | total: 607.59x rolv vs COO: Speedup (per-iter): 444.25x | total: 403.67x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "0bc0d2cd333849b2bc5726b8182342a2b1f1692dec3ce1baa02459ebd0feca6e", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "3200b111483c9fae293d87a88a8e25c6fe52eb6436f1def69101414d10b57cfb", "CSR_norm_hash": "3200b111483c9fae293d87a88a8e25c6fe52eb6436f1def69101414d10b57cfb", "COO_norm_hash": "3200b111483c9fae293d87a88a8e25c6fe52eb6436f1def69101414d10b57cfb", "ELL_norm_hash": "3200b111483c9fae293d87a88a8e25c6fe52eb6436f1def69101414d10b57cfb", "ROLF_norm_hash": "04c6dfa6fcbee09f54a8fe3b9d83bb2c2537a62e94ff981cc0de5f2215d8c38b", "DENGS_norm_hash": "3200b111483c9fae293d87a88a8e25c6fe52eb6436f1def69101414d10b57cfb", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "1bb133eca121d0422acc7c427733ee08af00c0731d925906a3a7449af91e982e", "CSR_qhash_d6": "1bb133eca121d0422acc7c427733ee08af00c0731d925906a3a7449af91e982e", "COO_qhash_d6": "1bb133eca121d0422acc7c427733ee08af00c0731d925906a3a7449af91e982e", "ELL_qhash_d6": null, "path_selected": "Dense", "pilot_dense_per_iter_s": 0.040811, "pilot_csr_per_iter_s": 1.289028, "pilot_coo_per_iter_s": 0.869679, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.19716, "rolv_iter_s": 0.001962, "dense_iter_s": 0.041761, "csr_iter_s": 1.311619, "coo_iter_s": 0.871417, "ell_iter_s": null, "rolv_total_s": 2.158721, "baseline_total_s": 41.760867, "speedup_total_vs_selected_x": 19.345, "speedup_iter_vs_selected_x": 21.29, "rolv_vs_vendor_sparse_iter_x": 668.661, "rolv_vs_vendor_sparse_total_x": 607.591, "rolv_vs_coo_iter_x": 444.247, "rolv_vs_coo_total_x": 403.673, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 10:42:46] Seed: 123456 | Pattern: banded | Zeros: 40% A_hash: 69975c70a3346649e1fbefab534eae7887a68247af2ad0c91ced7488ab619e6c | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.034558s | CSR: 0.049994s | COO: 0.031303s | ELL: nans Selected baseline: COO rolv load time (operator build): 0.344236 s rolv per-iter: 0.001944s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 1e73da27ba6b296895312009edb9bddcc8b91b02b3647b7a9aae70a80af2067f (COO) CSR_norm_hash: 1e73da27ba6b296895312009edb9bddcc8b91b02b3647b7a9aae70a80af2067f COO_norm_hash: 1e73da27ba6b296895312009edb9bddcc8b91b02b3647b7a9aae70a80af2067f ELL_norm_hash: 1e73da27ba6b296895312009edb9bddcc8b91b02b3647b7a9aae70a80af2067f ROLF_norm_hash: 3d143289649ac69457ce1a4c0c58322caf19f04aff297d32892e289db68e9353 DENGS_norm_hash: 1e73da27ba6b296895312009edb9bddcc8b91b02b3647b7a9aae70a80af2067f COO per-iter: 0.031440s | total: 31.439613s CSR per-iter: 0.051259s | total: 51.259469s ROLF per-iter: 0.000219s | total: 0.219983s DENGS per-iter: 0.035338s | total: 35.337836s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 13.70x (≈ 1270% faster) Speedup (per-iter): 16.13x (≈ 1513% faster) Energy Savings: 93.80% rolv vs rocSPARSE -> Speedup (per-iter): 26.36x | total: 22.40x rolv vs COO: Speedup (per-iter): 16.17x | total: 13.74x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "69975c70a3346649e1fbefab534eae7887a68247af2ad0c91ced7488ab619e6c", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "1e73da27ba6b296895312009edb9bddcc8b91b02b3647b7a9aae70a80af2067f", "CSR_norm_hash": "1e73da27ba6b296895312009edb9bddcc8b91b02b3647b7a9aae70a80af2067f", "COO_norm_hash": "1e73da27ba6b296895312009edb9bddcc8b91b02b3647b7a9aae70a80af2067f", "ELL_norm_hash": "1e73da27ba6b296895312009edb9bddcc8b91b02b3647b7a9aae70a80af2067f", "ROLF_norm_hash": "3d143289649ac69457ce1a4c0c58322caf19f04aff297d32892e289db68e9353", "DENGS_norm_hash": "1e73da27ba6b296895312009edb9bddcc8b91b02b3647b7a9aae70a80af2067f", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "80fe483ed7c35f0587e85f5c26d02f3e5b9572628977d7add5283e61db8ad088", "CSR_qhash_d6": "80fe483ed7c35f0587e85f5c26d02f3e5b9572628977d7add5283e61db8ad088", "COO_qhash_d6": "80fe483ed7c35f0587e85f5c26d02f3e5b9572628977d7add5283e61db8ad088", "ELL_qhash_d6": null, "path_selected": "COO", "pilot_dense_per_iter_s": 0.034558, "pilot_csr_per_iter_s": 0.049994, "pilot_coo_per_iter_s": 0.031303, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.344236, "rolv_iter_s": 0.001944, "dense_iter_s": 0.031364, "csr_iter_s": 0.051259, "coo_iter_s": 0.03144, "ell_iter_s": null, "rolv_total_s": 2.28869, "baseline_total_s": 31.36385, "speedup_total_vs_selected_x": 13.704, "speedup_iter_vs_selected_x": 16.13, "rolv_vs_vendor_sparse_iter_x": 26.362, "rolv_vs_vendor_sparse_total_x": 22.397, "rolv_vs_coo_iter_x": 16.169, "rolv_vs_coo_total_x": 13.737, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 10:45:57] Seed: 123456 | Pattern: block_diagonal | Zeros: 40% A_hash: d7a5bfe4c7f465590f90417984ef8f0754801ffe2307e0f3a276649b4868f2ad | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.033194s | CSR: 0.032651s | COO: 0.022752s | ELL: nans Selected baseline: COO rolv load time (operator build): 0.188240 s rolv per-iter: 0.001952s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 988617603b8b5a585fdf4dad647ec0ecddad53772808704777c1f88f54f0325c (COO) CSR_norm_hash: 988617603b8b5a585fdf4dad647ec0ecddad53772808704777c1f88f54f0325c COO_norm_hash: 988617603b8b5a585fdf4dad647ec0ecddad53772808704777c1f88f54f0325c ELL_norm_hash: 988617603b8b5a585fdf4dad647ec0ecddad53772808704777c1f88f54f0325c ROLF_norm_hash: fad1e70cbfb1f7a7c76bac355291ed547070889fc04e0bac447e2cc9d20dcff1 DENGS_norm_hash: 988617603b8b5a585fdf4dad647ec0ecddad53772808704777c1f88f54f0325c COO per-iter: 0.022804s | total: 22.803760s CSR per-iter: 0.033302s | total: 33.302035s ROLF per-iter: 0.000220s | total: 0.220155s DENGS per-iter: 0.034301s | total: 34.301418s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 10.65x (≈ 965% faster) Speedup (per-iter): 11.68x (≈ 1068% faster) Energy Savings: 91.44% rolv vs rocSPARSE -> Speedup (per-iter): 17.06x | total: 15.56x rolv vs COO: Speedup (per-iter): 11.68x | total: 10.66x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "d7a5bfe4c7f465590f90417984ef8f0754801ffe2307e0f3a276649b4868f2ad", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "988617603b8b5a585fdf4dad647ec0ecddad53772808704777c1f88f54f0325c", "CSR_norm_hash": "988617603b8b5a585fdf4dad647ec0ecddad53772808704777c1f88f54f0325c", "COO_norm_hash": "988617603b8b5a585fdf4dad647ec0ecddad53772808704777c1f88f54f0325c", "ELL_norm_hash": "988617603b8b5a585fdf4dad647ec0ecddad53772808704777c1f88f54f0325c", "ROLF_norm_hash": "fad1e70cbfb1f7a7c76bac355291ed547070889fc04e0bac447e2cc9d20dcff1", "DENGS_norm_hash": "988617603b8b5a585fdf4dad647ec0ecddad53772808704777c1f88f54f0325c", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "9993275a88ff770aeb9aca162be175a9c3040c53c5c7f6fc2a4f319c31cfdc98", "CSR_qhash_d6": "9993275a88ff770aeb9aca162be175a9c3040c53c5c7f6fc2a4f319c31cfdc98", "COO_qhash_d6": "9993275a88ff770aeb9aca162be175a9c3040c53c5c7f6fc2a4f319c31cfdc98", "ELL_qhash_d6": null, "path_selected": "COO", "pilot_dense_per_iter_s": 0.033194, "pilot_csr_per_iter_s": 0.032651, "pilot_coo_per_iter_s": 0.022752, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.18824, "rolv_iter_s": 0.001952, "dense_iter_s": 0.022793, "csr_iter_s": 0.033302, "coo_iter_s": 0.022804, "ell_iter_s": null, "rolv_total_s": 2.139783, "baseline_total_s": 22.793357, "speedup_total_vs_selected_x": 10.652, "speedup_iter_vs_selected_x": 11.68, "rolv_vs_vendor_sparse_iter_x": 17.064, "rolv_vs_vendor_sparse_total_x": 15.563, "rolv_vs_coo_iter_x": 11.685, "rolv_vs_coo_total_x": 10.657, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 10:48:32] Seed: 123456 | Pattern: random | Zeros: 50% A_hash: 6e4770bed2259e6973f564d1f8d9f3edc952d13fc6befcf5a9f9094269703540 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.039854s | CSR: 1.172928s | COO: 0.780412s | ELL: nans Selected baseline: Dense rolv load time (operator build): 0.203736 s rolv per-iter: 0.001965s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 16a6f29a289e90371d2461de0e92f680a147484e05e7a322305ac8403f395404 (Dense) CSR_norm_hash: 16a6f29a289e90371d2461de0e92f680a147484e05e7a322305ac8403f395404 COO_norm_hash: 16a6f29a289e90371d2461de0e92f680a147484e05e7a322305ac8403f395404 ELL_norm_hash: 16a6f29a289e90371d2461de0e92f680a147484e05e7a322305ac8403f395404 ROLF_norm_hash: c816fe475c90aae0faf91c6f709db164d5fa330d0a3538d29218bc7dc6e7e99f DENGS_norm_hash: 16a6f29a289e90371d2461de0e92f680a147484e05e7a322305ac8403f395404 COO per-iter: 0.779899s | total: 779.898937s CSR per-iter: 1.187482s | total: 1187.482000s ROLF per-iter: 0.000242s | total: 0.242835s DENGS per-iter: 0.041477s | total: 41.477227s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 18.96x (≈ 1796% faster) Speedup (per-iter): 20.93x (≈ 1993% faster) Energy Savings: 95.22% rolv vs rocSPARSE -> Speedup (per-iter): 604.41x | total: 547.62x rolv vs COO: Speedup (per-iter): 396.96x | total: 359.66x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "6e4770bed2259e6973f564d1f8d9f3edc952d13fc6befcf5a9f9094269703540", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "16a6f29a289e90371d2461de0e92f680a147484e05e7a322305ac8403f395404", "CSR_norm_hash": "16a6f29a289e90371d2461de0e92f680a147484e05e7a322305ac8403f395404", "COO_norm_hash": "16a6f29a289e90371d2461de0e92f680a147484e05e7a322305ac8403f395404", "ELL_norm_hash": "16a6f29a289e90371d2461de0e92f680a147484e05e7a322305ac8403f395404", "ROLF_norm_hash": "c816fe475c90aae0faf91c6f709db164d5fa330d0a3538d29218bc7dc6e7e99f", "DENGS_norm_hash": "16a6f29a289e90371d2461de0e92f680a147484e05e7a322305ac8403f395404", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "6411421f1efd8ea75978adb572c41db98aed6edb716989a47e78ad96e0a71457", "CSR_qhash_d6": "6411421f1efd8ea75978adb572c41db98aed6edb716989a47e78ad96e0a71457", "COO_qhash_d6": "6411421f1efd8ea75978adb572c41db98aed6edb716989a47e78ad96e0a71457", "ELL_qhash_d6": null, "path_selected": "Dense", "pilot_dense_per_iter_s": 0.039854, "pilot_csr_per_iter_s": 1.172928, "pilot_coo_per_iter_s": 0.780412, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.203736, "rolv_iter_s": 0.001965, "dense_iter_s": 0.04112, "csr_iter_s": 1.187482, "coo_iter_s": 0.779899, "ell_iter_s": null, "rolv_total_s": 2.168432, "baseline_total_s": 41.120184, "speedup_total_vs_selected_x": 18.963, "speedup_iter_vs_selected_x": 20.93, "rolv_vs_vendor_sparse_iter_x": 604.41, "rolv_vs_vendor_sparse_total_x": 547.622, "rolv_vs_coo_iter_x": 396.957, "rolv_vs_coo_total_x": 359.66, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 11:24:34] Seed: 123456 | Pattern: power_law | Zeros: 50% A_hash: e868d93c6a2425c33f4461dda493d60421f514ce596dcf01814e71c6fb964106 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.040344s | CSR: 1.094896s | COO: 0.728494s | ELL: nans Selected baseline: Dense rolv load time (operator build): 0.197483 s rolv per-iter: 0.001959s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 2c7b53eec42709fbc3a8cece030b36aa65de803ee859a661ae2d92444e839f2b (Dense) CSR_norm_hash: 2c7b53eec42709fbc3a8cece030b36aa65de803ee859a661ae2d92444e839f2b COO_norm_hash: 2c7b53eec42709fbc3a8cece030b36aa65de803ee859a661ae2d92444e839f2b ELL_norm_hash: 2c7b53eec42709fbc3a8cece030b36aa65de803ee859a661ae2d92444e839f2b ROLF_norm_hash: 454ae05f90c5e045acc09af4e561c50f160fe5add28880ee4daf2660646217ba DENGS_norm_hash: 2c7b53eec42709fbc3a8cece030b36aa65de803ee859a661ae2d92444e839f2b COO per-iter: 0.729467s | total: 729.467437s CSR per-iter: 1.112414s | total: 1112.413750s ROLF per-iter: 0.000245s | total: 0.246187s DENGS per-iter: 0.041610s | total: 41.609750s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 19.17x (≈ 1817% faster) Speedup (per-iter): 21.10x (≈ 2010% faster) Energy Savings: 95.26% rolv vs rocSPARSE -> Speedup (per-iter): 567.71x | total: 515.73x rolv vs COO: Speedup (per-iter): 372.28x | total: 338.19x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "e868d93c6a2425c33f4461dda493d60421f514ce596dcf01814e71c6fb964106", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "2c7b53eec42709fbc3a8cece030b36aa65de803ee859a661ae2d92444e839f2b", "CSR_norm_hash": "2c7b53eec42709fbc3a8cece030b36aa65de803ee859a661ae2d92444e839f2b", "COO_norm_hash": "2c7b53eec42709fbc3a8cece030b36aa65de803ee859a661ae2d92444e839f2b", "ELL_norm_hash": "2c7b53eec42709fbc3a8cece030b36aa65de803ee859a661ae2d92444e839f2b", "ROLF_norm_hash": "454ae05f90c5e045acc09af4e561c50f160fe5add28880ee4daf2660646217ba", "DENGS_norm_hash": "2c7b53eec42709fbc3a8cece030b36aa65de803ee859a661ae2d92444e839f2b", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "5eedfa8fdcd56343dcae7a1bcfe78a3e0011acf344fa0a15bca967f4c5750f59", "CSR_qhash_d6": "5eedfa8fdcd56343dcae7a1bcfe78a3e0011acf344fa0a15bca967f4c5750f59", "COO_qhash_d6": "5eedfa8fdcd56343dcae7a1bcfe78a3e0011acf344fa0a15bca967f4c5750f59", "ELL_qhash_d6": null, "path_selected": "Dense", "pilot_dense_per_iter_s": 0.040344, "pilot_csr_per_iter_s": 1.094896, "pilot_coo_per_iter_s": 0.728494, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.197483, "rolv_iter_s": 0.001959, "dense_iter_s": 0.041347, "csr_iter_s": 1.112414, "coo_iter_s": 0.729467, "ell_iter_s": null, "rolv_total_s": 2.156957, "baseline_total_s": 41.347148, "speedup_total_vs_selected_x": 19.169, "speedup_iter_vs_selected_x": 21.101, "rolv_vs_vendor_sparse_iter_x": 567.711, "rolv_vs_vendor_sparse_total_x": 515.733, "rolv_vs_coo_iter_x": 372.277, "rolv_vs_coo_total_x": 338.193, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 11:58:24] Seed: 123456 | Pattern: banded | Zeros: 50% A_hash: 36930b864e45f6c7bc4c05a36ceed9e5546aba4f26c38e27ec94b84500ab052f | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.034089s | CSR: 0.042786s | COO: 0.028860s | ELL: nans Selected baseline: COO rolv load time (operator build): 0.188691 s rolv per-iter: 0.001941s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 0fe031672a78ac00d079d51b2c3b1ad3e4eb1c0428bd1bc66b4a2f100f6a7234 (COO) CSR_norm_hash: 0fe031672a78ac00d079d51b2c3b1ad3e4eb1c0428bd1bc66b4a2f100f6a7234 COO_norm_hash: 0fe031672a78ac00d079d51b2c3b1ad3e4eb1c0428bd1bc66b4a2f100f6a7234 ELL_norm_hash: 0fe031672a78ac00d079d51b2c3b1ad3e4eb1c0428bd1bc66b4a2f100f6a7234 ROLF_norm_hash: 0621305b553d14c934dc235dd9616ff421a67e87342f046923dc5da9bed27029 DENGS_norm_hash: 0fe031672a78ac00d079d51b2c3b1ad3e4eb1c0428bd1bc66b4a2f100f6a7234 COO per-iter: 0.029019s | total: 29.019338s CSR per-iter: 0.043816s | total: 43.816020s ROLF per-iter: 0.000220s | total: 0.220847s DENGS per-iter: 0.035206s | total: 35.205934s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 13.63x (≈ 1263% faster) Speedup (per-iter): 14.96x (≈ 1396% faster) Energy Savings: 93.31% rolv vs rocSPARSE -> Speedup (per-iter): 22.58x | total: 20.58x rolv vs COO: Speedup (per-iter): 14.95x | total: 13.63x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "36930b864e45f6c7bc4c05a36ceed9e5546aba4f26c38e27ec94b84500ab052f", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "0fe031672a78ac00d079d51b2c3b1ad3e4eb1c0428bd1bc66b4a2f100f6a7234", "CSR_norm_hash": "0fe031672a78ac00d079d51b2c3b1ad3e4eb1c0428bd1bc66b4a2f100f6a7234", "COO_norm_hash": "0fe031672a78ac00d079d51b2c3b1ad3e4eb1c0428bd1bc66b4a2f100f6a7234", "ELL_norm_hash": "0fe031672a78ac00d079d51b2c3b1ad3e4eb1c0428bd1bc66b4a2f100f6a7234", "ROLF_norm_hash": "0621305b553d14c934dc235dd9616ff421a67e87342f046923dc5da9bed27029", "DENGS_norm_hash": "0fe031672a78ac00d079d51b2c3b1ad3e4eb1c0428bd1bc66b4a2f100f6a7234", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "91a6790e57791bfb5eb9aeab2ad3128bc32fc47fb2b6dcd8728760921b04c533", "CSR_qhash_d6": "91a6790e57791bfb5eb9aeab2ad3128bc32fc47fb2b6dcd8728760921b04c533", "COO_qhash_d6": "91a6790e57791bfb5eb9aeab2ad3128bc32fc47fb2b6dcd8728760921b04c533", "ELL_qhash_d6": null, "path_selected": "COO", "pilot_dense_per_iter_s": 0.034089, "pilot_csr_per_iter_s": 0.042786, "pilot_coo_per_iter_s": 0.02886, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.188691, "rolv_iter_s": 0.001941, "dense_iter_s": 0.029022, "csr_iter_s": 0.043816, "coo_iter_s": 0.029019, "ell_iter_s": null, "rolv_total_s": 2.129302, "baseline_total_s": 29.022242, "speedup_total_vs_selected_x": 13.63, "speedup_iter_vs_selected_x": 14.955, "rolv_vs_vendor_sparse_iter_x": 22.578, "rolv_vs_vendor_sparse_total_x": 20.578, "rolv_vs_coo_iter_x": 14.954, "rolv_vs_coo_total_x": 13.629, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 12:01:25] Seed: 123456 | Pattern: block_diagonal | Zeros: 50% A_hash: 8db5189cd07996217967440640b6d42a07f04d0966354d2bccdba45b8f0e85b6 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.033326s | CSR: 0.028134s | COO: 0.020416s | ELL: nans Selected baseline: COO rolv load time (operator build): 0.187951 s rolv per-iter: 0.001954s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 03f3a34956a323cf0aaab8acfc428958d3cdffa022221a76104fbccce492ff66 (COO) CSR_norm_hash: 03f3a34956a323cf0aaab8acfc428958d3cdffa022221a76104fbccce492ff66 COO_norm_hash: 03f3a34956a323cf0aaab8acfc428958d3cdffa022221a76104fbccce492ff66 ELL_norm_hash: 03f3a34956a323cf0aaab8acfc428958d3cdffa022221a76104fbccce492ff66 ROLF_norm_hash: 1b772b5257f1ebb4c0e1b3eb7d087b0a42bbbeb650b2fd543326e966e21a2cb4 DENGS_norm_hash: 03f3a34956a323cf0aaab8acfc428958d3cdffa022221a76104fbccce492ff66 COO per-iter: 0.020461s | total: 20.461258s CSR per-iter: 0.028644s | total: 28.644160s ROLF per-iter: 0.000213s | total: 0.213368s DENGS per-iter: 0.034264s | total: 34.264289s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 9.58x (≈ 858% faster) Speedup (per-iter): 10.50x (≈ 950% faster) Energy Savings: 90.48% rolv vs rocSPARSE -> Speedup (per-iter): 14.66x | total: 13.38x rolv vs COO: Speedup (per-iter): 10.47x | total: 9.55x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "8db5189cd07996217967440640b6d42a07f04d0966354d2bccdba45b8f0e85b6", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "03f3a34956a323cf0aaab8acfc428958d3cdffa022221a76104fbccce492ff66", "CSR_norm_hash": "03f3a34956a323cf0aaab8acfc428958d3cdffa022221a76104fbccce492ff66", "COO_norm_hash": "03f3a34956a323cf0aaab8acfc428958d3cdffa022221a76104fbccce492ff66", "ELL_norm_hash": "03f3a34956a323cf0aaab8acfc428958d3cdffa022221a76104fbccce492ff66", "ROLF_norm_hash": "1b772b5257f1ebb4c0e1b3eb7d087b0a42bbbeb650b2fd543326e966e21a2cb4", "DENGS_norm_hash": "03f3a34956a323cf0aaab8acfc428958d3cdffa022221a76104fbccce492ff66", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "b5f52a9ef8ffd7efcef97cbd495082fcb11cd7cd2264e12ccaebab8950e1f08e", "CSR_qhash_d6": "b5f52a9ef8ffd7efcef97cbd495082fcb11cd7cd2264e12ccaebab8950e1f08e", "COO_qhash_d6": "b5f52a9ef8ffd7efcef97cbd495082fcb11cd7cd2264e12ccaebab8950e1f08e", "ELL_qhash_d6": null, "path_selected": "COO", "pilot_dense_per_iter_s": 0.033326, "pilot_csr_per_iter_s": 0.028134, "pilot_coo_per_iter_s": 0.020416, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.187951, "rolv_iter_s": 0.001954, "dense_iter_s": 0.020518, "csr_iter_s": 0.028644, "coo_iter_s": 0.020461, "ell_iter_s": null, "rolv_total_s": 2.141545, "baseline_total_s": 20.51833, "speedup_total_vs_selected_x": 9.581, "speedup_iter_vs_selected_x": 10.503, "rolv_vs_vendor_sparse_iter_x": 14.662, "rolv_vs_vendor_sparse_total_x": 13.375, "rolv_vs_coo_iter_x": 10.474, "rolv_vs_coo_total_x": 9.554, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 12:03:53] Seed: 123456 | Pattern: random | Zeros: 60% A_hash: 3a128a12c751e2a52a9f05427ad881a4beeb441b1aa828f2c83dec9767075e14 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.039484s | CSR: 0.959474s | COO: 0.626496s | ELL: nans Selected baseline: Dense rolv load time (operator build): 0.194216 s rolv per-iter: 0.001946s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 82ff97b0c2c9d6b4e6a850bdbeec16cf158da8950cbefe522f043e059a8a944e (Dense) CSR_norm_hash: 82ff97b0c2c9d6b4e6a850bdbeec16cf158da8950cbefe522f043e059a8a944e COO_norm_hash: 82ff97b0c2c9d6b4e6a850bdbeec16cf158da8950cbefe522f043e059a8a944e ELL_norm_hash: 82ff97b0c2c9d6b4e6a850bdbeec16cf158da8950cbefe522f043e059a8a944e ROLF_norm_hash: 53ee64013770fdb75b32a6e92743de71f64c87429155a238a486d4601379ff1b DENGS_norm_hash: 82ff97b0c2c9d6b4e6a850bdbeec16cf158da8950cbefe522f043e059a8a944e COO per-iter: 0.628151s | total: 628.151375s CSR per-iter: 0.966774s | total: 966.774188s ROLF per-iter: 0.000238s | total: 0.238523s DENGS per-iter: 0.040861s | total: 40.860602s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 18.88x (≈ 1788% faster) Speedup (per-iter): 20.77x (≈ 1977% faster) Energy Savings: 95.18% rolv vs rocSPARSE -> Speedup (per-iter): 496.77x | total: 451.69x rolv vs COO: Speedup (per-iter): 322.77x | total: 293.48x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "3a128a12c751e2a52a9f05427ad881a4beeb441b1aa828f2c83dec9767075e14", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "82ff97b0c2c9d6b4e6a850bdbeec16cf158da8950cbefe522f043e059a8a944e", "CSR_norm_hash": "82ff97b0c2c9d6b4e6a850bdbeec16cf158da8950cbefe522f043e059a8a944e", "COO_norm_hash": "82ff97b0c2c9d6b4e6a850bdbeec16cf158da8950cbefe522f043e059a8a944e", "ELL_norm_hash": "82ff97b0c2c9d6b4e6a850bdbeec16cf158da8950cbefe522f043e059a8a944e", "ROLF_norm_hash": "53ee64013770fdb75b32a6e92743de71f64c87429155a238a486d4601379ff1b", "DENGS_norm_hash": "82ff97b0c2c9d6b4e6a850bdbeec16cf158da8950cbefe522f043e059a8a944e", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "474ef2e5789ba96a76b96db5a6f8e76990d35228b24cd5d7660008bef1c3606c", "CSR_qhash_d6": "474ef2e5789ba96a76b96db5a6f8e76990d35228b24cd5d7660008bef1c3606c", "COO_qhash_d6": "474ef2e5789ba96a76b96db5a6f8e76990d35228b24cd5d7660008bef1c3606c", "ELL_qhash_d6": null, "path_selected": "Dense", "pilot_dense_per_iter_s": 0.039484, "pilot_csr_per_iter_s": 0.959474, "pilot_coo_per_iter_s": 0.626496, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.194216, "rolv_iter_s": 0.001946, "dense_iter_s": 0.040416, "csr_iter_s": 0.966774, "coo_iter_s": 0.628151, "ell_iter_s": null, "rolv_total_s": 2.140349, "baseline_total_s": 40.416102, "speedup_total_vs_selected_x": 18.883, "speedup_iter_vs_selected_x": 20.767, "rolv_vs_vendor_sparse_iter_x": 496.767, "rolv_vs_vendor_sparse_total_x": 451.69, "rolv_vs_coo_iter_x": 322.769, "rolv_vs_coo_total_x": 293.481, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 12:33:30] Seed: 123456 | Pattern: power_law | Zeros: 60% A_hash: 9d19ea5f391575455f95a6f93a0dc330f0816afb109185aa39e76d5e5e3f84a5 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.039558s | CSR: 0.896425s | COO: 0.586250s | ELL: nans Selected baseline: Dense rolv load time (operator build): 0.193746 s rolv per-iter: 0.001961s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 3397dfb188f303cce8ca1e8cfc9ceaf57b34d9574df64e8d752935e89f273568 (Dense) CSR_norm_hash: 3397dfb188f303cce8ca1e8cfc9ceaf57b34d9574df64e8d752935e89f273568 COO_norm_hash: 3397dfb188f303cce8ca1e8cfc9ceaf57b34d9574df64e8d752935e89f273568 ELL_norm_hash: 3397dfb188f303cce8ca1e8cfc9ceaf57b34d9574df64e8d752935e89f273568 ROLF_norm_hash: d20759f7172d2f43c943a230dae84150fb48c05b8b882f77935ccb62e51e5023 DENGS_norm_hash: 3397dfb188f303cce8ca1e8cfc9ceaf57b34d9574df64e8d752935e89f273568 COO per-iter: 0.586949s | total: 586.948625s CSR per-iter: 0.908240s | total: 908.240188s ROLF per-iter: 0.000243s | total: 0.243242s DENGS per-iter: 0.041096s | total: 41.095828s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 18.90x (≈ 1790% faster) Speedup (per-iter): 20.77x (≈ 1977% faster) Energy Savings: 95.19% rolv vs rocSPARSE -> Speedup (per-iter): 463.17x | total: 421.52x rolv vs COO: Speedup (per-iter): 299.32x | total: 272.41x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "9d19ea5f391575455f95a6f93a0dc330f0816afb109185aa39e76d5e5e3f84a5", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "3397dfb188f303cce8ca1e8cfc9ceaf57b34d9574df64e8d752935e89f273568", "CSR_norm_hash": "3397dfb188f303cce8ca1e8cfc9ceaf57b34d9574df64e8d752935e89f273568", "COO_norm_hash": "3397dfb188f303cce8ca1e8cfc9ceaf57b34d9574df64e8d752935e89f273568", "ELL_norm_hash": "3397dfb188f303cce8ca1e8cfc9ceaf57b34d9574df64e8d752935e89f273568", "ROLF_norm_hash": "d20759f7172d2f43c943a230dae84150fb48c05b8b882f77935ccb62e51e5023", "DENGS_norm_hash": "3397dfb188f303cce8ca1e8cfc9ceaf57b34d9574df64e8d752935e89f273568", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "054e2c6d65e7082adb830742e210da6458ef0c3b993c9efeb6f8de4af5491a0b", "CSR_qhash_d6": "054e2c6d65e7082adb830742e210da6458ef0c3b993c9efeb6f8de4af5491a0b", "COO_qhash_d6": "054e2c6d65e7082adb830742e210da6458ef0c3b993c9efeb6f8de4af5491a0b", "ELL_qhash_d6": null, "path_selected": "Dense", "pilot_dense_per_iter_s": 0.039558, "pilot_csr_per_iter_s": 0.896425, "pilot_coo_per_iter_s": 0.58625, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.193746, "rolv_iter_s": 0.001961, "dense_iter_s": 0.040733, "csr_iter_s": 0.90824, "coo_iter_s": 0.586949, "ell_iter_s": null, "rolv_total_s": 2.154674, "baseline_total_s": 40.732828, "speedup_total_vs_selected_x": 18.904, "speedup_iter_vs_selected_x": 20.772, "rolv_vs_vendor_sparse_iter_x": 463.169, "rolv_vs_vendor_sparse_total_x": 421.521, "rolv_vs_coo_iter_x": 299.322, "rolv_vs_coo_total_x": 272.407, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 13:01:21] Seed: 123456 | Pattern: banded | Zeros: 60% A_hash: e78e035e07d681d9c88788fb30448528322d3759de0292aef1030acc8d438be2 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.034519s | CSR: 0.035773s | COO: 0.025788s | ELL: nans Selected baseline: COO rolv load time (operator build): 0.189412 s rolv per-iter: 0.001953s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: a917726a5ab831eb4b9c1cbef78f9dab9bf38f1875cc27bd0c7e1a74d85cd51a (COO) CSR_norm_hash: a917726a5ab831eb4b9c1cbef78f9dab9bf38f1875cc27bd0c7e1a74d85cd51a COO_norm_hash: a917726a5ab831eb4b9c1cbef78f9dab9bf38f1875cc27bd0c7e1a74d85cd51a ELL_norm_hash: a917726a5ab831eb4b9c1cbef78f9dab9bf38f1875cc27bd0c7e1a74d85cd51a ROLF_norm_hash: 875e17c75ff856b76232765109555a25b7fbfd369a523cfcfc37fecd0cf0a765 DENGS_norm_hash: a917726a5ab831eb4b9c1cbef78f9dab9bf38f1875cc27bd0c7e1a74d85cd51a COO per-iter: 0.025822s | total: 25.821992s CSR per-iter: 0.036532s | total: 36.531539s ROLF per-iter: 0.000219s | total: 0.219839s DENGS per-iter: 0.035182s | total: 35.181711s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 12.06x (≈ 1106% faster) Speedup (per-iter): 13.23x (≈ 1223% faster) Energy Savings: 92.44% rolv vs rocSPARSE -> Speedup (per-iter): 18.70x | total: 17.05x rolv vs COO: Speedup (per-iter): 13.22x | total: 12.05x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "e78e035e07d681d9c88788fb30448528322d3759de0292aef1030acc8d438be2", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "a917726a5ab831eb4b9c1cbef78f9dab9bf38f1875cc27bd0c7e1a74d85cd51a", "CSR_norm_hash": "a917726a5ab831eb4b9c1cbef78f9dab9bf38f1875cc27bd0c7e1a74d85cd51a", "COO_norm_hash": "a917726a5ab831eb4b9c1cbef78f9dab9bf38f1875cc27bd0c7e1a74d85cd51a", "ELL_norm_hash": "a917726a5ab831eb4b9c1cbef78f9dab9bf38f1875cc27bd0c7e1a74d85cd51a", "ROLF_norm_hash": "875e17c75ff856b76232765109555a25b7fbfd369a523cfcfc37fecd0cf0a765", "DENGS_norm_hash": "a917726a5ab831eb4b9c1cbef78f9dab9bf38f1875cc27bd0c7e1a74d85cd51a", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "6c590cb1c86449e9a55609b2add184da816091d6e591141b6417902275b2cba6", "CSR_qhash_d6": "6c590cb1c86449e9a55609b2add184da816091d6e591141b6417902275b2cba6", "COO_qhash_d6": "6c590cb1c86449e9a55609b2add184da816091d6e591141b6417902275b2cba6", "ELL_qhash_d6": null, "path_selected": "COO", "pilot_dense_per_iter_s": 0.034519, "pilot_csr_per_iter_s": 0.035773, "pilot_coo_per_iter_s": 0.025788, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.189412, "rolv_iter_s": 0.001953, "dense_iter_s": 0.025846, "csr_iter_s": 0.036532, "coo_iter_s": 0.025822, "ell_iter_s": null, "rolv_total_s": 2.14259, "baseline_total_s": 25.846469, "speedup_total_vs_selected_x": 12.063, "speedup_iter_vs_selected_x": 13.233, "rolv_vs_vendor_sparse_iter_x": 18.704, "rolv_vs_vendor_sparse_total_x": 17.05, "rolv_vs_coo_iter_x": 13.221, "rolv_vs_coo_total_x": 12.052, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 13:04:07] Seed: 123456 | Pattern: block_diagonal | Zeros: 60% A_hash: 2b99793bda656b5689cc9f5b049fc1a55ae8c234e0386e439c7204b281ffc158 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.033211s | CSR: 0.023548s | COO: 0.017100s | ELL: nans Selected baseline: COO rolv load time (operator build): 0.188862 s rolv per-iter: 0.001948s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 36ee25dfb91d647bc72a00e82673d5af290686eb222c108851af0797263cbfc4 (COO) CSR_norm_hash: 36ee25dfb91d647bc72a00e82673d5af290686eb222c108851af0797263cbfc4 COO_norm_hash: 36ee25dfb91d647bc72a00e82673d5af290686eb222c108851af0797263cbfc4 ELL_norm_hash: 36ee25dfb91d647bc72a00e82673d5af290686eb222c108851af0797263cbfc4 ROLF_norm_hash: 968c6fc2b0accf1cdc4ab0d5115cfd7051533a5d02294705346da1e28a71c50b DENGS_norm_hash: 36ee25dfb91d647bc72a00e82673d5af290686eb222c108851af0797263cbfc4 COO per-iter: 0.017170s | total: 17.170432s CSR per-iter: 0.024001s | total: 24.000824s ROLF per-iter: 0.000212s | total: 0.212729s DENGS per-iter: 0.034141s | total: 34.141016s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 8.02x (≈ 702% faster) Speedup (per-iter): 8.80x (≈ 780% faster) Energy Savings: 88.63% rolv vs rocSPARSE -> Speedup (per-iter): 12.32x | total: 11.23x rolv vs COO: Speedup (per-iter): 8.81x | total: 8.03x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "2b99793bda656b5689cc9f5b049fc1a55ae8c234e0386e439c7204b281ffc158", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "36ee25dfb91d647bc72a00e82673d5af290686eb222c108851af0797263cbfc4", "CSR_norm_hash": "36ee25dfb91d647bc72a00e82673d5af290686eb222c108851af0797263cbfc4", "COO_norm_hash": "36ee25dfb91d647bc72a00e82673d5af290686eb222c108851af0797263cbfc4", "ELL_norm_hash": "36ee25dfb91d647bc72a00e82673d5af290686eb222c108851af0797263cbfc4", "ROLF_norm_hash": "968c6fc2b0accf1cdc4ab0d5115cfd7051533a5d02294705346da1e28a71c50b", "DENGS_norm_hash": "36ee25dfb91d647bc72a00e82673d5af290686eb222c108851af0797263cbfc4", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "4d9bc3a8c04e309be3d194ede6251942ddf9e71860fd8aa14f66686b71fd0279", "CSR_qhash_d6": "4d9bc3a8c04e309be3d194ede6251942ddf9e71860fd8aa14f66686b71fd0279", "COO_qhash_d6": "4d9bc3a8c04e309be3d194ede6251942ddf9e71860fd8aa14f66686b71fd0279", "ELL_qhash_d6": null, "path_selected": "COO", "pilot_dense_per_iter_s": 0.033211, "pilot_csr_per_iter_s": 0.023548, "pilot_coo_per_iter_s": 0.0171, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.188862, "rolv_iter_s": 0.001948, "dense_iter_s": 0.017136, "csr_iter_s": 0.024001, "coo_iter_s": 0.01717, "ell_iter_s": null, "rolv_total_s": 2.137011, "baseline_total_s": 17.13609, "speedup_total_vs_selected_x": 8.019, "speedup_iter_vs_selected_x": 8.796, "rolv_vs_vendor_sparse_iter_x": 12.32, "rolv_vs_vendor_sparse_total_x": 11.231, "rolv_vs_coo_iter_x": 8.814, "rolv_vs_coo_total_x": 8.035, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 13:06:21] Seed: 123456 | Pattern: random | Zeros: 70% A_hash: b6d397e4d0e8ebd4f3a13d59f635831bd762ee60284807ed9d008435058ec326 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.038749s | CSR: 0.736745s | COO: 0.476107s | ELL: nans Selected baseline: Dense rolv load time (operator build): 0.189396 s rolv per-iter: 0.001947s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 722aa1f103b022093a749ccf9de9cf9003cb678bf7f25648ca7f11ed9adde915 (Dense) CSR_norm_hash: 722aa1f103b022093a749ccf9de9cf9003cb678bf7f25648ca7f11ed9adde915 COO_norm_hash: 722aa1f103b022093a749ccf9de9cf9003cb678bf7f25648ca7f11ed9adde915 ELL_norm_hash: 722aa1f103b022093a749ccf9de9cf9003cb678bf7f25648ca7f11ed9adde915 ROLF_norm_hash: a503fe60776ce90abef86e50f774bce0642268c79ed36f1ed2a0cfdb456b6090 DENGS_norm_hash: 722aa1f103b022093a749ccf9de9cf9003cb678bf7f25648ca7f11ed9adde915 COO per-iter: 0.474754s | total: 474.754062s CSR per-iter: 0.744779s | total: 744.778625s ROLF per-iter: 0.000235s | total: 0.236061s DENGS per-iter: 0.039967s | total: 39.967223s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 18.52x (≈ 1752% faster) Speedup (per-iter): 20.32x (≈ 1932% faster) Energy Savings: 95.08% rolv vs rocSPARSE -> Speedup (per-iter): 382.49x | total: 348.58x rolv vs COO: Speedup (per-iter): 243.81x | total: 222.20x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "b6d397e4d0e8ebd4f3a13d59f635831bd762ee60284807ed9d008435058ec326", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "722aa1f103b022093a749ccf9de9cf9003cb678bf7f25648ca7f11ed9adde915", "CSR_norm_hash": "722aa1f103b022093a749ccf9de9cf9003cb678bf7f25648ca7f11ed9adde915", "COO_norm_hash": "722aa1f103b022093a749ccf9de9cf9003cb678bf7f25648ca7f11ed9adde915", "ELL_norm_hash": "722aa1f103b022093a749ccf9de9cf9003cb678bf7f25648ca7f11ed9adde915", "ROLF_norm_hash": "a503fe60776ce90abef86e50f774bce0642268c79ed36f1ed2a0cfdb456b6090", "DENGS_norm_hash": "722aa1f103b022093a749ccf9de9cf9003cb678bf7f25648ca7f11ed9adde915", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "82da032e08a558947b8496a6df588242839c731dd132f6c51b2ccad1158e1e9e", "CSR_qhash_d6": "82da032e08a558947b8496a6df588242839c731dd132f6c51b2ccad1158e1e9e", "COO_qhash_d6": "82da032e08a558947b8496a6df588242839c731dd132f6c51b2ccad1158e1e9e", "ELL_qhash_d6": null, "path_selected": "Dense", "pilot_dense_per_iter_s": 0.038749, "pilot_csr_per_iter_s": 0.736745, "pilot_coo_per_iter_s": 0.476107, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.189396, "rolv_iter_s": 0.001947, "dense_iter_s": 0.039576, "csr_iter_s": 0.744779, "coo_iter_s": 0.474754, "ell_iter_s": null, "rolv_total_s": 2.136601, "baseline_total_s": 39.576336, "speedup_total_vs_selected_x": 18.523, "speedup_iter_vs_selected_x": 20.325, "rolv_vs_vendor_sparse_iter_x": 382.486, "rolv_vs_vendor_sparse_total_x": 348.581, "rolv_vs_coo_iter_x": 243.813, "rolv_vs_coo_total_x": 222.201, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 13:29:25] Seed: 123456 | Pattern: power_law | Zeros: 70% A_hash: 64b353290cc661d8798233b459b02627e318c8b6cd03fb9400cdc258605a7257 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.039186s | CSR: 0.691317s | COO: 0.443855s | ELL: nans Selected baseline: Dense rolv load time (operator build): 0.194456 s rolv per-iter: 0.001960s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 32c0bbfd6d7a5688cd46fb2b36d62e508c168c1fca43ba3125721e0a93b9b9dc (Dense) CSR_norm_hash: 32c0bbfd6d7a5688cd46fb2b36d62e508c168c1fca43ba3125721e0a93b9b9dc COO_norm_hash: 32c0bbfd6d7a5688cd46fb2b36d62e508c168c1fca43ba3125721e0a93b9b9dc ELL_norm_hash: 32c0bbfd6d7a5688cd46fb2b36d62e508c168c1fca43ba3125721e0a93b9b9dc ROLF_norm_hash: 72204d453e28b46f2f06b0ac265feda1932e914601162d00494a9d1bfa846619 DENGS_norm_hash: 32c0bbfd6d7a5688cd46fb2b36d62e508c168c1fca43ba3125721e0a93b9b9dc COO per-iter: 0.444273s | total: 444.273156s CSR per-iter: 0.696841s | total: 696.840937s ROLF per-iter: 0.000239s | total: 0.239301s DENGS per-iter: 0.040210s | total: 40.210113s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 18.48x (≈ 1748% faster) Speedup (per-iter): 20.31x (≈ 1931% faster) Energy Savings: 95.08% rolv vs rocSPARSE -> Speedup (per-iter): 355.50x | total: 323.41x rolv vs COO: Speedup (per-iter): 226.65x | total: 206.19x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "64b353290cc661d8798233b459b02627e318c8b6cd03fb9400cdc258605a7257", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "32c0bbfd6d7a5688cd46fb2b36d62e508c168c1fca43ba3125721e0a93b9b9dc", "CSR_norm_hash": "32c0bbfd6d7a5688cd46fb2b36d62e508c168c1fca43ba3125721e0a93b9b9dc", "COO_norm_hash": "32c0bbfd6d7a5688cd46fb2b36d62e508c168c1fca43ba3125721e0a93b9b9dc", "ELL_norm_hash": "32c0bbfd6d7a5688cd46fb2b36d62e508c168c1fca43ba3125721e0a93b9b9dc", "ROLF_norm_hash": "72204d453e28b46f2f06b0ac265feda1932e914601162d00494a9d1bfa846619", "DENGS_norm_hash": "32c0bbfd6d7a5688cd46fb2b36d62e508c168c1fca43ba3125721e0a93b9b9dc", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "b28317e48cc7768973ac901f2af18502a2b95897bacec4775b55b8b869b4083f", "CSR_qhash_d6": "b28317e48cc7768973ac901f2af18502a2b95897bacec4775b55b8b869b4083f", "COO_qhash_d6": "b28317e48cc7768973ac901f2af18502a2b95897bacec4775b55b8b869b4083f", "ELL_qhash_d6": null, "path_selected": "Dense", "pilot_dense_per_iter_s": 0.039186, "pilot_csr_per_iter_s": 0.691317, "pilot_coo_per_iter_s": 0.443855, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.194456, "rolv_iter_s": 0.00196, "dense_iter_s": 0.039809, "csr_iter_s": 0.696841, "coo_iter_s": 0.444273, "ell_iter_s": null, "rolv_total_s": 2.15464, "baseline_total_s": 39.809414, "speedup_total_vs_selected_x": 18.476, "speedup_iter_vs_selected_x": 20.309, "rolv_vs_vendor_sparse_iter_x": 355.498, "rolv_vs_vendor_sparse_total_x": 323.414, "rolv_vs_coo_iter_x": 226.649, "rolv_vs_coo_total_x": 206.194, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 13:51:10] Seed: 123456 | Pattern: banded | Zeros: 70% A_hash: 6de52c734dc3dd3e441813467d3974c05babbe147880af95cae93106e22a77bd | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.034102s | CSR: 0.028236s | COO: 0.019748s | ELL: nans Selected baseline: COO rolv load time (operator build): 0.188387 s rolv per-iter: 0.001950s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: afa0b5ccf007ed78efa389e675d49ed3175a5e895800ce2c51b65ef34c1c93f8 (COO) CSR_norm_hash: afa0b5ccf007ed78efa389e675d49ed3175a5e895800ce2c51b65ef34c1c93f8 COO_norm_hash: afa0b5ccf007ed78efa389e675d49ed3175a5e895800ce2c51b65ef34c1c93f8 ELL_norm_hash: afa0b5ccf007ed78efa389e675d49ed3175a5e895800ce2c51b65ef34c1c93f8 ROLF_norm_hash: 0f39b126fa71c26f2672d66e7c24c5b70fa1e325cbf4ccf3333c8d2df71581bf DENGS_norm_hash: afa0b5ccf007ed78efa389e675d49ed3175a5e895800ce2c51b65ef34c1c93f8 COO per-iter: 0.019814s | total: 19.813516s CSR per-iter: 0.028746s | total: 28.746150s ROLF per-iter: 0.000218s | total: 0.218054s DENGS per-iter: 0.035194s | total: 35.194129s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 9.24x (≈ 824% faster) Speedup (per-iter): 10.13x (≈ 913% faster) Energy Savings: 90.13% rolv vs rocSPARSE -> Speedup (per-iter): 14.74x | total: 13.44x rolv vs COO: Speedup (per-iter): 10.16x | total: 9.26x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "6de52c734dc3dd3e441813467d3974c05babbe147880af95cae93106e22a77bd", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "afa0b5ccf007ed78efa389e675d49ed3175a5e895800ce2c51b65ef34c1c93f8", "CSR_norm_hash": "afa0b5ccf007ed78efa389e675d49ed3175a5e895800ce2c51b65ef34c1c93f8", "COO_norm_hash": "afa0b5ccf007ed78efa389e675d49ed3175a5e895800ce2c51b65ef34c1c93f8", "ELL_norm_hash": "afa0b5ccf007ed78efa389e675d49ed3175a5e895800ce2c51b65ef34c1c93f8", "ROLF_norm_hash": "0f39b126fa71c26f2672d66e7c24c5b70fa1e325cbf4ccf3333c8d2df71581bf", "DENGS_norm_hash": "afa0b5ccf007ed78efa389e675d49ed3175a5e895800ce2c51b65ef34c1c93f8", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "34587fe968cb118281d6e320a80b1d361638e54fc3de87df1dbf85b3f83c9fef", "CSR_qhash_d6": "34587fe968cb118281d6e320a80b1d361638e54fc3de87df1dbf85b3f83c9fef", "COO_qhash_d6": "34587fe968cb118281d6e320a80b1d361638e54fc3de87df1dbf85b3f83c9fef", "ELL_qhash_d6": null, "path_selected": "COO", "pilot_dense_per_iter_s": 0.034102, "pilot_csr_per_iter_s": 0.028236, "pilot_coo_per_iter_s": 0.019748, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.188387, "rolv_iter_s": 0.00195, "dense_iter_s": 0.019766, "csr_iter_s": 0.028746, "coo_iter_s": 0.019814, "ell_iter_s": null, "rolv_total_s": 2.138754, "baseline_total_s": 19.766223, "speedup_total_vs_selected_x": 9.242, "speedup_iter_vs_selected_x": 10.135, "rolv_vs_vendor_sparse_iter_x": 14.739, "rolv_vs_vendor_sparse_total_x": 13.441, "rolv_vs_coo_iter_x": 10.159, "rolv_vs_coo_total_x": 9.264, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 13:53:35] Seed: 123456 | Pattern: block_diagonal | Zeros: 70% A_hash: 605ad79227a409511ccd935bac7446d55792ae15e0550623f778311797a2ba80 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.033115s | CSR: 0.018917s | COO: 0.013868s | ELL: nans Selected baseline: COO rolv load time (operator build): 0.187606 s rolv per-iter: 0.001949s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: afb0000c8a8069b1416a99dc37be6c761f158e933ad2069f4c2a71f2a7f8ef75 (COO) CSR_norm_hash: afb0000c8a8069b1416a99dc37be6c761f158e933ad2069f4c2a71f2a7f8ef75 COO_norm_hash: afb0000c8a8069b1416a99dc37be6c761f158e933ad2069f4c2a71f2a7f8ef75 ELL_norm_hash: afb0000c8a8069b1416a99dc37be6c761f158e933ad2069f4c2a71f2a7f8ef75 ROLF_norm_hash: 71235fd9a701d922ba41c2836607dbd277b18a94a84c0fe54ad3e3cb46430593 DENGS_norm_hash: afb0000c8a8069b1416a99dc37be6c761f158e933ad2069f4c2a71f2a7f8ef75 COO per-iter: 0.013882s | total: 13.882259s CSR per-iter: 0.019216s | total: 19.215646s ROLF per-iter: 0.000212s | total: 0.212458s DENGS per-iter: 0.034091s | total: 34.091449s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 6.48x (≈ 548% faster) Speedup (per-iter): 7.10x (≈ 610% faster) Energy Savings: 85.92% rolv vs rocSPARSE -> Speedup (per-iter): 9.86x | total: 8.99x rolv vs COO: Speedup (per-iter): 7.12x | total: 6.50x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "605ad79227a409511ccd935bac7446d55792ae15e0550623f778311797a2ba80", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "afb0000c8a8069b1416a99dc37be6c761f158e933ad2069f4c2a71f2a7f8ef75", "CSR_norm_hash": "afb0000c8a8069b1416a99dc37be6c761f158e933ad2069f4c2a71f2a7f8ef75", "COO_norm_hash": "afb0000c8a8069b1416a99dc37be6c761f158e933ad2069f4c2a71f2a7f8ef75", "ELL_norm_hash": "afb0000c8a8069b1416a99dc37be6c761f158e933ad2069f4c2a71f2a7f8ef75", "ROLF_norm_hash": "71235fd9a701d922ba41c2836607dbd277b18a94a84c0fe54ad3e3cb46430593", "DENGS_norm_hash": "afb0000c8a8069b1416a99dc37be6c761f158e933ad2069f4c2a71f2a7f8ef75", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "7bc340a2266a60176174c6afbc41e54623a3acb3c008de5aeff26276a7332fb6", "CSR_qhash_d6": "7bc340a2266a60176174c6afbc41e54623a3acb3c008de5aeff26276a7332fb6", "COO_qhash_d6": "7bc340a2266a60176174c6afbc41e54623a3acb3c008de5aeff26276a7332fb6", "ELL_qhash_d6": null, "path_selected": "COO", "pilot_dense_per_iter_s": 0.033115, "pilot_csr_per_iter_s": 0.018917, "pilot_coo_per_iter_s": 0.013868, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.187606, "rolv_iter_s": 0.001949, "dense_iter_s": 0.013846, "csr_iter_s": 0.019216, "coo_iter_s": 0.013882, "ell_iter_s": null, "rolv_total_s": 2.136825, "baseline_total_s": 13.846479, "speedup_total_vs_selected_x": 6.48, "speedup_iter_vs_selected_x": 7.104, "rolv_vs_vendor_sparse_iter_x": 9.858, "rolv_vs_vendor_sparse_total_x": 8.993, "rolv_vs_coo_iter_x": 7.122, "rolv_vs_coo_total_x": 6.497, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 13:55:38] Seed: 123456 | Pattern: random | Zeros: 80% A_hash: fe8ecd469d65375943070e2c9f72b2cb8ffc99f59b8e95e01ee55ff351e8a5b5 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.038051s | CSR: 0.507866s | COO: 0.320311s | ELL: nans Selected baseline: Dense rolv load time (operator build): 0.189788 s rolv per-iter: 0.001961s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: e7c01a70a75e7c23f6388af2f7da803ea0bdb8bf7a90f33bfbd68ec38023b5fb (Dense) CSR_norm_hash: e7c01a70a75e7c23f6388af2f7da803ea0bdb8bf7a90f33bfbd68ec38023b5fb COO_norm_hash: e7c01a70a75e7c23f6388af2f7da803ea0bdb8bf7a90f33bfbd68ec38023b5fb ELL_norm_hash: e7c01a70a75e7c23f6388af2f7da803ea0bdb8bf7a90f33bfbd68ec38023b5fb ROLF_norm_hash: 2c97f0d1200024546d7641880cac8a28d456f33e2b9d9806d0408d519cd56f37 DENGS_norm_hash: e7c01a70a75e7c23f6388af2f7da803ea0bdb8bf7a90f33bfbd68ec38023b5fb COO per-iter: 0.320794s | total: 320.793687s CSR per-iter: 0.512847s | total: 512.846906s ROLF per-iter: 0.000234s | total: 0.234549s DENGS per-iter: 0.038902s | total: 38.901582s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 17.94x (≈ 1694% faster) Speedup (per-iter): 19.68x (≈ 1868% faster) Energy Savings: 94.92% rolv vs rocSPARSE -> Speedup (per-iter): 261.55x | total: 238.47x rolv vs COO: Speedup (per-iter): 163.60x | total: 149.16x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "fe8ecd469d65375943070e2c9f72b2cb8ffc99f59b8e95e01ee55ff351e8a5b5", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "e7c01a70a75e7c23f6388af2f7da803ea0bdb8bf7a90f33bfbd68ec38023b5fb", "CSR_norm_hash": "e7c01a70a75e7c23f6388af2f7da803ea0bdb8bf7a90f33bfbd68ec38023b5fb", "COO_norm_hash": "e7c01a70a75e7c23f6388af2f7da803ea0bdb8bf7a90f33bfbd68ec38023b5fb", "ELL_norm_hash": "e7c01a70a75e7c23f6388af2f7da803ea0bdb8bf7a90f33bfbd68ec38023b5fb", "ROLF_norm_hash": "2c97f0d1200024546d7641880cac8a28d456f33e2b9d9806d0408d519cd56f37", "DENGS_norm_hash": "e7c01a70a75e7c23f6388af2f7da803ea0bdb8bf7a90f33bfbd68ec38023b5fb", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "73400dd11bba92807f9dd80ee8c7bdc5e578106e1ea67465c31cebca0c1dc833", "CSR_qhash_d6": "73400dd11bba92807f9dd80ee8c7bdc5e578106e1ea67465c31cebca0c1dc833", "COO_qhash_d6": "73400dd11bba92807f9dd80ee8c7bdc5e578106e1ea67465c31cebca0c1dc833", "ELL_qhash_d6": null, "path_selected": "Dense", "pilot_dense_per_iter_s": 0.038051, "pilot_csr_per_iter_s": 0.507866, "pilot_coo_per_iter_s": 0.320311, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.189788, "rolv_iter_s": 0.001961, "dense_iter_s": 0.038591, "csr_iter_s": 0.512847, "coo_iter_s": 0.320794, "ell_iter_s": null, "rolv_total_s": 2.150599, "baseline_total_s": 38.591273, "speedup_total_vs_selected_x": 17.944, "speedup_iter_vs_selected_x": 19.681, "rolv_vs_vendor_sparse_iter_x": 261.548, "rolv_vs_vendor_sparse_total_x": 238.467, "rolv_vs_coo_iter_x": 163.603, "rolv_vs_coo_total_x": 149.165, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 14:12:00] Seed: 123456 | Pattern: power_law | Zeros: 80% A_hash: f5319945ed9e0de80929153636dd5033761020445fb403b1998eb9214d00e127 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.038282s | CSR: 0.477782s | COO: 0.299696s | ELL: nans Selected baseline: Dense rolv load time (operator build): 0.188725 s rolv per-iter: 0.001945s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 0bbec08c037006aa33c812b530df9811cc2768a08858d9d8f610d0e9dc3f1048 (Dense) CSR_norm_hash: 0bbec08c037006aa33c812b530df9811cc2768a08858d9d8f610d0e9dc3f1048 COO_norm_hash: 0bbec08c037006aa33c812b530df9811cc2768a08858d9d8f610d0e9dc3f1048 ELL_norm_hash: 0bbec08c037006aa33c812b530df9811cc2768a08858d9d8f610d0e9dc3f1048 ROLF_norm_hash: 222530270f5d311d6cd32e2219aefd7aa591c2a33d3143d33a5f01187d0d6e1b DENGS_norm_hash: 0bbec08c037006aa33c812b530df9811cc2768a08858d9d8f610d0e9dc3f1048 COO per-iter: 0.300233s | total: 300.233094s CSR per-iter: 0.480361s | total: 480.360813s ROLF per-iter: 0.000235s | total: 0.235684s DENGS per-iter: 0.039212s | total: 39.211727s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 18.20x (≈ 1720% faster) Speedup (per-iter): 19.96x (≈ 1896% faster) Energy Savings: 94.99% rolv vs rocSPARSE -> Speedup (per-iter): 246.93x | total: 225.10x rolv vs COO: Speedup (per-iter): 154.34x | total: 140.69x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "f5319945ed9e0de80929153636dd5033761020445fb403b1998eb9214d00e127", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "0bbec08c037006aa33c812b530df9811cc2768a08858d9d8f610d0e9dc3f1048", "CSR_norm_hash": "0bbec08c037006aa33c812b530df9811cc2768a08858d9d8f610d0e9dc3f1048", "COO_norm_hash": "0bbec08c037006aa33c812b530df9811cc2768a08858d9d8f610d0e9dc3f1048", "ELL_norm_hash": "0bbec08c037006aa33c812b530df9811cc2768a08858d9d8f610d0e9dc3f1048", "ROLF_norm_hash": "222530270f5d311d6cd32e2219aefd7aa591c2a33d3143d33a5f01187d0d6e1b", "DENGS_norm_hash": "0bbec08c037006aa33c812b530df9811cc2768a08858d9d8f610d0e9dc3f1048", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "66317c61bd76f15b22946d59ad005fceac533aed498ef9ff62ad0904ec173c58", "CSR_qhash_d6": "66317c61bd76f15b22946d59ad005fceac533aed498ef9ff62ad0904ec173c58", "COO_qhash_d6": "66317c61bd76f15b22946d59ad005fceac533aed498ef9ff62ad0904ec173c58", "ELL_qhash_d6": null, "path_selected": "Dense", "pilot_dense_per_iter_s": 0.038282, "pilot_csr_per_iter_s": 0.477782, "pilot_coo_per_iter_s": 0.299696, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.188725, "rolv_iter_s": 0.001945, "dense_iter_s": 0.038834, "csr_iter_s": 0.480361, "coo_iter_s": 0.300233, "ell_iter_s": null, "rolv_total_s": 2.134031, "baseline_total_s": 38.834016, "speedup_total_vs_selected_x": 18.197, "speedup_iter_vs_selected_x": 19.963, "rolv_vs_vendor_sparse_iter_x": 246.933, "rolv_vs_vendor_sparse_total_x": 225.095, "rolv_vs_coo_iter_x": 154.337, "rolv_vs_coo_total_x": 140.688, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 14:27:29] Seed: 123456 | Pattern: banded | Zeros: 80% A_hash: b2fc7f83b499ca9e4b29ed3cc68b966b4b322cf7926c12186e98ae033e84be58 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.033762s | CSR: 0.020752s | COO: 0.013558s | ELL: nans Selected baseline: COO rolv load time (operator build): 0.187645 s rolv per-iter: 0.001936s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 7c5ad9bbb7deb3f95a8b23ba1ddfb19f857bf6120ba461e883e8789abd82d6c6 (COO) CSR_norm_hash: 7c5ad9bbb7deb3f95a8b23ba1ddfb19f857bf6120ba461e883e8789abd82d6c6 COO_norm_hash: 7c5ad9bbb7deb3f95a8b23ba1ddfb19f857bf6120ba461e883e8789abd82d6c6 ELL_norm_hash: 7c5ad9bbb7deb3f95a8b23ba1ddfb19f857bf6120ba461e883e8789abd82d6c6 ROLF_norm_hash: e59dd960a001f4488b5b3c7cdcaa8d75712839372bc52a94bab002baba02fd9d DENGS_norm_hash: 7c5ad9bbb7deb3f95a8b23ba1ddfb19f857bf6120ba461e883e8789abd82d6c6 COO per-iter: 0.013633s | total: 13.633490s CSR per-iter: 0.021095s | total: 21.095234s ROLF per-iter: 0.000214s | total: 0.215017s DENGS per-iter: 0.034992s | total: 34.991766s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 6.39x (≈ 539% faster) Speedup (per-iter): 7.01x (≈ 601% faster) Energy Savings: 85.73% rolv vs rocSPARSE -> Speedup (per-iter): 10.89x | total: 9.93x rolv vs COO: Speedup (per-iter): 7.04x | total: 6.42x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "b2fc7f83b499ca9e4b29ed3cc68b966b4b322cf7926c12186e98ae033e84be58", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "7c5ad9bbb7deb3f95a8b23ba1ddfb19f857bf6120ba461e883e8789abd82d6c6", "CSR_norm_hash": "7c5ad9bbb7deb3f95a8b23ba1ddfb19f857bf6120ba461e883e8789abd82d6c6", "COO_norm_hash": "7c5ad9bbb7deb3f95a8b23ba1ddfb19f857bf6120ba461e883e8789abd82d6c6", "ELL_norm_hash": "7c5ad9bbb7deb3f95a8b23ba1ddfb19f857bf6120ba461e883e8789abd82d6c6", "ROLF_norm_hash": "e59dd960a001f4488b5b3c7cdcaa8d75712839372bc52a94bab002baba02fd9d", "DENGS_norm_hash": "7c5ad9bbb7deb3f95a8b23ba1ddfb19f857bf6120ba461e883e8789abd82d6c6", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "3d7a5452a9f38bbfb38ee0d4922ca8020b2506f811ff5ec119664b4fe4236084", "CSR_qhash_d6": "3d7a5452a9f38bbfb38ee0d4922ca8020b2506f811ff5ec119664b4fe4236084", "COO_qhash_d6": "3d7a5452a9f38bbfb38ee0d4922ca8020b2506f811ff5ec119664b4fe4236084", "ELL_qhash_d6": null, "path_selected": "COO", "pilot_dense_per_iter_s": 0.033762, "pilot_csr_per_iter_s": 0.020752, "pilot_coo_per_iter_s": 0.013558, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.187645, "rolv_iter_s": 0.001936, "dense_iter_s": 0.013572, "csr_iter_s": 0.021095, "coo_iter_s": 0.013633, "ell_iter_s": null, "rolv_total_s": 2.123954, "baseline_total_s": 13.571749, "speedup_total_vs_selected_x": 6.39, "speedup_iter_vs_selected_x": 7.009, "rolv_vs_vendor_sparse_iter_x": 10.895, "rolv_vs_vendor_sparse_total_x": 9.932, "rolv_vs_coo_iter_x": 7.041, "rolv_vs_coo_total_x": 6.419, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 14:29:33] Seed: 123456 | Pattern: block_diagonal | Zeros: 80% A_hash: 4b02e483523fbec343feac2b8fed3820615bb6832dda42a3da7b63ccf1ef0014 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.033382s | CSR: 0.014278s | COO: 0.011023s | ELL: nans Selected baseline: COO rolv load time (operator build): 0.187499 s rolv per-iter: 0.001950s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 0afabd3c3f898124c4d2adf90b4ec9ca28ee520d74975002112bb8408f023a82 (COO) CSR_norm_hash: 0afabd3c3f898124c4d2adf90b4ec9ca28ee520d74975002112bb8408f023a82 COO_norm_hash: 0afabd3c3f898124c4d2adf90b4ec9ca28ee520d74975002112bb8408f023a82 ELL_norm_hash: 0afabd3c3f898124c4d2adf90b4ec9ca28ee520d74975002112bb8408f023a82 ROLF_norm_hash: 50f833cd0edb871dab70fd1076fd14286647900fd24dbd3131295de2f48e9f7b DENGS_norm_hash: 0afabd3c3f898124c4d2adf90b4ec9ca28ee520d74975002112bb8408f023a82 COO per-iter: 0.011038s | total: 11.037725s CSR per-iter: 0.014493s | total: 14.493321s ROLF per-iter: 0.000210s | total: 0.210356s DENGS per-iter: 0.033969s | total: 33.969008s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 5.15x (≈ 415% faster) Speedup (per-iter): 5.65x (≈ 465% faster) Energy Savings: 82.30% rolv vs rocSPARSE -> Speedup (per-iter): 7.43x | total: 6.78x rolv vs COO: Speedup (per-iter): 5.66x | total: 5.16x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "4b02e483523fbec343feac2b8fed3820615bb6832dda42a3da7b63ccf1ef0014", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "0afabd3c3f898124c4d2adf90b4ec9ca28ee520d74975002112bb8408f023a82", "CSR_norm_hash": "0afabd3c3f898124c4d2adf90b4ec9ca28ee520d74975002112bb8408f023a82", "COO_norm_hash": "0afabd3c3f898124c4d2adf90b4ec9ca28ee520d74975002112bb8408f023a82", "ELL_norm_hash": "0afabd3c3f898124c4d2adf90b4ec9ca28ee520d74975002112bb8408f023a82", "ROLF_norm_hash": "50f833cd0edb871dab70fd1076fd14286647900fd24dbd3131295de2f48e9f7b", "DENGS_norm_hash": "0afabd3c3f898124c4d2adf90b4ec9ca28ee520d74975002112bb8408f023a82", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "4eaaed0b681ad4346a051280bb2504683f2650e05430ced90b415a8515706ae4", "CSR_qhash_d6": "4eaaed0b681ad4346a051280bb2504683f2650e05430ced90b415a8515706ae4", "COO_qhash_d6": "4eaaed0b681ad4346a051280bb2504683f2650e05430ced90b415a8515706ae4", "ELL_qhash_d6": null, "path_selected": "COO", "pilot_dense_per_iter_s": 0.033382, "pilot_csr_per_iter_s": 0.014278, "pilot_coo_per_iter_s": 0.011023, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.187499, "rolv_iter_s": 0.00195, "dense_iter_s": 0.011017, "csr_iter_s": 0.014493, "coo_iter_s": 0.011038, "ell_iter_s": null, "rolv_total_s": 2.137202, "baseline_total_s": 11.016573, "speedup_total_vs_selected_x": 5.155, "speedup_iter_vs_selected_x": 5.65, "rolv_vs_vendor_sparse_iter_x": 7.434, "rolv_vs_vendor_sparse_total_x": 6.781, "rolv_vs_coo_iter_x": 5.661, "rolv_vs_coo_total_x": 5.165, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 14:31:25] Seed: 123456 | Pattern: random | Zeros: 90% A_hash: 252a6d9ec7eeab4eb29b6c652bffba9f11178919caadeccd14c45d00311e1433 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.036971s | CSR: 0.267385s | COO: 0.164223s | ELL: nans Selected baseline: Dense rolv load time (operator build): 0.189758 s rolv per-iter: 0.001941s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 0ea34324829b59e6d5a810b043219ca106a8eb538079e8849cd5903c80796f83 (Dense) CSR_norm_hash: 0ea34324829b59e6d5a810b043219ca106a8eb538079e8849cd5903c80796f83 COO_norm_hash: 0ea34324829b59e6d5a810b043219ca106a8eb538079e8849cd5903c80796f83 ELL_norm_hash: 0ea34324829b59e6d5a810b043219ca106a8eb538079e8849cd5903c80796f83 ROLF_norm_hash: 321943d79a8e5a4c00b38e7a32c69fc5870129ed248809d076b51f4519f056ca DENGS_norm_hash: 0ea34324829b59e6d5a810b043219ca106a8eb538079e8849cd5903c80796f83 COO per-iter: 0.164503s | total: 164.503219s CSR per-iter: 0.271926s | total: 271.926469s ROLF per-iter: 0.000231s | total: 0.231361s DENGS per-iter: 0.037809s | total: 37.808746s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 17.55x (≈ 1655% faster) Speedup (per-iter): 19.26x (≈ 1826% faster) Energy Savings: 94.81% rolv vs rocSPARSE -> Speedup (per-iter): 140.13x | total: 127.65x rolv vs COO: Speedup (per-iter): 84.77x | total: 77.22x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "252a6d9ec7eeab4eb29b6c652bffba9f11178919caadeccd14c45d00311e1433", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "0ea34324829b59e6d5a810b043219ca106a8eb538079e8849cd5903c80796f83", "CSR_norm_hash": "0ea34324829b59e6d5a810b043219ca106a8eb538079e8849cd5903c80796f83", "COO_norm_hash": "0ea34324829b59e6d5a810b043219ca106a8eb538079e8849cd5903c80796f83", "ELL_norm_hash": "0ea34324829b59e6d5a810b043219ca106a8eb538079e8849cd5903c80796f83", "ROLF_norm_hash": "321943d79a8e5a4c00b38e7a32c69fc5870129ed248809d076b51f4519f056ca", "DENGS_norm_hash": "0ea34324829b59e6d5a810b043219ca106a8eb538079e8849cd5903c80796f83", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "d07e9d8e4b761c9e713843d9e5ad22646d68738c9c9f78b846a77a566e81f77a", "CSR_qhash_d6": "d07e9d8e4b761c9e713843d9e5ad22646d68738c9c9f78b846a77a566e81f77a", "COO_qhash_d6": "d07e9d8e4b761c9e713843d9e5ad22646d68738c9c9f78b846a77a566e81f77a", "ELL_qhash_d6": null, "path_selected": "Dense", "pilot_dense_per_iter_s": 0.036971, "pilot_csr_per_iter_s": 0.267385, "pilot_coo_per_iter_s": 0.164223, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.189758, "rolv_iter_s": 0.001941, "dense_iter_s": 0.037383, "csr_iter_s": 0.271926, "coo_iter_s": 0.164503, "ell_iter_s": null, "rolv_total_s": 2.130308, "baseline_total_s": 37.382867, "speedup_total_vs_selected_x": 17.548, "speedup_iter_vs_selected_x": 19.264, "rolv_vs_vendor_sparse_iter_x": 140.129, "rolv_vs_vendor_sparse_total_x": 127.647, "rolv_vs_coo_iter_x": 84.771, "rolv_vs_coo_total_x": 77.22, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 14:40:53] Seed: 123456 | Pattern: power_law | Zeros: 90% A_hash: d1784f30a29c88bb759e8e0ce2e1d3a72ec63f8f7d0190e4b7c74bf9b0f76e26 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.037157s | CSR: 0.252115s | COO: 0.154278s | ELL: nans Selected baseline: Dense rolv load time (operator build): 0.189832 s rolv per-iter: 0.001955s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: ff7b7b9d919c85aa942dd3c65988841f5aedc3f475953f6a39377245c2e213f6 (Dense) CSR_norm_hash: ff7b7b9d919c85aa942dd3c65988841f5aedc3f475953f6a39377245c2e213f6 COO_norm_hash: ff7b7b9d919c85aa942dd3c65988841f5aedc3f475953f6a39377245c2e213f6 ELL_norm_hash: ff7b7b9d919c85aa942dd3c65988841f5aedc3f475953f6a39377245c2e213f6 ROLF_norm_hash: d25ac0c9367d63ba2b4533bd33d053eefdea95762e4d5dafaf6a2ce9ff4cb49f DENGS_norm_hash: ff7b7b9d919c85aa942dd3c65988841f5aedc3f475953f6a39377245c2e213f6 COO per-iter: 0.154339s | total: 154.338531s CSR per-iter: 0.256868s | total: 256.867781s ROLF per-iter: 0.000231s | total: 0.231724s DENGS per-iter: 0.038087s | total: 38.086641s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 17.54x (≈ 1654% faster) Speedup (per-iter): 19.24x (≈ 1824% faster) Energy Savings: 94.80% rolv vs rocSPARSE -> Speedup (per-iter): 131.42x | total: 119.79x rolv vs COO: Speedup (per-iter): 78.96x | total: 71.97x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "d1784f30a29c88bb759e8e0ce2e1d3a72ec63f8f7d0190e4b7c74bf9b0f76e26", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "ff7b7b9d919c85aa942dd3c65988841f5aedc3f475953f6a39377245c2e213f6", "CSR_norm_hash": "ff7b7b9d919c85aa942dd3c65988841f5aedc3f475953f6a39377245c2e213f6", "COO_norm_hash": "ff7b7b9d919c85aa942dd3c65988841f5aedc3f475953f6a39377245c2e213f6", "ELL_norm_hash": "ff7b7b9d919c85aa942dd3c65988841f5aedc3f475953f6a39377245c2e213f6", "ROLF_norm_hash": "d25ac0c9367d63ba2b4533bd33d053eefdea95762e4d5dafaf6a2ce9ff4cb49f", "DENGS_norm_hash": "ff7b7b9d919c85aa942dd3c65988841f5aedc3f475953f6a39377245c2e213f6", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "11f8da7a2080d0d605a1ebff2f877f43fdf7d15739e0c108fe9752e7a0e9c93a", "CSR_qhash_d6": "11f8da7a2080d0d605a1ebff2f877f43fdf7d15739e0c108fe9752e7a0e9c93a", "COO_qhash_d6": "11f8da7a2080d0d605a1ebff2f877f43fdf7d15739e0c108fe9752e7a0e9c93a", "ELL_qhash_d6": null, "path_selected": "Dense", "pilot_dense_per_iter_s": 0.037157, "pilot_csr_per_iter_s": 0.252115, "pilot_coo_per_iter_s": 0.154278, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.189832, "rolv_iter_s": 0.001955, "dense_iter_s": 0.037606, "csr_iter_s": 0.256868, "coo_iter_s": 0.154339, "ell_iter_s": null, "rolv_total_s": 2.144396, "baseline_total_s": 37.60575, "speedup_total_vs_selected_x": 17.537, "speedup_iter_vs_selected_x": 19.24, "rolv_vs_vendor_sparse_iter_x": 131.42, "rolv_vs_vendor_sparse_total_x": 119.786, "rolv_vs_coo_iter_x": 78.963, "rolv_vs_coo_total_x": 71.973, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 14:49:56] Seed: 123456 | Pattern: banded | Zeros: 90% A_hash: d70a4343e5b268957eb68d7e3674a43f240457ccfda08b4a2d80bc40ab643157 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.033844s | CSR: 0.013120s | COO: 0.009537s | ELL: nans Selected baseline: COO rolv load time (operator build): 0.187871 s rolv per-iter: 0.001948s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 6bab4e46cb1871e51f5424a844af2f1390bcb5c5fb0b3a7c6f421bd1bc78bc94 (COO) CSR_norm_hash: 6bab4e46cb1871e51f5424a844af2f1390bcb5c5fb0b3a7c6f421bd1bc78bc94 COO_norm_hash: 6bab4e46cb1871e51f5424a844af2f1390bcb5c5fb0b3a7c6f421bd1bc78bc94 ELL_norm_hash: 6bab4e46cb1871e51f5424a844af2f1390bcb5c5fb0b3a7c6f421bd1bc78bc94 ROLF_norm_hash: 7b612cfd403099d0e842a486b4d51c55fba1f9c7ff554a4a1ed5affedc0f75a1 DENGS_norm_hash: 6bab4e46cb1871e51f5424a844af2f1390bcb5c5fb0b3a7c6f421bd1bc78bc94 COO per-iter: 0.009605s | total: 9.604517s CSR per-iter: 0.013286s | total: 13.285565s ROLF per-iter: 0.000210s | total: 0.210630s DENGS per-iter: 0.034970s | total: 34.969887s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 4.49x (≈ 349% faster) Speedup (per-iter): 4.92x (≈ 392% faster) Energy Savings: 79.67% rolv vs rocSPARSE -> Speedup (per-iter): 6.82x | total: 6.22x rolv vs COO: Speedup (per-iter): 4.93x | total: 4.50x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "d70a4343e5b268957eb68d7e3674a43f240457ccfda08b4a2d80bc40ab643157", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "6bab4e46cb1871e51f5424a844af2f1390bcb5c5fb0b3a7c6f421bd1bc78bc94", "CSR_norm_hash": "6bab4e46cb1871e51f5424a844af2f1390bcb5c5fb0b3a7c6f421bd1bc78bc94", "COO_norm_hash": "6bab4e46cb1871e51f5424a844af2f1390bcb5c5fb0b3a7c6f421bd1bc78bc94", "ELL_norm_hash": "6bab4e46cb1871e51f5424a844af2f1390bcb5c5fb0b3a7c6f421bd1bc78bc94", "ROLF_norm_hash": "7b612cfd403099d0e842a486b4d51c55fba1f9c7ff554a4a1ed5affedc0f75a1", "DENGS_norm_hash": "6bab4e46cb1871e51f5424a844af2f1390bcb5c5fb0b3a7c6f421bd1bc78bc94", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "c850461ae5bfa1c3183f8c5ad70eadff35c42a0cf45abfac99892c98541b884e", "CSR_qhash_d6": "c850461ae5bfa1c3183f8c5ad70eadff35c42a0cf45abfac99892c98541b884e", "COO_qhash_d6": "c850461ae5bfa1c3183f8c5ad70eadff35c42a0cf45abfac99892c98541b884e", "ELL_qhash_d6": null, "path_selected": "COO", "pilot_dense_per_iter_s": 0.033844, "pilot_csr_per_iter_s": 0.01312, "pilot_coo_per_iter_s": 0.009537, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.187871, "rolv_iter_s": 0.001948, "dense_iter_s": 0.009579, "csr_iter_s": 0.013286, "coo_iter_s": 0.009605, "ell_iter_s": null, "rolv_total_s": 2.135535, "baseline_total_s": 9.578621, "speedup_total_vs_selected_x": 4.485, "speedup_iter_vs_selected_x": 4.918, "rolv_vs_vendor_sparse_iter_x": 6.821, "rolv_vs_vendor_sparse_total_x": 6.221, "rolv_vs_coo_iter_x": 4.931, "rolv_vs_coo_total_x": 4.497, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 14:51:44] Seed: 123456 | Pattern: block_diagonal | Zeros: 90% A_hash: ef3c072370841e3130690e4f6793ea35e3e0c704fce673efdbae340a03091d07 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.032874s | CSR: 0.009518s | COO: 0.006847s | ELL: nans Selected baseline: COO rolv load time (operator build): 0.187597 s rolv per-iter: 0.001949s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 3954f5e832c294f5f63bb74e0179a360d788ef7079faeb84f69713613cc4ea79 (COO) CSR_norm_hash: 3954f5e832c294f5f63bb74e0179a360d788ef7079faeb84f69713613cc4ea79 COO_norm_hash: 3954f5e832c294f5f63bb74e0179a360d788ef7079faeb84f69713613cc4ea79 ELL_norm_hash: 3954f5e832c294f5f63bb74e0179a360d788ef7079faeb84f69713613cc4ea79 ROLF_norm_hash: 043a6ad806a02a4dbfd07edb33a20bdebb586c17d7abfcd015ef28379fdadb27 DENGS_norm_hash: 3954f5e832c294f5f63bb74e0179a360d788ef7079faeb84f69713613cc4ea79 COO per-iter: 0.006879s | total: 6.878592s CSR per-iter: 0.009626s | total: 9.625719s ROLF per-iter: 0.000206s | total: 0.206346s DENGS per-iter: 0.033961s | total: 33.960617s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 3.21x (≈ 221% faster) Speedup (per-iter): 3.52x (≈ 252% faster) Energy Savings: 71.59% rolv vs rocSPARSE -> Speedup (per-iter): 4.94x | total: 4.51x rolv vs COO: Speedup (per-iter): 3.53x | total: 3.22x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "ef3c072370841e3130690e4f6793ea35e3e0c704fce673efdbae340a03091d07", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "3954f5e832c294f5f63bb74e0179a360d788ef7079faeb84f69713613cc4ea79", "CSR_norm_hash": "3954f5e832c294f5f63bb74e0179a360d788ef7079faeb84f69713613cc4ea79", "COO_norm_hash": "3954f5e832c294f5f63bb74e0179a360d788ef7079faeb84f69713613cc4ea79", "ELL_norm_hash": "3954f5e832c294f5f63bb74e0179a360d788ef7079faeb84f69713613cc4ea79", "ROLF_norm_hash": "043a6ad806a02a4dbfd07edb33a20bdebb586c17d7abfcd015ef28379fdadb27", "DENGS_norm_hash": "3954f5e832c294f5f63bb74e0179a360d788ef7079faeb84f69713613cc4ea79", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "cffe32a36e15addac2c5b2fef97a992d6d9c8731c108477cdc00f463498f0e00", "CSR_qhash_d6": "cffe32a36e15addac2c5b2fef97a992d6d9c8731c108477cdc00f463498f0e00", "COO_qhash_d6": "cffe32a36e15addac2c5b2fef97a992d6d9c8731c108477cdc00f463498f0e00", "ELL_qhash_d6": null, "path_selected": "COO", "pilot_dense_per_iter_s": 0.032874, "pilot_csr_per_iter_s": 0.009518, "pilot_coo_per_iter_s": 0.006847, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.187597, "rolv_iter_s": 0.001949, "dense_iter_s": 0.006859, "csr_iter_s": 0.009626, "coo_iter_s": 0.006879, "ell_iter_s": null, "rolv_total_s": 2.136246, "baseline_total_s": 6.85862, "speedup_total_vs_selected_x": 3.211, "speedup_iter_vs_selected_x": 3.52, "rolv_vs_vendor_sparse_iter_x": 4.94, "rolv_vs_vendor_sparse_total_x": 4.506, "rolv_vs_coo_iter_x": 3.53, "rolv_vs_coo_total_x": 3.22, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 14:53:22] Seed: 123456 | Pattern: random | Zeros: 95% A_hash: c926d3fc034ec0adbed3fa6ecc74c1e0c4191486cd48fd095fa3c179c6ef96db | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.035870s | CSR: 0.167568s | COO: 0.085325s | ELL: nans Selected baseline: Dense rolv load time (operator build): 0.188521 s rolv per-iter: 0.001951s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: f5570aa0c2e30ca57e4bfba6db1ba2ed338b13cf6c3a3861cd9efa7d20b15d71 (Dense) CSR_norm_hash: f5570aa0c2e30ca57e4bfba6db1ba2ed338b13cf6c3a3861cd9efa7d20b15d71 COO_norm_hash: f5570aa0c2e30ca57e4bfba6db1ba2ed338b13cf6c3a3861cd9efa7d20b15d71 ELL_norm_hash: f5570aa0c2e30ca57e4bfba6db1ba2ed338b13cf6c3a3861cd9efa7d20b15d71 ROLF_norm_hash: 438099e42e0c75acf4c952287040709c2d6b5b9b38bb0103770643c2625437f3 DENGS_norm_hash: f5570aa0c2e30ca57e4bfba6db1ba2ed338b13cf6c3a3861cd9efa7d20b15d71 COO per-iter: 0.085350s | total: 85.350430s CSR per-iter: 0.172228s | total: 172.227984s ROLF per-iter: 0.000228s | total: 0.228468s DENGS per-iter: 0.037135s | total: 37.134816s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 17.18x (≈ 1618% faster) Speedup (per-iter): 18.84x (≈ 1784% faster) Energy Savings: 94.69% rolv vs rocSPARSE -> Speedup (per-iter): 88.28x | total: 80.50x rolv vs COO: Speedup (per-iter): 43.75x | total: 39.89x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "c926d3fc034ec0adbed3fa6ecc74c1e0c4191486cd48fd095fa3c179c6ef96db", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "f5570aa0c2e30ca57e4bfba6db1ba2ed338b13cf6c3a3861cd9efa7d20b15d71", "CSR_norm_hash": "f5570aa0c2e30ca57e4bfba6db1ba2ed338b13cf6c3a3861cd9efa7d20b15d71", "COO_norm_hash": "f5570aa0c2e30ca57e4bfba6db1ba2ed338b13cf6c3a3861cd9efa7d20b15d71", "ELL_norm_hash": "f5570aa0c2e30ca57e4bfba6db1ba2ed338b13cf6c3a3861cd9efa7d20b15d71", "ROLF_norm_hash": "438099e42e0c75acf4c952287040709c2d6b5b9b38bb0103770643c2625437f3", "DENGS_norm_hash": "f5570aa0c2e30ca57e4bfba6db1ba2ed338b13cf6c3a3861cd9efa7d20b15d71", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "c341e71c1d6aaaff215d32bf3ce2823faac0512f379a99a19bf0274553f15740", "CSR_qhash_d6": "c341e71c1d6aaaff215d32bf3ce2823faac0512f379a99a19bf0274553f15740", "COO_qhash_d6": "c341e71c1d6aaaff215d32bf3ce2823faac0512f379a99a19bf0274553f15740", "ELL_qhash_d6": null, "path_selected": "Dense", "pilot_dense_per_iter_s": 0.03587, "pilot_csr_per_iter_s": 0.167568, "pilot_coo_per_iter_s": 0.085325, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.188521, "rolv_iter_s": 0.001951, "dense_iter_s": 0.036757, "csr_iter_s": 0.172228, "coo_iter_s": 0.08535, "ell_iter_s": null, "rolv_total_s": 2.139553, "baseline_total_s": 36.756574, "speedup_total_vs_selected_x": 17.18, "speedup_iter_vs_selected_x": 18.84, "rolv_vs_vendor_sparse_iter_x": 88.275, "rolv_vs_vendor_sparse_total_x": 80.497, "rolv_vs_coo_iter_x": 43.746, "rolv_vs_coo_total_x": 39.892, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 14:59:42] Seed: 123456 | Pattern: power_law | Zeros: 95% A_hash: 6417a2a60f09c4389956722addb9e641d9618bcfe0eae0e987dfe602fdefb429 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.036163s | CSR: 0.163311s | COO: 0.080119s | ELL: nans Selected baseline: Dense rolv load time (operator build): 0.189327 s rolv per-iter: 0.001955s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: e1a90360ba8f3a6ce55dfff4d1515996f6b04b522f024ed9ab8124c98d9cb2cf (Dense) CSR_norm_hash: e1a90360ba8f3a6ce55dfff4d1515996f6b04b522f024ed9ab8124c98d9cb2cf COO_norm_hash: e1a90360ba8f3a6ce55dfff4d1515996f6b04b522f024ed9ab8124c98d9cb2cf ELL_norm_hash: e1a90360ba8f3a6ce55dfff4d1515996f6b04b522f024ed9ab8124c98d9cb2cf ROLF_norm_hash: d02f72a24d8562f072ac59b90a05de2bb21f3e093177220577b1c675fdd1e16e DENGS_norm_hash: e1a90360ba8f3a6ce55dfff4d1515996f6b04b522f024ed9ab8124c98d9cb2cf COO per-iter: 0.080128s | total: 80.128023s CSR per-iter: 0.167538s | total: 167.537594s ROLF per-iter: 0.000228s | total: 0.228459s DENGS per-iter: 0.037425s | total: 37.425320s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 17.27x (≈ 1627% faster) Speedup (per-iter): 18.94x (≈ 1794% faster) Energy Savings: 94.72% rolv vs rocSPARSE -> Speedup (per-iter): 85.71x | total: 78.14x rolv vs COO: Speedup (per-iter): 40.99x | total: 37.37x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "6417a2a60f09c4389956722addb9e641d9618bcfe0eae0e987dfe602fdefb429", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "e1a90360ba8f3a6ce55dfff4d1515996f6b04b522f024ed9ab8124c98d9cb2cf", "CSR_norm_hash": "e1a90360ba8f3a6ce55dfff4d1515996f6b04b522f024ed9ab8124c98d9cb2cf", "COO_norm_hash": "e1a90360ba8f3a6ce55dfff4d1515996f6b04b522f024ed9ab8124c98d9cb2cf", "ELL_norm_hash": "e1a90360ba8f3a6ce55dfff4d1515996f6b04b522f024ed9ab8124c98d9cb2cf", "ROLF_norm_hash": "d02f72a24d8562f072ac59b90a05de2bb21f3e093177220577b1c675fdd1e16e", "DENGS_norm_hash": "e1a90360ba8f3a6ce55dfff4d1515996f6b04b522f024ed9ab8124c98d9cb2cf", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "100de79f25f9eba2174c313cdd119f1094c254144c65bddf7539fe46bd2b2afb", "CSR_qhash_d6": "100de79f25f9eba2174c313cdd119f1094c254144c65bddf7539fe46bd2b2afb", "COO_qhash_d6": "100de79f25f9eba2174c313cdd119f1094c254144c65bddf7539fe46bd2b2afb", "ELL_qhash_d6": null, "path_selected": "Dense", "pilot_dense_per_iter_s": 0.036163, "pilot_csr_per_iter_s": 0.163311, "pilot_coo_per_iter_s": 0.080119, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.189327, "rolv_iter_s": 0.001955, "dense_iter_s": 0.037021, "csr_iter_s": 0.167538, "coo_iter_s": 0.080128, "ell_iter_s": null, "rolv_total_s": 2.144039, "baseline_total_s": 37.021, "speedup_total_vs_selected_x": 17.267, "speedup_iter_vs_selected_x": 18.939, "rolv_vs_vendor_sparse_iter_x": 85.71, "rolv_vs_vendor_sparse_total_x": 78.141, "rolv_vs_coo_iter_x": 40.992, "rolv_vs_coo_total_x": 37.372, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 15:05:54] Seed: 123456 | Pattern: banded | Zeros: 95% A_hash: f9841b629a96caca12ae5093b69047a66277d824418f1f09df0d2ec6bec61381 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.033834s | CSR: 0.008960s | COO: 0.006184s | ELL: nans Selected baseline: COO rolv load time (operator build): 0.188498 s rolv per-iter: 0.001949s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: a00aed09a9ca1a80f3acaafa376dc0e568949aeb81fcc547c28bc98683803a08 (COO) CSR_norm_hash: a00aed09a9ca1a80f3acaafa376dc0e568949aeb81fcc547c28bc98683803a08 COO_norm_hash: a00aed09a9ca1a80f3acaafa376dc0e568949aeb81fcc547c28bc98683803a08 ELL_norm_hash: a00aed09a9ca1a80f3acaafa376dc0e568949aeb81fcc547c28bc98683803a08 ROLF_norm_hash: da09351c75b5339ca035bd5d494f359e4c85eb4b75b0e3f8b5462fc1a53761ba DENGS_norm_hash: a00aed09a9ca1a80f3acaafa376dc0e568949aeb81fcc547c28bc98683803a08 COO per-iter: 0.006202s | total: 6.201649s CSR per-iter: 0.009038s | total: 9.037995s ROLF per-iter: 0.000208s | total: 0.208480s DENGS per-iter: 0.034887s | total: 34.886969s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 2.90x (≈ 190% faster) Speedup (per-iter): 3.18x (≈ 218% faster) Energy Savings: 68.51% rolv vs rocSPARSE -> Speedup (per-iter): 4.64x | total: 4.23x rolv vs COO: Speedup (per-iter): 3.18x | total: 2.90x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "f9841b629a96caca12ae5093b69047a66277d824418f1f09df0d2ec6bec61381", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "a00aed09a9ca1a80f3acaafa376dc0e568949aeb81fcc547c28bc98683803a08", "CSR_norm_hash": "a00aed09a9ca1a80f3acaafa376dc0e568949aeb81fcc547c28bc98683803a08", "COO_norm_hash": "a00aed09a9ca1a80f3acaafa376dc0e568949aeb81fcc547c28bc98683803a08", "ELL_norm_hash": "a00aed09a9ca1a80f3acaafa376dc0e568949aeb81fcc547c28bc98683803a08", "ROLF_norm_hash": "da09351c75b5339ca035bd5d494f359e4c85eb4b75b0e3f8b5462fc1a53761ba", "DENGS_norm_hash": "a00aed09a9ca1a80f3acaafa376dc0e568949aeb81fcc547c28bc98683803a08", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "9c378462296ec1cacef0a364ddaeebca041f1cfda7d5705308d0e32cbdf1f369", "CSR_qhash_d6": "9c378462296ec1cacef0a364ddaeebca041f1cfda7d5705308d0e32cbdf1f369", "COO_qhash_d6": "9c378462296ec1cacef0a364ddaeebca041f1cfda7d5705308d0e32cbdf1f369", "ELL_qhash_d6": null, "path_selected": "COO", "pilot_dense_per_iter_s": 0.033834, "pilot_csr_per_iter_s": 0.00896, "pilot_coo_per_iter_s": 0.006184, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.188498, "rolv_iter_s": 0.001949, "dense_iter_s": 0.006189, "csr_iter_s": 0.009038, "coo_iter_s": 0.006202, "ell_iter_s": null, "rolv_total_s": 2.137228, "baseline_total_s": 6.18901, "speedup_total_vs_selected_x": 2.896, "speedup_iter_vs_selected_x": 3.176, "rolv_vs_vendor_sparse_iter_x": 4.638, "rolv_vs_vendor_sparse_total_x": 4.229, "rolv_vs_coo_iter_x": 3.182, "rolv_vs_coo_total_x": 2.902, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 15:07:30] Seed: 123456 | Pattern: block_diagonal | Zeros: 95% A_hash: 743ed1c8dc03a5de5d0b131edc508c8c9e30dc02e5406aeb9cb6e8c0ce493874 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.032914s | CSR: 0.006953s | COO: 0.004728s | ELL: nans Selected baseline: COO rolv load time (operator build): 0.187722 s rolv per-iter: 0.001946s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: cb31498379c907686529a18f196926f32e7b1052704f4ed5aaebf0f0e0ed14b1 (COO) CSR_norm_hash: cb31498379c907686529a18f196926f32e7b1052704f4ed5aaebf0f0e0ed14b1 COO_norm_hash: cb31498379c907686529a18f196926f32e7b1052704f4ed5aaebf0f0e0ed14b1 ELL_norm_hash: cb31498379c907686529a18f196926f32e7b1052704f4ed5aaebf0f0e0ed14b1 ROLF_norm_hash: a23f52d144ae759bed22286b940fe6340572d8562f2e0cdd90f665c8a5c99fa6 DENGS_norm_hash: cb31498379c907686529a18f196926f32e7b1052704f4ed5aaebf0f0e0ed14b1 COO per-iter: 0.004744s | total: 4.743615s CSR per-iter: 0.007010s | total: 7.009609s ROLF per-iter: 0.000204s | total: 0.205104s DENGS per-iter: 0.033868s | total: 33.867609s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 2.22x (≈ 122% faster) Speedup (per-iter): 2.43x (≈ 143% faster) Energy Savings: 58.90% rolv vs rocSPARSE -> Speedup (per-iter): 3.60x | total: 3.29x rolv vs COO: Speedup (per-iter): 2.44x | total: 2.22x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "743ed1c8dc03a5de5d0b131edc508c8c9e30dc02e5406aeb9cb6e8c0ce493874", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "cb31498379c907686529a18f196926f32e7b1052704f4ed5aaebf0f0e0ed14b1", "CSR_norm_hash": "cb31498379c907686529a18f196926f32e7b1052704f4ed5aaebf0f0e0ed14b1", "COO_norm_hash": "cb31498379c907686529a18f196926f32e7b1052704f4ed5aaebf0f0e0ed14b1", "ELL_norm_hash": "cb31498379c907686529a18f196926f32e7b1052704f4ed5aaebf0f0e0ed14b1", "ROLF_norm_hash": "a23f52d144ae759bed22286b940fe6340572d8562f2e0cdd90f665c8a5c99fa6", "DENGS_norm_hash": "cb31498379c907686529a18f196926f32e7b1052704f4ed5aaebf0f0e0ed14b1", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "efbe0ca36ca8f3c6721275268db28beadf0ad8a4b2a7aa76452f596348100e65", "CSR_qhash_d6": "efbe0ca36ca8f3c6721275268db28beadf0ad8a4b2a7aa76452f596348100e65", "COO_qhash_d6": "efbe0ca36ca8f3c6721275268db28beadf0ad8a4b2a7aa76452f596348100e65", "ELL_qhash_d6": null, "path_selected": "COO", "pilot_dense_per_iter_s": 0.032914, "pilot_csr_per_iter_s": 0.006953, "pilot_coo_per_iter_s": 0.004728, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.187722, "rolv_iter_s": 0.001946, "dense_iter_s": 0.004735, "csr_iter_s": 0.00701, "coo_iter_s": 0.004744, "ell_iter_s": null, "rolv_total_s": 2.133522, "baseline_total_s": 4.734759, "speedup_total_vs_selected_x": 2.219, "speedup_iter_vs_selected_x": 2.433, "rolv_vs_vendor_sparse_iter_x": 3.602, "rolv_vs_vendor_sparse_total_x": 3.285, "rolv_vs_coo_iter_x": 2.438, "rolv_vs_coo_total_x": 2.223, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 15:09:00] Seed: 123456 | Pattern: random | Zeros: 99% A_hash: 9fde8b5d279f5d4d8297c2b0a4f006d0bf2475b62e6dabc7da09b547c8edbc8a | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.035463s | CSR: 0.066250s | COO: 0.020270s | ELL: nans Selected baseline: COO rolv load time (operator build): 0.187375 s rolv per-iter: 0.001948s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: cc8c3cc839a9d0c5930d0938d01d9bd2a7f5d66f47fcce4e53dec39e0dc7faa9 (COO) CSR_norm_hash: cc8c3cc839a9d0c5930d0938d01d9bd2a7f5d66f47fcce4e53dec39e0dc7faa9 COO_norm_hash: cc8c3cc839a9d0c5930d0938d01d9bd2a7f5d66f47fcce4e53dec39e0dc7faa9 ELL_norm_hash: cc8c3cc839a9d0c5930d0938d01d9bd2a7f5d66f47fcce4e53dec39e0dc7faa9 ROLF_norm_hash: cd16ca2eb816f4e3a4324002312ca0268a38b732796aab7888293ef7f24633b7 DENGS_norm_hash: cc8c3cc839a9d0c5930d0938d01d9bd2a7f5d66f47fcce4e53dec39e0dc7faa9 COO per-iter: 0.020439s | total: 20.438904s CSR per-iter: 0.068173s | total: 68.172727s ROLF per-iter: 0.000218s | total: 0.218390s DENGS per-iter: 0.036556s | total: 36.556469s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 9.54x (≈ 854% faster) Speedup (per-iter): 10.46x (≈ 946% faster) Energy Savings: 90.44% rolv vs rocSPARSE -> Speedup (per-iter): 35.00x | total: 31.93x rolv vs COO: Speedup (per-iter): 10.49x | total: 9.57x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "9fde8b5d279f5d4d8297c2b0a4f006d0bf2475b62e6dabc7da09b547c8edbc8a", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "cc8c3cc839a9d0c5930d0938d01d9bd2a7f5d66f47fcce4e53dec39e0dc7faa9", "CSR_norm_hash": "cc8c3cc839a9d0c5930d0938d01d9bd2a7f5d66f47fcce4e53dec39e0dc7faa9", "COO_norm_hash": "cc8c3cc839a9d0c5930d0938d01d9bd2a7f5d66f47fcce4e53dec39e0dc7faa9", "ELL_norm_hash": "cc8c3cc839a9d0c5930d0938d01d9bd2a7f5d66f47fcce4e53dec39e0dc7faa9", "ROLF_norm_hash": "cd16ca2eb816f4e3a4324002312ca0268a38b732796aab7888293ef7f24633b7", "DENGS_norm_hash": "cc8c3cc839a9d0c5930d0938d01d9bd2a7f5d66f47fcce4e53dec39e0dc7faa9", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "18a55821ec1e53d19620c8b3fdf1bd7dc27837e4d3f0bc4b8bb33ab2e24117eb", "CSR_qhash_d6": "18a55821ec1e53d19620c8b3fdf1bd7dc27837e4d3f0bc4b8bb33ab2e24117eb", "COO_qhash_d6": "18a55821ec1e53d19620c8b3fdf1bd7dc27837e4d3f0bc4b8bb33ab2e24117eb", "ELL_qhash_d6": null, "path_selected": "COO", "pilot_dense_per_iter_s": 0.035463, "pilot_csr_per_iter_s": 0.06625, "pilot_coo_per_iter_s": 0.02027, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.187375, "rolv_iter_s": 0.001948, "dense_iter_s": 0.020374, "csr_iter_s": 0.068173, "coo_iter_s": 0.020439, "ell_iter_s": null, "rolv_total_s": 2.135294, "baseline_total_s": 20.373879, "speedup_total_vs_selected_x": 9.541, "speedup_iter_vs_selected_x": 10.459, "rolv_vs_vendor_sparse_iter_x": 34.998, "rolv_vs_vendor_sparse_total_x": 31.927, "rolv_vs_coo_iter_x": 10.493, "rolv_vs_coo_total_x": 9.572, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 15:12:09] Seed: 123456 | Pattern: power_law | Zeros: 99% A_hash: 3884cba828aa7a1488fc132da5edbcb037e4d5cda60d2548cbb05d1438117888 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.035735s | CSR: 0.063122s | COO: 0.019269s | ELL: nans Selected baseline: COO rolv load time (operator build): 0.187962 s rolv per-iter: 0.001952s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: c3e9496be8a189fc75d306823ff6359f8fa3b5aa5557bbc72bae19d4a7ed7d5d (COO) CSR_norm_hash: c3e9496be8a189fc75d306823ff6359f8fa3b5aa5557bbc72bae19d4a7ed7d5d COO_norm_hash: c3e9496be8a189fc75d306823ff6359f8fa3b5aa5557bbc72bae19d4a7ed7d5d ELL_norm_hash: c3e9496be8a189fc75d306823ff6359f8fa3b5aa5557bbc72bae19d4a7ed7d5d ROLF_norm_hash: 4727304f2130748ccc072bc431f0db2c0cbe41049c05e3323613e00e8f930359 DENGS_norm_hash: c3e9496be8a189fc75d306823ff6359f8fa3b5aa5557bbc72bae19d4a7ed7d5d COO per-iter: 0.019416s | total: 19.415646s CSR per-iter: 0.064838s | total: 64.838012s ROLF per-iter: 0.000214s | total: 0.214380s DENGS per-iter: 0.036803s | total: 36.802875s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 9.03x (≈ 803% faster) Speedup (per-iter): 9.89x (≈ 889% faster) Energy Savings: 89.89% rolv vs rocSPARSE -> Speedup (per-iter): 33.22x | total: 30.30x rolv vs COO: Speedup (per-iter): 9.95x | total: 9.07x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "3884cba828aa7a1488fc132da5edbcb037e4d5cda60d2548cbb05d1438117888", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "c3e9496be8a189fc75d306823ff6359f8fa3b5aa5557bbc72bae19d4a7ed7d5d", "CSR_norm_hash": "c3e9496be8a189fc75d306823ff6359f8fa3b5aa5557bbc72bae19d4a7ed7d5d", "COO_norm_hash": "c3e9496be8a189fc75d306823ff6359f8fa3b5aa5557bbc72bae19d4a7ed7d5d", "ELL_norm_hash": "c3e9496be8a189fc75d306823ff6359f8fa3b5aa5557bbc72bae19d4a7ed7d5d", "ROLF_norm_hash": "4727304f2130748ccc072bc431f0db2c0cbe41049c05e3323613e00e8f930359", "DENGS_norm_hash": "c3e9496be8a189fc75d306823ff6359f8fa3b5aa5557bbc72bae19d4a7ed7d5d", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "570879d26c1763504bd05013868ba03d6e5c343019d405fbb6664799c3446a0a", "CSR_qhash_d6": "570879d26c1763504bd05013868ba03d6e5c343019d405fbb6664799c3446a0a", "COO_qhash_d6": "570879d26c1763504bd05013868ba03d6e5c343019d405fbb6664799c3446a0a", "ELL_qhash_d6": null, "path_selected": "COO", "pilot_dense_per_iter_s": 0.035735, "pilot_csr_per_iter_s": 0.063122, "pilot_coo_per_iter_s": 0.019269, "pilot_ell_per_iter_s": null, "rolv_build_s": 0.187962, "rolv_iter_s": 0.001952, "dense_iter_s": 0.019312, "csr_iter_s": 0.064838, "coo_iter_s": 0.019416, "ell_iter_s": null, "rolv_total_s": 2.13979, "baseline_total_s": 19.311652, "speedup_total_vs_selected_x": 9.025, "speedup_iter_vs_selected_x": 9.894, "rolv_vs_vendor_sparse_iter_x": 33.219, "rolv_vs_vendor_sparse_total_x": 30.301, "rolv_vs_coo_iter_x": 9.947, "rolv_vs_coo_total_x": 9.074, "rolv_vs_ell_iter_x": null, "rolv_vs_ell_total_x": null, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 15:15:12] Seed: 123456 | Pattern: banded | Zeros: 99% A_hash: 1b643fe5ac4811868b9b5bfee7d7ed4d02a612b4add98ac2d0f399d014599b67 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.034502s | CSR: 0.005416s | COO: 0.003279s | ELL: 0.017051s Selected baseline: COO rolv load time (operator build): 0.191091 s rolv per-iter: 0.001939s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 2974cb2e57d28e2c52f6b64f51fb5c2c10ab828f6e7279b652ec7d4c7fa407ad (COO) CSR_norm_hash: 2974cb2e57d28e2c52f6b64f51fb5c2c10ab828f6e7279b652ec7d4c7fa407ad COO_norm_hash: 2974cb2e57d28e2c52f6b64f51fb5c2c10ab828f6e7279b652ec7d4c7fa407ad ELL_norm_hash: 02dfc8baaa24486fd1107fecaf28fdedc65dc9dd6c70abc3d632dd99dafe2225 ROLF_norm_hash: 832a9e85bddee020a7655ba9a16031e464caa87f1ceb92d8567b75780ccb7c6d DENGS_norm_hash: 2974cb2e57d28e2c52f6b64f51fb5c2c10ab828f6e7279b652ec7d4c7fa407ad COO per-iter: 0.003297s | total: 3.296891s CSR per-iter: 0.005442s | total: 5.441580s ELL per-iter: 0.017241s | total: 20.847531s ROLF per-iter: 0.000210s | total: 0.210070s DENGS per-iter: 0.035172s | total: 35.171750s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 1.54x (≈ 54% faster) Speedup (per-iter): 1.69x (≈ 69% faster) Energy Savings: 40.98% rolv vs rocSPARSE -> Speedup (per-iter): 2.81x | total: 2.55x rolv vs COO: Speedup (per-iter): 1.70x | total: 1.55x rolv vs ELL: Speedup (per-iter): 8.89x | total: 9.79x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "1b643fe5ac4811868b9b5bfee7d7ed4d02a612b4add98ac2d0f399d014599b67", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "2974cb2e57d28e2c52f6b64f51fb5c2c10ab828f6e7279b652ec7d4c7fa407ad", "CSR_norm_hash": "2974cb2e57d28e2c52f6b64f51fb5c2c10ab828f6e7279b652ec7d4c7fa407ad", "COO_norm_hash": "2974cb2e57d28e2c52f6b64f51fb5c2c10ab828f6e7279b652ec7d4c7fa407ad", "ELL_norm_hash": "02dfc8baaa24486fd1107fecaf28fdedc65dc9dd6c70abc3d632dd99dafe2225", "ROLF_norm_hash": "832a9e85bddee020a7655ba9a16031e464caa87f1ceb92d8567b75780ccb7c6d", "DENGS_norm_hash": "2974cb2e57d28e2c52f6b64f51fb5c2c10ab828f6e7279b652ec7d4c7fa407ad", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "36fabc0207e13e7ae08a8c6db45526167c20c66464e356ba7cd4969570c1cb52", "CSR_qhash_d6": "36fabc0207e13e7ae08a8c6db45526167c20c66464e356ba7cd4969570c1cb52", "COO_qhash_d6": "36fabc0207e13e7ae08a8c6db45526167c20c66464e356ba7cd4969570c1cb52", "ELL_qhash_d6": "ae4eccb37f4ad0156b7703f13d762a8d45312be12fa4d502e74261e2181f376d", "path_selected": "COO", "pilot_dense_per_iter_s": 0.034502, "pilot_csr_per_iter_s": 0.005416, "pilot_coo_per_iter_s": 0.003279, "pilot_ell_per_iter_s": 0.017051, "rolv_build_s": 0.191091, "rolv_iter_s": 0.001939, "dense_iter_s": 0.003286, "csr_iter_s": 0.005442, "coo_iter_s": 0.003297, "ell_iter_s": 0.017241, "rolv_total_s": 2.130281, "baseline_total_s": 3.285578, "speedup_total_vs_selected_x": 1.542, "speedup_iter_vs_selected_x": 1.694, "rolv_vs_vendor_sparse_iter_x": 2.806, "rolv_vs_vendor_sparse_total_x": 2.554, "rolv_vs_coo_iter_x": 1.7, "rolv_vs_coo_total_x": 1.548, "rolv_vs_ell_iter_x": 8.891, "rolv_vs_ell_total_x": 9.786, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-11 15:16:57] Seed: 123456 | Pattern: block_diagonal | Zeros: 99% A_hash: d78e202117fb1b5ee60605254db62aa72b0d2b72a9d6ceec1a84ad78c44df368 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.033309s | CSR: 0.004720s | COO: 0.002994s | ELL: 0.016554s Selected baseline: COO rolv load time (operator build): 0.191635 s rolv per-iter: 0.001946s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 81df416748e59ff8fb7b3e31c4a3a74db121c9fc011e70c1604e496e3b107c2b (COO) CSR_norm_hash: 81df416748e59ff8fb7b3e31c4a3a74db121c9fc011e70c1604e496e3b107c2b COO_norm_hash: 81df416748e59ff8fb7b3e31c4a3a74db121c9fc011e70c1604e496e3b107c2b ELL_norm_hash: 0c8917844825832c8bc528d96f54a16cdd751ca874f855153da4000a9c097fc0 ROLF_norm_hash: aedd4cf9d03fcc98a2d4b33b9c56bc89ad6ca996f94ce0f83609d26a240cfcd3 DENGS_norm_hash: 81df416748e59ff8fb7b3e31c4a3a74db121c9fc011e70c1604e496e3b107c2b COO per-iter: 0.003017s | total: 3.017451s CSR per-iter: 0.004744s | total: 4.744417s ELL per-iter: 0.016723s | total: 19.996827s ROLF per-iter: 0.000204s | total: 0.204547s DENGS per-iter: 0.034060s | total: 34.060012s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 1.40x (≈ 40% faster) Speedup (per-iter): 1.54x (≈ 54% faster) Energy Savings: 35.20% rolv vs rocSPARSE -> Speedup (per-iter): 2.44x | total: 2.22x rolv vs COO: Speedup (per-iter): 1.55x | total: 1.41x rolv vs ELL: Speedup (per-iter): 8.60x | total: 9.36x {"platform": "ROCm", "device": "AMD Instinct MI300X", "adapted_batch": false, "effective_batch": 5000, "dense_label": "rocBLAS", "sparse_label": "rocSPARSE", "input_hash_A": "d78e202117fb1b5ee60605254db62aa72b0d2b72a9d6ceec1a84ad78c44df368", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "81df416748e59ff8fb7b3e31c4a3a74db121c9fc011e70c1604e496e3b107c2b", "CSR_norm_hash": "81df416748e59ff8fb7b3e31c4a3a74db121c9fc011e70c1604e496e3b107c2b", "COO_norm_hash": "81df416748e59ff8fb7b3e31c4a3a74db121c9fc011e70c1604e496e3b107c2b", "ELL_norm_hash": "0c8917844825832c8bc528d96f54a16cdd751ca874f855153da4000a9c097fc0", "ROLF_norm_hash": "aedd4cf9d03fcc98a2d4b33b9c56bc89ad6ca996f94ce0f83609d26a240cfcd3", "DENGS_norm_hash": "81df416748e59ff8fb7b3e31c4a3a74db121c9fc011e70c1604e496e3b107c2b", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "72ae2e8b1b11989b240bd407be5eae8d1ffdbf75beeacce50c056cdb544c412f", "CSR_qhash_d6": "72ae2e8b1b11989b240bd407be5eae8d1ffdbf75beeacce50c056cdb544c412f", "COO_qhash_d6": "72ae2e8b1b11989b240bd407be5eae8d1ffdbf75beeacce50c056cdb544c412f", "ELL_qhash_d6": "75bb10fb7f6c1797427495d49fffc80547fb8d6e1dec04103c1da7429b25a8c6", "path_selected": "COO", "pilot_dense_per_iter_s": 0.033309, "pilot_csr_per_iter_s": 0.00472, "pilot_coo_per_iter_s": 0.002994, "pilot_ell_per_iter_s": 0.016554, "rolv_build_s": 0.191635, "rolv_iter_s": 0.001946, "dense_iter_s": 0.003003, "csr_iter_s": 0.004744, "coo_iter_s": 0.003017, "ell_iter_s": 0.016723, "rolv_total_s": 2.13729, "baseline_total_s": 3.002549, "speedup_total_vs_selected_x": 1.405, "speedup_iter_vs_selected_x": 1.543, "rolv_vs_vendor_sparse_iter_x": 2.438, "rolv_vs_vendor_sparse_total_x": 2.22, "rolv_vs_coo_iter_x": 1.551, "rolv_vs_coo_total_x": 1.412, "rolv_vs_ell_iter_x": 8.595, "rolv_vs_ell_total_x": 9.356, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

=== FOOTER REPORT (ROCm) ===

Aggregate speedup (total vs selected): 12.05x (≈ 1105% faster)
Aggregate speedup (per-iter vs selected): 13.29x (≈ 1229% faster)
Aggregate energy savings (proxy vs selected): 86.2%
Verification: TF32 off, deterministic algorithms, CSR canonicalization, CPU-fp64 normalization and SHA-256 hashing. {"platform": "ROCm", "device": "AMD Instinct MI300X", "aggregate_speedup_total_vs_selected_x": 12.055, "aggregate_speedup_iter_vs_selected_x": 13.294, "aggregate_energy_savings_pct": 86.247, "verification": "TF32 off, deterministic algorithms, CSR canonicalization, CPU-fp64 normalization, SHA-256 hashing"}

=== Timing & Energy Measurement Explanation ===

Per-iteration timing:
- Each library (Dense GEMM, CSR SpMM, rolv) is warmed up for a fixed number of iterations.
- Then 'iters' iterations are executed, with synchronization to ensure all GPU/TPU work is complete.
- The average time per iteration is reported as _iter_s.
Build/setup time:
- For rolv, operator construction (tiling, quantization, surrogate build) is timed separately as rolv_build_s.
- Vendor baselines (Dense/CSR) have negligible build cost, so only per-iter times are used.
Total time:
- For each library, total runtime = build/setup time + (per-iter time × number of iterations).
- Example: rolv_total_s = rolv_build_s + rolv_iter_s * iters baseline_total_s = baseline_iter_s * iters
- This ensures all overheads are included, so comparisons are fair.
Speedup calculation:
- Speedup (per-iter) = baseline_iter_s / rolv_iter_s
- Speedup (total) = baseline_total_s / rolv_total_s
- Both metrics are reported to show raw kernel efficiency and end-to-end cost.
Energy measurement:
- Proxy energy savings are computed from per-iter times: energy_savings_pct = 100 × (1 - rolv_iter_s / baseline_iter_s)
- If telemetry is enabled (NVML/ROCm SMI), instantaneous power samples (W) are integrated over time to yield Joules (trapz).
- Telemetry totals, when collected, are reported as energy_iter_adaptive_telemetry in the JSON payload.
Fairness guarantee:
- All libraries run the same matrix/vector inputs (identical seeds, identical input hashes).
- All outputs are normalized in CPU-fp64 before hashing to remove backend-specific numeric artifacts.
- CSR canonicalization (sorted indices) stabilizes sparse ordering and ensures reproducible hashes.
- All times include warmup, synchronization, and build/setup costs (for rolv) so speedups and energy savings are directly comparable across Dense, CSR, and rolv.

Imagination is the Only Limitation to Innovation

Rolv E. Heggenhougen

NVIDIA Benchmarks B200 12/08/2025

[2025-12-08 19:48:14] Seed: 123456 | Pattern: random | Zeros: 40% A_hash: e3644a901043856adaa3b878146a5978eda600732465e78134f6121ad2135eab | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 /tmp/ipykernel_454/2276430012.py:728: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /pytorch/aten/src/ATen/SparseCsrTensorImpl.cpp:53.) A_csr_raw = A_dense.to_sparse_csr() Baseline pilots per-iter -> Dense: 0.062046s | CSR: 0.514377s Selected baseline: Dense GEMM (cuBLAS) ROLV load time (operator build): 0.843197 s ROLV per-iter: 0.001008s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 11b6241f09adfebda8a84e36dfbfba9192af8d759dbd0b8612db6923472fac6c (Dense GEMM (cuBLAS)) CSR_norm_hash: 270fb3df92e7c7303d022876a20fef09f0bfd25676616d937a2c05c11561d528 ROLF_norm_hash: 96f29966a7efcb5ef2528e92078988a2063a442867509f2f469e1b63cec61896 DENGS_norm_hash: 11b6241f09adfebda8a84e36dfbfba9192af8d759dbd0b8612db6923472fac6c ROLF per-iter: 0.000329s | total: 0.337095s DENGS per-iter: 0.062037s | total: 62.037008s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 33.51x (≈ 3251% faster) Speedup (per-iter): 61.55x (≈ 6055% faster) Energy Savings: 98.38% ROLV vs cuSPARSE -> Speedup (per-iter): 510.33x | total: 277.86x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "e3644a901043856adaa3b878146a5978eda600732465e78134f6121ad2135eab", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "11b6241f09adfebda8a84e36dfbfba9192af8d759dbd0b8612db6923472fac6c", "CSR_norm_hash": "270fb3df92e7c7303d022876a20fef09f0bfd25676616d937a2c05c11561d528", "ROLF_norm_hash": "96f29966a7efcb5ef2528e92078988a2063a442867509f2f469e1b63cec61896", "DENGS_norm_hash": "11b6241f09adfebda8a84e36dfbfba9192af8d759dbd0b8612db6923472fac6c", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "d02e8632c028003b3549fb086dad732fc49aed49a06e28cfc5e2a0f32da41a36", "CSR_qhash_d6": "98b9b6f78ef43ad11838e877d54107cbefd420faf92bf941d302b43ae3f9cc95", "path_selected": "Dense", "pilot_dense_per_iter_s": 0.062046, "pilot_csr_per_iter_s": 0.514377, "rolv_build_s": 0.843197, "rolv_iter_s": 0.001008, "dense_iter_s": 0.062036, "csr_iter_s": 0.514326, "rolv_total_s": 1.851025, "baseline_total_s": 62.036344, "speedup_total_vs_selected_x": 33.515, "speedup_iter_vs_selected_x": 61.554, "rolv_vs_vendor_sparse_iter_x": 510.331, "rolv_vs_vendor_sparse_total_x": 277.86, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 19:59:38] Seed: 123456 | Pattern: power_law | Zeros: 40% A_hash: 0bc0d2cd333849b2bc5726b8182342a2b1f1692dec3ce1baa02459ebd0feca6e | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062058s | CSR: 0.475749s Selected baseline: Dense GEMM (cuBLAS) ROLV load time (operator build): 1.040151 s ROLV per-iter: 0.001008s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 3200b111483c9fae293d87a88a8e25c6fe52eb6436f1def69101414d10b57cfb (Dense GEMM (cuBLAS)) CSR_norm_hash: f3acfca79d4467ed4ca44cf55f8167239e8cc9300b3c9b660b81a3ab356ebe64 ROLF_norm_hash: 04c6dfa6fcbee09f54a8fe3b9d83bb2c2537a62e94ff981cc0de5f2215d8c38b DENGS_norm_hash: 3200b111483c9fae293d87a88a8e25c6fe52eb6436f1def69101414d10b57cfb ROLF per-iter: 0.000337s | total: 0.337602s DENGS per-iter: 0.062034s | total: 62.033844s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 30.29x (≈ 2929% faster) Speedup (per-iter): 61.56x (≈ 6056% faster) Energy Savings: 98.38% ROLV vs cuSPARSE -> Speedup (per-iter): 472.17x | total: 232.36x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "0bc0d2cd333849b2bc5726b8182342a2b1f1692dec3ce1baa02459ebd0feca6e", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "3200b111483c9fae293d87a88a8e25c6fe52eb6436f1def69101414d10b57cfb", "CSR_norm_hash": "f3acfca79d4467ed4ca44cf55f8167239e8cc9300b3c9b660b81a3ab356ebe64", "ROLF_norm_hash": "04c6dfa6fcbee09f54a8fe3b9d83bb2c2537a62e94ff981cc0de5f2215d8c38b", "DENGS_norm_hash": "3200b111483c9fae293d87a88a8e25c6fe52eb6436f1def69101414d10b57cfb", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "1bb133eca121d0422acc7c427733ee08af00c0731d925906a3a7449af91e982e", "CSR_qhash_d6": "c3d1100cbf06a9e6be9c1f7129ef8f5afd1c8a9bb708131ab4fcefe154077cb2", "path_selected": "Dense", "pilot_dense_per_iter_s": 0.062058, "pilot_csr_per_iter_s": 0.475749, "rolv_build_s": 1.040151, "rolv_iter_s": 0.001008, "dense_iter_s": 0.062041, "csr_iter_s": 0.475875, "rolv_total_s": 2.047999, "baseline_total_s": 62.041453, "speedup_total_vs_selected_x": 30.294, "speedup_iter_vs_selected_x": 61.558, "rolv_vs_vendor_sparse_iter_x": 472.169, "rolv_vs_vendor_sparse_total_x": 232.361, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 20:12:08] Seed: 123456 | Pattern: banded | Zeros: 40% A_hash: 69975c70a3346649e1fbefab534eae7887a68247af2ad0c91ced7488ab619e6c | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062061s | CSR: 0.018675s Selected baseline: Sparse CSR (cuSPARSE) ROLV load time (operator build): 0.763800 s ROLV per-iter: 0.001007s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 3b5530207d627894ec73effd7480860caa3e0b2fc748448e9b3c8e221a36c56c (Sparse CSR (cuSPARSE)) CSR_norm_hash: 3b5530207d627894ec73effd7480860caa3e0b2fc748448e9b3c8e221a36c56c ROLF_norm_hash: 3d143289649ac69457ce1a4c0c58322caf19f04aff297d32892e289db68e9353 DENGS_norm_hash: 1e73da27ba6b296895312009edb9bddcc8b91b02b3647b7a9aae70a80af2067f ROLF per-iter: 0.000329s | total: 0.329374s DENGS per-iter: 0.062044s | total: 62.043703s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 10.55x (≈ 955% faster) Speedup (per-iter): 18.54x (≈ 1754% faster) Energy Savings: 94.61% ROLV vs cuSPARSE -> Speedup (per-iter): 18.54x | total: 10.55x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "69975c70a3346649e1fbefab534eae7887a68247af2ad0c91ced7488ab619e6c", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "1e73da27ba6b296895312009edb9bddcc8b91b02b3647b7a9aae70a80af2067f", "CSR_norm_hash": "3b5530207d627894ec73effd7480860caa3e0b2fc748448e9b3c8e221a36c56c", "ROLF_norm_hash": "3d143289649ac69457ce1a4c0c58322caf19f04aff297d32892e289db68e9353", "DENGS_norm_hash": "1e73da27ba6b296895312009edb9bddcc8b91b02b3647b7a9aae70a80af2067f", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "80fe483ed7c35f0587e85f5c26d02f3e5b9572628977d7add5283e61db8ad088", "CSR_qhash_d6": "8ffdad69ce312c554f449b23b85d2351b2fdf108f114d8d0e88c032df791c40a", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.062061, "pilot_csr_per_iter_s": 0.018675, "rolv_build_s": 0.7638, "rolv_iter_s": 0.001007, "dense_iter_s": 0.018678, "csr_iter_s": 0.01868, "rolv_total_s": 1.771138, "baseline_total_s": 18.678244, "speedup_total_vs_selected_x": 10.546, "speedup_iter_vs_selected_x": 18.542, "rolv_vs_vendor_sparse_iter_x": 18.544, "rolv_vs_vendor_sparse_total_x": 10.547, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 20:14:12] Seed: 123456 | Pattern: block_diagonal | Zeros: 40% A_hash: d7a5bfe4c7f465590f90417984ef8f0754801ffe2307e0f3a276649b4868f2ad | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062063s | CSR: 0.011846s Selected baseline: Sparse CSR (cuSPARSE) ROLV load time (operator build): 0.757459 s ROLV per-iter: 0.001008s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 6cdf881646cf23f7374e162e7ecd71808779d65bf0559be36bfeef1515e23b70 (Sparse CSR (cuSPARSE)) CSR_norm_hash: 6cdf881646cf23f7374e162e7ecd71808779d65bf0559be36bfeef1515e23b70 ROLF_norm_hash: fad1e70cbfb1f7a7c76bac355291ed547070889fc04e0bac447e2cc9d20dcff1 DENGS_norm_hash: 988617603b8b5a585fdf4dad647ec0ecddad53772808704777c1f88f54f0325c ROLF per-iter: 0.000329s | total: 0.329354s DENGS per-iter: 0.062039s | total: 62.039133s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 6.71x (≈ 571% faster) Speedup (per-iter): 11.75x (≈ 1075% faster) Energy Savings: 91.49% ROLV vs cuSPARSE -> Speedup (per-iter): 11.75x | total: 6.71x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "d7a5bfe4c7f465590f90417984ef8f0754801ffe2307e0f3a276649b4868f2ad", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "988617603b8b5a585fdf4dad647ec0ecddad53772808704777c1f88f54f0325c", "CSR_norm_hash": "6cdf881646cf23f7374e162e7ecd71808779d65bf0559be36bfeef1515e23b70", "ROLF_norm_hash": "fad1e70cbfb1f7a7c76bac355291ed547070889fc04e0bac447e2cc9d20dcff1", "DENGS_norm_hash": "988617603b8b5a585fdf4dad647ec0ecddad53772808704777c1f88f54f0325c", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "9993275a88ff770aeb9aca162be175a9c3040c53c5c7f6fc2a4f319c31cfdc98", "CSR_qhash_d6": "bec512c28f1a76488b43de210873552d371d1e1285d835b2156453823ffcc9f8", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.062063, "pilot_csr_per_iter_s": 0.011846, "rolv_build_s": 0.757459, "rolv_iter_s": 0.001008, "dense_iter_s": 0.011845, "csr_iter_s": 0.011847, "rolv_total_s": 1.765793, "baseline_total_s": 11.844898, "speedup_total_vs_selected_x": 6.708, "speedup_iter_vs_selected_x": 11.747, "rolv_vs_vendor_sparse_iter_x": 11.749, "rolv_vs_vendor_sparse_total_x": 6.709, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 20:16:43] Seed: 123456 | Pattern: random | Zeros: 50% A_hash: 6e4770bed2259e6973f564d1f8d9f3edc952d13fc6befcf5a9f9094269703540 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062073s | CSR: 0.418370s Selected baseline: Dense GEMM (cuBLAS) ROLV load time (operator build): 0.882566 s ROLV per-iter: 0.001009s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 16a6f29a289e90371d2461de0e92f680a147484e05e7a322305ac8403f395404 (Dense GEMM (cuBLAS)) CSR_norm_hash: 6ae93e2063f2fcea3fdbe5cdeebba8d417404d60b2b8f5f25fc3255d24b07cb4 ROLF_norm_hash: c816fe475c90aae0faf91c6f709db164d5fa330d0a3538d29218bc7dc6e7e99f DENGS_norm_hash: 16a6f29a289e90371d2461de0e92f680a147484e05e7a322305ac8403f395404 ROLF per-iter: 0.000329s | total: 0.329300s DENGS per-iter: 0.062034s | total: 62.034195s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 32.80x (≈ 3180% faster) Speedup (per-iter): 61.48x (≈ 6048% faster) Energy Savings: 98.37% ROLV vs cuSPARSE -> Speedup (per-iter): 414.48x | total: 221.10x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "6e4770bed2259e6973f564d1f8d9f3edc952d13fc6befcf5a9f9094269703540", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "16a6f29a289e90371d2461de0e92f680a147484e05e7a322305ac8403f395404", "CSR_norm_hash": "6ae93e2063f2fcea3fdbe5cdeebba8d417404d60b2b8f5f25fc3255d24b07cb4", "ROLF_norm_hash": "c816fe475c90aae0faf91c6f709db164d5fa330d0a3538d29218bc7dc6e7e99f", "DENGS_norm_hash": "16a6f29a289e90371d2461de0e92f680a147484e05e7a322305ac8403f395404", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "6411421f1efd8ea75978adb572c41db98aed6edb716989a47e78ad96e0a71457", "CSR_qhash_d6": "8361a051cd7b2f166fe6a4cb14abf936113f3a42fe807a5d79665f7624fe2e83", "path_selected": "Dense", "pilot_dense_per_iter_s": 0.062073, "pilot_csr_per_iter_s": 0.41837, "rolv_build_s": 0.882566, "rolv_iter_s": 0.001009, "dense_iter_s": 0.062035, "csr_iter_s": 0.418226, "rolv_total_s": 1.891606, "baseline_total_s": 62.03525, "speedup_total_vs_selected_x": 32.795, "speedup_iter_vs_selected_x": 61.479, "rolv_vs_vendor_sparse_iter_x": 414.479, "rolv_vs_vendor_sparse_total_x": 221.095, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 20:27:47] Seed: 123456 | Pattern: power_law | Zeros: 50% A_hash: e868d93c6a2425c33f4461dda493d60421f514ce596dcf01814e71c6fb964106 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062221s | CSR: 0.387766s Selected baseline: Dense GEMM (cuBLAS) ROLV load time (operator build): 0.748624 s ROLV per-iter: 0.001007s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 2c7b53eec42709fbc3a8cece030b36aa65de803ee859a661ae2d92444e839f2b (Dense GEMM (cuBLAS)) CSR_norm_hash: b4787e9e1a523741e6db580f2a69205b98f09bf52468891eba17f3c0ec51e819 ROLF_norm_hash: 454ae05f90c5e045acc09af4e561c50f160fe5add28880ee4daf2660646217ba DENGS_norm_hash: 2c7b53eec42709fbc3a8cece030b36aa65de803ee859a661ae2d92444e839f2b ROLF per-iter: 0.000329s | total: 0.329389s DENGS per-iter: 0.062034s | total: 62.033574s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 35.33x (≈ 3433% faster) Speedup (per-iter): 61.58x (≈ 6058% faster) Energy Savings: 98.38% ROLV vs cuSPARSE -> Speedup (per-iter): 384.91x | total: 220.82x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "e868d93c6a2425c33f4461dda493d60421f514ce596dcf01814e71c6fb964106", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "2c7b53eec42709fbc3a8cece030b36aa65de803ee859a661ae2d92444e839f2b", "CSR_norm_hash": "b4787e9e1a523741e6db580f2a69205b98f09bf52468891eba17f3c0ec51e819", "ROLF_norm_hash": "454ae05f90c5e045acc09af4e561c50f160fe5add28880ee4daf2660646217ba", "DENGS_norm_hash": "2c7b53eec42709fbc3a8cece030b36aa65de803ee859a661ae2d92444e839f2b", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "5eedfa8fdcd56343dcae7a1bcfe78a3e0011acf344fa0a15bca967f4c5750f59", "CSR_qhash_d6": "119299c325b362570715c1694f9ffd42b2f9d39b9c620f3eeb736c3c83a997a9", "path_selected": "Dense", "pilot_dense_per_iter_s": 0.062221, "pilot_csr_per_iter_s": 0.387766, "rolv_build_s": 0.748624, "rolv_iter_s": 0.001007, "dense_iter_s": 0.062033, "csr_iter_s": 0.387767, "rolv_total_s": 1.756056, "baseline_total_s": 62.032719, "speedup_total_vs_selected_x": 35.325, "speedup_iter_vs_selected_x": 61.575, "rolv_vs_vendor_sparse_iter_x": 384.906, "rolv_vs_vendor_sparse_total_x": 220.817, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 20:38:15] Seed: 123456 | Pattern: banded | Zeros: 50% A_hash: 36930b864e45f6c7bc4c05a36ceed9e5546aba4f26c38e27ec94b84500ab052f | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062073s | CSR: 0.015665s Selected baseline: Sparse CSR (cuSPARSE) ROLV load time (operator build): 0.732619 s ROLV per-iter: 0.001007s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 7c4936017595a5d5a9b18a0030a7ae531dbabdab18f298e21264722303eabded (Sparse CSR (cuSPARSE)) CSR_norm_hash: 7c4936017595a5d5a9b18a0030a7ae531dbabdab18f298e21264722303eabded ROLF_norm_hash: 0621305b553d14c934dc235dd9616ff421a67e87342f046923dc5da9bed27029 DENGS_norm_hash: 0fe031672a78ac00d079d51b2c3b1ad3e4eb1c0428bd1bc66b4a2f100f6a7234 ROLF per-iter: 0.000329s | total: 0.329341s DENGS per-iter: 0.062046s | total: 62.045805s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 9.01x (≈ 801% faster) Speedup (per-iter): 15.56x (≈ 1456% faster) Energy Savings: 93.57% ROLV vs cuSPARSE -> Speedup (per-iter): 15.56x | total: 9.01x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "36930b864e45f6c7bc4c05a36ceed9e5546aba4f26c38e27ec94b84500ab052f", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "0fe031672a78ac00d079d51b2c3b1ad3e4eb1c0428bd1bc66b4a2f100f6a7234", "CSR_norm_hash": "7c4936017595a5d5a9b18a0030a7ae531dbabdab18f298e21264722303eabded", "ROLF_norm_hash": "0621305b553d14c934dc235dd9616ff421a67e87342f046923dc5da9bed27029", "DENGS_norm_hash": "0fe031672a78ac00d079d51b2c3b1ad3e4eb1c0428bd1bc66b4a2f100f6a7234", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "91a6790e57791bfb5eb9aeab2ad3128bc32fc47fb2b6dcd8728760921b04c533", "CSR_qhash_d6": "984ba9fb69efd5425438639a22cd5f4a99b1a62be888e965a39f355997aeab0a", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.062073, "pilot_csr_per_iter_s": 0.015665, "rolv_build_s": 0.732619, "rolv_iter_s": 0.001007, "dense_iter_s": 0.015668, "csr_iter_s": 0.01567, "rolv_total_s": 1.739868, "baseline_total_s": 15.668132, "speedup_total_vs_selected_x": 9.005, "speedup_iter_vs_selected_x": 15.555, "rolv_vs_vendor_sparse_iter_x": 15.558, "rolv_vs_vendor_sparse_total_x": 9.007, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 20:40:03] Seed: 123456 | Pattern: block_diagonal | Zeros: 50% A_hash: 8db5189cd07996217967440640b6d42a07f04d0966354d2bccdba45b8f0e85b6 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062054s | CSR: 0.009887s Selected baseline: Sparse CSR (cuSPARSE) ROLV load time (operator build): 0.756429 s ROLV per-iter: 0.001007s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 7633dbd945f281a61e67ced1bd751d6c973d82fffc40b55d3e47154f22949fea (Sparse CSR (cuSPARSE)) CSR_norm_hash: 7633dbd945f281a61e67ced1bd751d6c973d82fffc40b55d3e47154f22949fea ROLF_norm_hash: 1b772b5257f1ebb4c0e1b3eb7d087b0a42bbbeb650b2fd543326e966e21a2cb4 DENGS_norm_hash: 03f3a34956a323cf0aaab8acfc428958d3cdffa022221a76104fbccce492ff66 ROLF per-iter: 0.000329s | total: 0.329397s DENGS per-iter: 0.062039s | total: 62.038797s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 5.60x (≈ 460% faster) Speedup (per-iter): 9.81x (≈ 881% faster) Energy Savings: 89.80% ROLV vs cuSPARSE -> Speedup (per-iter): 9.82x | total: 5.61x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "8db5189cd07996217967440640b6d42a07f04d0966354d2bccdba45b8f0e85b6", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "03f3a34956a323cf0aaab8acfc428958d3cdffa022221a76104fbccce492ff66", "CSR_norm_hash": "7633dbd945f281a61e67ced1bd751d6c973d82fffc40b55d3e47154f22949fea", "ROLF_norm_hash": "1b772b5257f1ebb4c0e1b3eb7d087b0a42bbbeb650b2fd543326e966e21a2cb4", "DENGS_norm_hash": "03f3a34956a323cf0aaab8acfc428958d3cdffa022221a76104fbccce492ff66", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "b5f52a9ef8ffd7efcef97cbd495082fcb11cd7cd2264e12ccaebab8950e1f08e", "CSR_qhash_d6": "8ba828b6c218e1b6e5a5863ba57843db6f5d02ffa09983ee72dfb5fb77938f1f", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.062054, "pilot_csr_per_iter_s": 0.009887, "rolv_build_s": 0.756429, "rolv_iter_s": 0.001007, "dense_iter_s": 0.009879, "csr_iter_s": 0.009892, "rolv_total_s": 1.763774, "baseline_total_s": 9.879288, "speedup_total_vs_selected_x": 5.601, "speedup_iter_vs_selected_x": 9.807, "rolv_vs_vendor_sparse_iter_x": 9.82, "rolv_vs_vendor_sparse_total_x": 5.609, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 20:41:56] Seed: 123456 | Pattern: random | Zeros: 60% A_hash: 3a128a12c751e2a52a9f05427ad881a4beeb441b1aa828f2c83dec9767075e14 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062034s | CSR: 0.326881s Selected baseline: Dense GEMM (cuBLAS) ROLV load time (operator build): 0.750711 s ROLV per-iter: 0.001006s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 82ff97b0c2c9d6b4e6a850bdbeec16cf158da8950cbefe522f043e059a8a944e (Dense GEMM (cuBLAS)) CSR_norm_hash: b66301fbd8653ce218a327674263790f73785fcea907370e07dce35259e5a064 ROLF_norm_hash: 53ee64013770fdb75b32a6e92743de71f64c87429155a238a486d4601379ff1b DENGS_norm_hash: 82ff97b0c2c9d6b4e6a850bdbeec16cf158da8950cbefe522f043e059a8a944e ROLF per-iter: 0.000329s | total: 0.329290s DENGS per-iter: 0.062037s | total: 62.037078s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 35.32x (≈ 3432% faster) Speedup (per-iter): 61.68x (≈ 6068% faster) Energy Savings: 98.38% ROLV vs cuSPARSE -> Speedup (per-iter): 325.05x | total: 186.13x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "3a128a12c751e2a52a9f05427ad881a4beeb441b1aa828f2c83dec9767075e14", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "82ff97b0c2c9d6b4e6a850bdbeec16cf158da8950cbefe522f043e059a8a944e", "CSR_norm_hash": "b66301fbd8653ce218a327674263790f73785fcea907370e07dce35259e5a064", "ROLF_norm_hash": "53ee64013770fdb75b32a6e92743de71f64c87429155a238a486d4601379ff1b", "DENGS_norm_hash": "82ff97b0c2c9d6b4e6a850bdbeec16cf158da8950cbefe522f043e059a8a944e", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "474ef2e5789ba96a76b96db5a6f8e76990d35228b24cd5d7660008bef1c3606c", "CSR_qhash_d6": "109b84881927eaa3a4515daa3a8637791b5b191bfaf330e73be508eacf3e007a", "path_selected": "Dense", "pilot_dense_per_iter_s": 0.062034, "pilot_csr_per_iter_s": 0.326881, "rolv_build_s": 0.750711, "rolv_iter_s": 0.001006, "dense_iter_s": 0.062034, "csr_iter_s": 0.326934, "rolv_total_s": 1.756492, "baseline_total_s": 62.033848, "speedup_total_vs_selected_x": 35.317, "speedup_iter_vs_selected_x": 61.677, "rolv_vs_vendor_sparse_iter_x": 325.055, "rolv_vs_vendor_sparse_total_x": 186.129, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 20:50:04] Seed: 123456 | Pattern: power_law | Zeros: 60% A_hash: 9d19ea5f391575455f95a6f93a0dc330f0816afb109185aa39e76d5e5e3f84a5 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062042s | CSR: 0.303561s Selected baseline: Dense GEMM (cuBLAS) ROLV load time (operator build): 0.758762 s ROLV per-iter: 0.001006s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 3397dfb188f303cce8ca1e8cfc9ceaf57b34d9574df64e8d752935e89f273568 (Dense GEMM (cuBLAS)) CSR_norm_hash: 020f7d2dc40ebe8b34be2ef43ca39dd08ddb44477f7aa1feb5d5daebfaca729c ROLF_norm_hash: d20759f7172d2f43c943a230dae84150fb48c05b8b882f77935ccb62e51e5023 DENGS_norm_hash: 3397dfb188f303cce8ca1e8cfc9ceaf57b34d9574df64e8d752935e89f273568 ROLF per-iter: 0.000329s | total: 0.329382s DENGS per-iter: 0.062034s | total: 62.034445s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 35.16x (≈ 3416% faster) Speedup (per-iter): 61.67x (≈ 6067% faster) Energy Savings: 98.38% ROLV vs cuSPARSE -> Speedup (per-iter): 301.79x | total: 172.03x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "9d19ea5f391575455f95a6f93a0dc330f0816afb109185aa39e76d5e5e3f84a5", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "3397dfb188f303cce8ca1e8cfc9ceaf57b34d9574df64e8d752935e89f273568", "CSR_norm_hash": "020f7d2dc40ebe8b34be2ef43ca39dd08ddb44477f7aa1feb5d5daebfaca729c", "ROLF_norm_hash": "d20759f7172d2f43c943a230dae84150fb48c05b8b882f77935ccb62e51e5023", "DENGS_norm_hash": "3397dfb188f303cce8ca1e8cfc9ceaf57b34d9574df64e8d752935e89f273568", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "054e2c6d65e7082adb830742e210da6458ef0c3b993c9efeb6f8de4af5491a0b", "CSR_qhash_d6": "0cef336c294e5d8ad89976d4d19db74c4ea9bdf070b2aabce6645ba9fefd7ef6", "path_selected": "Dense", "pilot_dense_per_iter_s": 0.062042, "pilot_csr_per_iter_s": 0.303561, "rolv_build_s": 0.758762, "rolv_iter_s": 0.001006, "dense_iter_s": 0.062038, "csr_iter_s": 0.303574, "rolv_total_s": 1.764663, "baseline_total_s": 62.038371, "speedup_total_vs_selected_x": 35.156, "speedup_iter_vs_selected_x": 61.674, "rolv_vs_vendor_sparse_iter_x": 301.793, "rolv_vs_vendor_sparse_total_x": 172.03, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 20:57:45] Seed: 123456 | Pattern: banded | Zeros: 60% A_hash: e78e035e07d681d9c88788fb30448528322d3759de0292aef1030acc8d438be2 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062050s | CSR: 0.012644s Selected baseline: Sparse CSR (cuSPARSE) ROLV load time (operator build): 0.722590 s ROLV per-iter: 0.001006s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 07488ff89f6ce97a4b816e4b09ba61a52fdd3e27d1e162de7cbf1029d1b6a88a (Sparse CSR (cuSPARSE)) CSR_norm_hash: 07488ff89f6ce97a4b816e4b09ba61a52fdd3e27d1e162de7cbf1029d1b6a88a ROLF_norm_hash: 875e17c75ff856b76232765109555a25b7fbfd369a523cfcfc37fecd0cf0a765 DENGS_norm_hash: a917726a5ab831eb4b9c1cbef78f9dab9bf38f1875cc27bd0c7e1a74d85cd51a ROLF per-iter: 0.000329s | total: 0.329263s DENGS per-iter: 0.062045s | total: 62.044512s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 7.32x (≈ 632% faster) Speedup (per-iter): 12.57x (≈ 1157% faster) Energy Savings: 92.05% ROLV vs cuSPARSE -> Speedup (per-iter): 12.57x | total: 7.32x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "e78e035e07d681d9c88788fb30448528322d3759de0292aef1030acc8d438be2", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "a917726a5ab831eb4b9c1cbef78f9dab9bf38f1875cc27bd0c7e1a74d85cd51a", "CSR_norm_hash": "07488ff89f6ce97a4b816e4b09ba61a52fdd3e27d1e162de7cbf1029d1b6a88a", "ROLF_norm_hash": "875e17c75ff856b76232765109555a25b7fbfd369a523cfcfc37fecd0cf0a765", "DENGS_norm_hash": "a917726a5ab831eb4b9c1cbef78f9dab9bf38f1875cc27bd0c7e1a74d85cd51a", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "6c590cb1c86449e9a55609b2add184da816091d6e591141b6417902275b2cba6", "CSR_qhash_d6": "041b8c19bc130ca1ebbaa9691d20a5abf0297a86f95eac44e0fdea8ac148aab5", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.06205, "pilot_csr_per_iter_s": 0.012644, "rolv_build_s": 0.72259, "rolv_iter_s": 0.001006, "dense_iter_s": 0.012646, "csr_iter_s": 0.012645, "rolv_total_s": 1.728454, "baseline_total_s": 12.646156, "speedup_total_vs_selected_x": 7.316, "speedup_iter_vs_selected_x": 12.572, "rolv_vs_vendor_sparse_iter_x": 12.571, "rolv_vs_vendor_sparse_total_x": 7.316, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 20:59:25] Seed: 123456 | Pattern: block_diagonal | Zeros: 60% A_hash: 2b99793bda656b5689cc9f5b049fc1a55ae8c234e0386e439c7204b281ffc158 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062056s | CSR: 0.007965s Selected baseline: Sparse CSR (cuSPARSE) ROLV load time (operator build): 0.746300 s ROLV per-iter: 0.001007s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: b1b8236d6825af4e301cca802ee097d2602013a69950f49dd35d67b262a9ba70 (Sparse CSR (cuSPARSE)) CSR_norm_hash: b1b8236d6825af4e301cca802ee097d2602013a69950f49dd35d67b262a9ba70 ROLF_norm_hash: 968c6fc2b0accf1cdc4ab0d5115cfd7051533a5d02294705346da1e28a71c50b DENGS_norm_hash: 36ee25dfb91d647bc72a00e82673d5af290686eb222c108851af0797263cbfc4 ROLF per-iter: 0.000329s | total: 0.329461s DENGS per-iter: 0.062040s | total: 62.040344s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 4.54x (≈ 354% faster) Speedup (per-iter): 7.91x (≈ 691% faster) Energy Savings: 87.36% ROLV vs cuSPARSE -> Speedup (per-iter): 7.91x | total: 4.54x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "2b99793bda656b5689cc9f5b049fc1a55ae8c234e0386e439c7204b281ffc158", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "36ee25dfb91d647bc72a00e82673d5af290686eb222c108851af0797263cbfc4", "CSR_norm_hash": "b1b8236d6825af4e301cca802ee097d2602013a69950f49dd35d67b262a9ba70", "ROLF_norm_hash": "968c6fc2b0accf1cdc4ab0d5115cfd7051533a5d02294705346da1e28a71c50b", "DENGS_norm_hash": "36ee25dfb91d647bc72a00e82673d5af290686eb222c108851af0797263cbfc4", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "4d9bc3a8c04e309be3d194ede6251942ddf9e71860fd8aa14f66686b71fd0279", "CSR_qhash_d6": "fd3d1e39dc3bae766e20cc42a42b340cfacafcd27bf6d79ed54fc1e671961651", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.062056, "pilot_csr_per_iter_s": 0.007965, "rolv_build_s": 0.7463, "rolv_iter_s": 0.001007, "dense_iter_s": 0.00796, "csr_iter_s": 0.007962, "rolv_total_s": 1.752843, "baseline_total_s": 7.960108, "speedup_total_vs_selected_x": 4.541, "speedup_iter_vs_selected_x": 7.908, "rolv_vs_vendor_sparse_iter_x": 7.91, "rolv_vs_vendor_sparse_total_x": 4.542, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 21:00:57] Seed: 123456 | Pattern: random | Zeros: 70% A_hash: b6d397e4d0e8ebd4f3a13d59f635831bd762ee60284807ed9d008435058ec326 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062057s | CSR: 0.239926s Selected baseline: Dense GEMM (cuBLAS) ROLV load time (operator build): 0.750362 s ROLV per-iter: 0.001007s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 722aa1f103b022093a749ccf9de9cf9003cb678bf7f25648ca7f11ed9adde915 (Dense GEMM (cuBLAS)) CSR_norm_hash: 04fcefd26ce84156909063ab27f8ffa57425def7fd2b784e2668796b39c2b473 ROLF_norm_hash: a503fe60776ce90abef86e50f774bce0642268c79ed36f1ed2a0cfdb456b6090 DENGS_norm_hash: 722aa1f103b022093a749ccf9de9cf9003cb678bf7f25648ca7f11ed9adde915 ROLF per-iter: 0.000329s | total: 0.329496s DENGS per-iter: 0.062037s | total: 62.037156s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 35.31x (≈ 3431% faster) Speedup (per-iter): 61.62x (≈ 6062% faster) Energy Savings: 98.38% ROLV vs cuSPARSE -> Speedup (per-iter): 238.43x | total: 136.61x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "b6d397e4d0e8ebd4f3a13d59f635831bd762ee60284807ed9d008435058ec326", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "722aa1f103b022093a749ccf9de9cf9003cb678bf7f25648ca7f11ed9adde915", "CSR_norm_hash": "04fcefd26ce84156909063ab27f8ffa57425def7fd2b784e2668796b39c2b473", "ROLF_norm_hash": "a503fe60776ce90abef86e50f774bce0642268c79ed36f1ed2a0cfdb456b6090", "DENGS_norm_hash": "722aa1f103b022093a749ccf9de9cf9003cb678bf7f25648ca7f11ed9adde915", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "82da032e08a558947b8496a6df588242839c731dd132f6c51b2ccad1158e1e9e", "CSR_qhash_d6": "80f87fb10e9a24eccb1b683cf6b386eed058b0deac6c880c1a86b28327ebc4e0", "path_selected": "Dense", "pilot_dense_per_iter_s": 0.062057, "pilot_csr_per_iter_s": 0.239926, "rolv_build_s": 0.750362, "rolv_iter_s": 0.001007, "dense_iter_s": 0.062036, "csr_iter_s": 0.24002, "rolv_total_s": 1.757033, "baseline_total_s": 62.035531, "speedup_total_vs_selected_x": 35.307, "speedup_iter_vs_selected_x": 61.624, "rolv_vs_vendor_sparse_iter_x": 238.43, "rolv_vs_vendor_sparse_total_x": 136.605, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 21:07:29] Seed: 123456 | Pattern: power_law | Zeros: 70% A_hash: 64b353290cc661d8798233b459b02627e318c8b6cd03fb9400cdc258605a7257 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062050s | CSR: 0.223276s Selected baseline: Dense GEMM (cuBLAS) ROLV load time (operator build): 1.151080 s ROLV per-iter: 0.001009s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 32c0bbfd6d7a5688cd46fb2b36d62e508c168c1fca43ba3125721e0a93b9b9dc (Dense GEMM (cuBLAS)) CSR_norm_hash: 2e44f0e19d9b94e808905bbe9f3d3c920aed03473e975532c43790263b8b228d ROLF_norm_hash: 72204d453e28b46f2f06b0ac265feda1932e914601162d00494a9d1bfa846619 DENGS_norm_hash: 32c0bbfd6d7a5688cd46fb2b36d62e508c168c1fca43ba3125721e0a93b9b9dc ROLF per-iter: 0.000329s | total: 0.329154s DENGS per-iter: 0.062038s | total: 62.038383s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 28.72x (≈ 2772% faster) Speedup (per-iter): 61.49x (≈ 6049% faster) Energy Savings: 98.37% ROLV vs cuSPARSE -> Speedup (per-iter): 221.39x | total: 103.41x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "64b353290cc661d8798233b459b02627e318c8b6cd03fb9400cdc258605a7257", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "32c0bbfd6d7a5688cd46fb2b36d62e508c168c1fca43ba3125721e0a93b9b9dc", "CSR_norm_hash": "2e44f0e19d9b94e808905bbe9f3d3c920aed03473e975532c43790263b8b228d", "ROLF_norm_hash": "72204d453e28b46f2f06b0ac265feda1932e914601162d00494a9d1bfa846619", "DENGS_norm_hash": "32c0bbfd6d7a5688cd46fb2b36d62e508c168c1fca43ba3125721e0a93b9b9dc", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "b28317e48cc7768973ac901f2af18502a2b95897bacec4775b55b8b869b4083f", "CSR_qhash_d6": "c764b23e1591e4c51b37fee46e921e443af4ed127d6e764301c3add5f86f3992", "path_selected": "Dense", "pilot_dense_per_iter_s": 0.06205, "pilot_csr_per_iter_s": 0.223276, "rolv_build_s": 1.15108, "rolv_iter_s": 0.001009, "dense_iter_s": 0.062036, "csr_iter_s": 0.223347, "rolv_total_s": 2.1599, "baseline_total_s": 62.036035, "speedup_total_vs_selected_x": 28.722, "speedup_iter_vs_selected_x": 61.494, "rolv_vs_vendor_sparse_iter_x": 221.394, "rolv_vs_vendor_sparse_total_x": 103.406, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 21:14:01] Seed: 123456 | Pattern: banded | Zeros: 70% A_hash: 6de52c734dc3dd3e441813467d3974c05babbe147880af95cae93106e22a77bd | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062031s | CSR: 0.009586s Selected baseline: Sparse CSR (cuSPARSE) ROLV load time (operator build): 0.748625 s ROLV per-iter: 0.001006s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: e90bfd484d6e25afc06fa511e93c3e2a01ff75f3a452bad2ade49d5585a35a8c (Sparse CSR (cuSPARSE)) CSR_norm_hash: e90bfd484d6e25afc06fa511e93c3e2a01ff75f3a452bad2ade49d5585a35a8c ROLF_norm_hash: 0f39b126fa71c26f2672d66e7c24c5b70fa1e325cbf4ccf3333c8d2df71581bf DENGS_norm_hash: afa0b5ccf007ed78efa389e675d49ed3175a5e895800ce2c51b65ef34c1c93f8 ROLF per-iter: 0.000329s | total: 0.329453s DENGS per-iter: 0.062046s | total: 62.045695s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 5.47x (≈ 447% faster) Speedup (per-iter): 9.54x (≈ 854% faster) Energy Savings: 89.52% ROLV vs cuSPARSE -> Speedup (per-iter): 9.53x | total: 5.46x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "6de52c734dc3dd3e441813467d3974c05babbe147880af95cae93106e22a77bd", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "afa0b5ccf007ed78efa389e675d49ed3175a5e895800ce2c51b65ef34c1c93f8", "CSR_norm_hash": "e90bfd484d6e25afc06fa511e93c3e2a01ff75f3a452bad2ade49d5585a35a8c", "ROLF_norm_hash": "0f39b126fa71c26f2672d66e7c24c5b70fa1e325cbf4ccf3333c8d2df71581bf", "DENGS_norm_hash": "afa0b5ccf007ed78efa389e675d49ed3175a5e895800ce2c51b65ef34c1c93f8", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "34587fe968cb118281d6e320a80b1d361638e54fc3de87df1dbf85b3f83c9fef", "CSR_qhash_d6": "1f6708025723cea6a42acd40a9941cdb7977691d299c1f970b62ed2ae636700c", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.062031, "pilot_csr_per_iter_s": 0.009586, "rolv_build_s": 0.748625, "rolv_iter_s": 0.001006, "dense_iter_s": 0.009606, "csr_iter_s": 0.009591, "rolv_total_s": 1.755038, "baseline_total_s": 9.605537, "speedup_total_vs_selected_x": 5.473, "speedup_iter_vs_selected_x": 9.544, "rolv_vs_vendor_sparse_iter_x": 9.53, "rolv_vs_vendor_sparse_total_x": 5.465, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 21:15:36] Seed: 123456 | Pattern: block_diagonal | Zeros: 70% A_hash: 605ad79227a409511ccd935bac7446d55792ae15e0550623f778311797a2ba80 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062048s | CSR: 0.006073s Selected baseline: Sparse CSR (cuSPARSE) ROLV load time (operator build): 0.755717 s ROLV per-iter: 0.001006s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: e55eaeda1938519fbb3576fd3659a04bbb4f8b188491e06f47140dd84dbbd64c (Sparse CSR (cuSPARSE)) CSR_norm_hash: e55eaeda1938519fbb3576fd3659a04bbb4f8b188491e06f47140dd84dbbd64c ROLF_norm_hash: 71235fd9a701d922ba41c2836607dbd277b18a94a84c0fe54ad3e3cb46430593 DENGS_norm_hash: afb0000c8a8069b1416a99dc37be6c761f158e933ad2069f4c2a71f2a7f8ef75 ROLF per-iter: 0.000329s | total: 0.329344s DENGS per-iter: 0.062039s | total: 62.039098s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 3.45x (≈ 245% faster) Speedup (per-iter): 6.04x (≈ 504% faster) Energy Savings: 83.44% ROLV vs cuSPARSE -> Speedup (per-iter): 6.05x | total: 3.46x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "605ad79227a409511ccd935bac7446d55792ae15e0550623f778311797a2ba80", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "afb0000c8a8069b1416a99dc37be6c761f158e933ad2069f4c2a71f2a7f8ef75", "CSR_norm_hash": "e55eaeda1938519fbb3576fd3659a04bbb4f8b188491e06f47140dd84dbbd64c", "ROLF_norm_hash": "71235fd9a701d922ba41c2836607dbd277b18a94a84c0fe54ad3e3cb46430593", "DENGS_norm_hash": "afb0000c8a8069b1416a99dc37be6c761f158e933ad2069f4c2a71f2a7f8ef75", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "7bc340a2266a60176174c6afbc41e54623a3acb3c008de5aeff26276a7332fb6", "CSR_qhash_d6": "a1d9c26e56d45d902aea275de0de4c140017e2f77fc217f76f0acc4c6a534b28", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.062048, "pilot_csr_per_iter_s": 0.006073, "rolv_build_s": 0.755717, "rolv_iter_s": 0.001006, "dense_iter_s": 0.006074, "csr_iter_s": 0.00609, "rolv_total_s": 1.761728, "baseline_total_s": 6.074383, "speedup_total_vs_selected_x": 3.448, "speedup_iter_vs_selected_x": 6.038, "rolv_vs_vendor_sparse_iter_x": 6.054, "rolv_vs_vendor_sparse_total_x": 3.457, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 21:17:07] Seed: 123456 | Pattern: random | Zeros: 80% A_hash: fe8ecd469d65375943070e2c9f72b2cb8ffc99f59b8e95e01ee55ff351e8a5b5 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062050s | CSR: 0.157310s Selected baseline: Dense GEMM (cuBLAS) ROLV load time (operator build): 0.810935 s ROLV per-iter: 0.001007s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: e7c01a70a75e7c23f6388af2f7da803ea0bdb8bf7a90f33bfbd68ec38023b5fb (Dense GEMM (cuBLAS)) CSR_norm_hash: e5b8197353d07db1ab9234776adabf080d908ebb1dfb9054f7f1dd4109fb63bc ROLF_norm_hash: 2c97f0d1200024546d7641880cac8a28d456f33e2b9d9806d0408d519cd56f37 DENGS_norm_hash: e7c01a70a75e7c23f6388af2f7da803ea0bdb8bf7a90f33bfbd68ec38023b5fb ROLF per-iter: 0.000329s | total: 0.329448s DENGS per-iter: 0.062038s | total: 62.038199s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 34.13x (≈ 3313% faster) Speedup (per-iter): 61.62x (≈ 6062% faster) Energy Savings: 98.38% ROLV vs cuSPARSE -> Speedup (per-iter): 156.27x | total: 86.55x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "fe8ecd469d65375943070e2c9f72b2cb8ffc99f59b8e95e01ee55ff351e8a5b5", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "e7c01a70a75e7c23f6388af2f7da803ea0bdb8bf7a90f33bfbd68ec38023b5fb", "CSR_norm_hash": "e5b8197353d07db1ab9234776adabf080d908ebb1dfb9054f7f1dd4109fb63bc", "ROLF_norm_hash": "2c97f0d1200024546d7641880cac8a28d456f33e2b9d9806d0408d519cd56f37", "DENGS_norm_hash": "e7c01a70a75e7c23f6388af2f7da803ea0bdb8bf7a90f33bfbd68ec38023b5fb", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "73400dd11bba92807f9dd80ee8c7bdc5e578106e1ea67465c31cebca0c1dc833", "CSR_qhash_d6": "dcf329d56d9c15537236a76138b0852a1803917fe480f352df93ef30032ca862", "path_selected": "Dense", "pilot_dense_per_iter_s": 0.06205, "pilot_csr_per_iter_s": 0.15731, "rolv_build_s": 0.810935, "rolv_iter_s": 0.001007, "dense_iter_s": 0.062037, "csr_iter_s": 0.157322, "rolv_total_s": 1.817689, "baseline_total_s": 62.037477, "speedup_total_vs_selected_x": 34.13, "speedup_iter_vs_selected_x": 61.621, "rolv_vs_vendor_sparse_iter_x": 156.267, "rolv_vs_vendor_sparse_total_x": 86.551, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 21:22:08] Seed: 123456 | Pattern: power_law | Zeros: 80% A_hash: f5319945ed9e0de80929153636dd5033761020445fb403b1998eb9214d00e127 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062042s | CSR: 0.146687s Selected baseline: Dense GEMM (cuBLAS) ROLV load time (operator build): 1.639742 s ROLV per-iter: 0.001008s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 0bbec08c037006aa33c812b530df9811cc2768a08858d9d8f610d0e9dc3f1048 (Dense GEMM (cuBLAS)) CSR_norm_hash: fb76d912e08337c2fb6164533d39b16a78d26d31d5a23c21bc5fcf11de5a90b7 ROLF_norm_hash: 222530270f5d311d6cd32e2219aefd7aa591c2a33d3143d33a5f01187d0d6e1b DENGS_norm_hash: 0bbec08c037006aa33c812b530df9811cc2768a08858d9d8f610d0e9dc3f1048 ROLF per-iter: 0.000329s | total: 0.329424s DENGS per-iter: 0.062038s | total: 62.038172s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 23.43x (≈ 2243% faster) Speedup (per-iter): 61.52x (≈ 6052% faster) Energy Savings: 98.37% ROLV vs cuSPARSE -> Speedup (per-iter): 145.51x | total: 55.41x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "f5319945ed9e0de80929153636dd5033761020445fb403b1998eb9214d00e127", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "0bbec08c037006aa33c812b530df9811cc2768a08858d9d8f610d0e9dc3f1048", "CSR_norm_hash": "fb76d912e08337c2fb6164533d39b16a78d26d31d5a23c21bc5fcf11de5a90b7", "ROLF_norm_hash": "222530270f5d311d6cd32e2219aefd7aa591c2a33d3143d33a5f01187d0d6e1b", "DENGS_norm_hash": "0bbec08c037006aa33c812b530df9811cc2768a08858d9d8f610d0e9dc3f1048", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "66317c61bd76f15b22946d59ad005fceac533aed498ef9ff62ad0904ec173c58", "CSR_qhash_d6": "b6a4c721e0fe7e51c81c42edf62abf2220d94463c774c9aa08c02ddd203941f7", "path_selected": "Dense", "pilot_dense_per_iter_s": 0.062042, "pilot_csr_per_iter_s": 0.146687, "rolv_build_s": 1.639742, "rolv_iter_s": 0.001008, "dense_iter_s": 0.062035, "csr_iter_s": 0.146718, "rolv_total_s": 2.648038, "baseline_total_s": 62.03527, "speedup_total_vs_selected_x": 23.427, "speedup_iter_vs_selected_x": 61.525, "rolv_vs_vendor_sparse_iter_x": 145.511, "rolv_vs_vendor_sparse_total_x": 55.406, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 21:27:25] Seed: 123456 | Pattern: banded | Zeros: 80% A_hash: b2fc7f83b499ca9e4b29ed3cc68b966b4b322cf7926c12186e98ae033e84be58 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062043s | CSR: 0.006498s Selected baseline: Sparse CSR (cuSPARSE) ROLV load time (operator build): 0.948909 s ROLV per-iter: 0.001007s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 4d8d3cad16f030600a156817b8744d12a0c8836b203892968b0604302ec16645 (Sparse CSR (cuSPARSE)) CSR_norm_hash: 4d8d3cad16f030600a156817b8744d12a0c8836b203892968b0604302ec16645 ROLF_norm_hash: e59dd960a001f4488b5b3c7cdcaa8d75712839372bc52a94bab002baba02fd9d DENGS_norm_hash: 7c5ad9bbb7deb3f95a8b23ba1ddfb19f857bf6120ba461e883e8789abd82d6c6 ROLF per-iter: 0.000329s | total: 0.329393s DENGS per-iter: 0.062046s | total: 62.045555s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 3.32x (≈ 232% faster) Speedup (per-iter): 6.45x (≈ 545% faster) Energy Savings: 84.50% ROLV vs cuSPARSE -> Speedup (per-iter): 6.46x | total: 3.33x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "b2fc7f83b499ca9e4b29ed3cc68b966b4b322cf7926c12186e98ae033e84be58", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "7c5ad9bbb7deb3f95a8b23ba1ddfb19f857bf6120ba461e883e8789abd82d6c6", "CSR_norm_hash": "4d8d3cad16f030600a156817b8744d12a0c8836b203892968b0604302ec16645", "ROLF_norm_hash": "e59dd960a001f4488b5b3c7cdcaa8d75712839372bc52a94bab002baba02fd9d", "DENGS_norm_hash": "7c5ad9bbb7deb3f95a8b23ba1ddfb19f857bf6120ba461e883e8789abd82d6c6", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "3d7a5452a9f38bbfb38ee0d4922ca8020b2506f811ff5ec119664b4fe4236084", "CSR_qhash_d6": "0bf1adf2826ddba58cfb34c67ed4ce264c3007a9d7e8b82460dabc3d560e699c", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.062043, "pilot_csr_per_iter_s": 0.006498, "rolv_build_s": 0.948909, "rolv_iter_s": 0.001007, "dense_iter_s": 0.006497, "csr_iter_s": 0.006508, "rolv_total_s": 1.955912, "baseline_total_s": 6.497034, "speedup_total_vs_selected_x": 3.322, "speedup_iter_vs_selected_x": 6.452, "rolv_vs_vendor_sparse_iter_x": 6.463, "rolv_vs_vendor_sparse_total_x": 3.327, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 21:28:53] Seed: 123456 | Pattern: block_diagonal | Zeros: 80% A_hash: 4b02e483523fbec343feac2b8fed3820615bb6832dda42a3da7b63ccf1ef0014 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062046s | CSR: 0.004192s Selected baseline: Sparse CSR (cuSPARSE) ROLV load time (operator build): 0.781035 s ROLV per-iter: 0.001006s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 85db7e29847ea47337867929779ed541af6034889ca6b685c2bfaa776144fc89 (Sparse CSR (cuSPARSE)) CSR_norm_hash: 85db7e29847ea47337867929779ed541af6034889ca6b685c2bfaa776144fc89 ROLF_norm_hash: 50f833cd0edb871dab70fd1076fd14286647900fd24dbd3131295de2f48e9f7b DENGS_norm_hash: 0afabd3c3f898124c4d2adf90b4ec9ca28ee520d74975002112bb8408f023a82 ROLF per-iter: 0.000329s | total: 0.329395s DENGS per-iter: 0.062041s | total: 62.040805s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 2.35x (≈ 135% faster) Speedup (per-iter): 4.17x (≈ 317% faster) Energy Savings: 76.01% ROLV vs cuSPARSE -> Speedup (per-iter): 4.17x | total: 2.35x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "4b02e483523fbec343feac2b8fed3820615bb6832dda42a3da7b63ccf1ef0014", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "0afabd3c3f898124c4d2adf90b4ec9ca28ee520d74975002112bb8408f023a82", "CSR_norm_hash": "85db7e29847ea47337867929779ed541af6034889ca6b685c2bfaa776144fc89", "ROLF_norm_hash": "50f833cd0edb871dab70fd1076fd14286647900fd24dbd3131295de2f48e9f7b", "DENGS_norm_hash": "0afabd3c3f898124c4d2adf90b4ec9ca28ee520d74975002112bb8408f023a82", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "4eaaed0b681ad4346a051280bb2504683f2650e05430ced90b415a8515706ae4", "CSR_qhash_d6": "c10f3c47e9dddb5dc4657b55066dafee57740d7981e917a318c6bf8be7057bf3", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.062046, "pilot_csr_per_iter_s": 0.004192, "rolv_build_s": 0.781035, "rolv_iter_s": 0.001006, "dense_iter_s": 0.004192, "csr_iter_s": 0.004193, "rolv_total_s": 1.78681, "baseline_total_s": 4.191962, "speedup_total_vs_selected_x": 2.346, "speedup_iter_vs_selected_x": 4.168, "rolv_vs_vendor_sparse_iter_x": 4.169, "rolv_vs_vendor_sparse_total_x": 2.347, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 21:30:19] Seed: 123456 | Pattern: random | Zeros: 90% A_hash: 252a6d9ec7eeab4eb29b6c652bffba9f11178919caadeccd14c45d00311e1433 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062044s | CSR: 0.077982s Selected baseline: Dense GEMM (cuBLAS) ROLV load time (operator build): 0.972721 s ROLV per-iter: 0.001007s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 0ea34324829b59e6d5a810b043219ca106a8eb538079e8849cd5903c80796f83 (Dense GEMM (cuBLAS)) CSR_norm_hash: 70a23b3823b0d3991a640ff0cdb8002b963f596dbf7eaa3314467cfc3826958d ROLF_norm_hash: 321943d79a8e5a4c00b38e7a32c69fc5870129ed248809d076b51f4519f056ca DENGS_norm_hash: 0ea34324829b59e6d5a810b043219ca106a8eb538079e8849cd5903c80796f83 ROLF per-iter: 0.000329s | total: 0.329469s DENGS per-iter: 0.062038s | total: 62.038461s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 31.34x (≈ 3034% faster) Speedup (per-iter): 61.62x (≈ 6062% faster) Energy Savings: 98.38% ROLV vs cuSPARSE -> Speedup (per-iter): 77.47x | total: 39.40x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "252a6d9ec7eeab4eb29b6c652bffba9f11178919caadeccd14c45d00311e1433", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "0ea34324829b59e6d5a810b043219ca106a8eb538079e8849cd5903c80796f83", "CSR_norm_hash": "70a23b3823b0d3991a640ff0cdb8002b963f596dbf7eaa3314467cfc3826958d", "ROLF_norm_hash": "321943d79a8e5a4c00b38e7a32c69fc5870129ed248809d076b51f4519f056ca", "DENGS_norm_hash": "0ea34324829b59e6d5a810b043219ca106a8eb538079e8849cd5903c80796f83", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "d07e9d8e4b761c9e713843d9e5ad22646d68738c9c9f78b846a77a566e81f77a", "CSR_qhash_d6": "615b0708dc3f1d8f0760862f37cdd6cfebe6b283a3e02b3e907e7c536c4634c6", "path_selected": "Dense", "pilot_dense_per_iter_s": 0.062044, "pilot_csr_per_iter_s": 0.077982, "rolv_build_s": 0.972721, "rolv_iter_s": 0.001007, "dense_iter_s": 0.062039, "csr_iter_s": 0.077991, "rolv_total_s": 1.97949, "baseline_total_s": 62.038777, "speedup_total_vs_selected_x": 31.341, "speedup_iter_vs_selected_x": 61.622, "rolv_vs_vendor_sparse_iter_x": 77.466, "rolv_vs_vendor_sparse_total_x": 39.399, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 21:34:00] Seed: 123456 | Pattern: power_law | Zeros: 90% A_hash: d1784f30a29c88bb759e8e0ce2e1d3a72ec63f8f7d0190e4b7c74bf9b0f76e26 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062036s | CSR: 0.072838s Selected baseline: Dense GEMM (cuBLAS) ROLV load time (operator build): 1.057907 s ROLV per-iter: 0.001008s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: ff7b7b9d919c85aa942dd3c65988841f5aedc3f475953f6a39377245c2e213f6 (Dense GEMM (cuBLAS)) CSR_norm_hash: a413e0c155d1b9850edaa4b8e32010bfa509df1ecb77bd073a4aa82fd2fef770 ROLF_norm_hash: d25ac0c9367d63ba2b4533bd33d053eefdea95762e4d5dafaf6a2ce9ff4cb49f DENGS_norm_hash: ff7b7b9d919c85aa942dd3c65988841f5aedc3f475953f6a39377245c2e213f6 ROLF per-iter: 0.000329s | total: 0.329366s DENGS per-iter: 0.062034s | total: 62.034211s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 30.02x (≈ 2902% faster) Speedup (per-iter): 61.51x (≈ 6051% faster) Energy Savings: 98.37% ROLV vs cuSPARSE -> Speedup (per-iter): 72.24x | total: 35.26x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "d1784f30a29c88bb759e8e0ce2e1d3a72ec63f8f7d0190e4b7c74bf9b0f76e26", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "ff7b7b9d919c85aa942dd3c65988841f5aedc3f475953f6a39377245c2e213f6", "CSR_norm_hash": "a413e0c155d1b9850edaa4b8e32010bfa509df1ecb77bd073a4aa82fd2fef770", "ROLF_norm_hash": "d25ac0c9367d63ba2b4533bd33d053eefdea95762e4d5dafaf6a2ce9ff4cb49f", "DENGS_norm_hash": "ff7b7b9d919c85aa942dd3c65988841f5aedc3f475953f6a39377245c2e213f6", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "11f8da7a2080d0d605a1ebff2f877f43fdf7d15739e0c108fe9752e7a0e9c93a", "CSR_qhash_d6": "7036227649a1548471b4afd68c57e3a0881fa85c6a8b7fa790a144e1aafb3bd7", "path_selected": "Dense", "pilot_dense_per_iter_s": 0.062036, "pilot_csr_per_iter_s": 0.072838, "rolv_build_s": 1.057907, "rolv_iter_s": 0.001008, "dense_iter_s": 0.062036, "csr_iter_s": 0.072857, "rolv_total_s": 2.066395, "baseline_total_s": 62.035832, "speedup_total_vs_selected_x": 30.021, "speedup_iter_vs_selected_x": 61.514, "rolv_vs_vendor_sparse_iter_x": 72.243, "rolv_vs_vendor_sparse_total_x": 35.258, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 21:37:56] Seed: 123456 | Pattern: banded | Zeros: 90% A_hash: d70a4343e5b268957eb68d7e3674a43f240457ccfda08b4a2d80bc40ab643157 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062043s | CSR: 0.003439s Selected baseline: Sparse CSR (cuSPARSE) ROLV load time (operator build): 0.812415 s ROLV per-iter: 0.001009s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: b8df9b2b2d2c8756b0a2e5f00e2e309fa2e8403a5f1e7290f433f0203e68df58 (Sparse CSR (cuSPARSE)) CSR_norm_hash: b8df9b2b2d2c8756b0a2e5f00e2e309fa2e8403a5f1e7290f433f0203e68df58 ROLF_norm_hash: 7b612cfd403099d0e842a486b4d51c55fba1f9c7ff554a4a1ed5affedc0f75a1 DENGS_norm_hash: 6bab4e46cb1871e51f5424a844af2f1390bcb5c5fb0b3a7c6f421bd1bc78bc94 ROLF per-iter: 0.000329s | total: 0.329381s DENGS per-iter: 0.062046s | total: 62.046457s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 1.89x (≈ 89% faster) Speedup (per-iter): 3.41x (≈ 241% faster) Energy Savings: 70.67% ROLV vs cuSPARSE -> Speedup (per-iter): 3.41x | total: 1.89x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "d70a4343e5b268957eb68d7e3674a43f240457ccfda08b4a2d80bc40ab643157", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "6bab4e46cb1871e51f5424a844af2f1390bcb5c5fb0b3a7c6f421bd1bc78bc94", "CSR_norm_hash": "b8df9b2b2d2c8756b0a2e5f00e2e309fa2e8403a5f1e7290f433f0203e68df58", "ROLF_norm_hash": "7b612cfd403099d0e842a486b4d51c55fba1f9c7ff554a4a1ed5affedc0f75a1", "DENGS_norm_hash": "6bab4e46cb1871e51f5424a844af2f1390bcb5c5fb0b3a7c6f421bd1bc78bc94", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "c850461ae5bfa1c3183f8c5ad70eadff35c42a0cf45abfac99892c98541b884e", "CSR_qhash_d6": "a7e095e393aea28c9a8c6534fc171b7f2ee300221e004fcacaa6d42b795acd44", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.062043, "pilot_csr_per_iter_s": 0.003439, "rolv_build_s": 0.812415, "rolv_iter_s": 0.001009, "dense_iter_s": 0.00344, "csr_iter_s": 0.003444, "rolv_total_s": 1.82135, "baseline_total_s": 3.439779, "speedup_total_vs_selected_x": 1.889, "speedup_iter_vs_selected_x": 3.409, "rolv_vs_vendor_sparse_iter_x": 3.414, "rolv_vs_vendor_sparse_total_x": 1.891, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 21:39:18] Seed: 123456 | Pattern: block_diagonal | Zeros: 90% A_hash: ef3c072370841e3130690e4f6793ea35e3e0c704fce673efdbae340a03091d07 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062047s | CSR: 0.002319s Selected baseline: Sparse CSR (cuSPARSE) ROLV load time (operator build): 0.764816 s ROLV per-iter: 0.001008s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: c4ae745d2e96aa98d3b351f897423007074eb39904008bc22b91b9d13c04799d (Sparse CSR (cuSPARSE)) CSR_norm_hash: c4ae745d2e96aa98d3b351f897423007074eb39904008bc22b91b9d13c04799d ROLF_norm_hash: 043a6ad806a02a4dbfd07edb33a20bdebb586c17d7abfcd015ef28379fdadb27 DENGS_norm_hash: 3954f5e832c294f5f63bb74e0179a360d788ef7079faeb84f69713613cc4ea79 ROLF per-iter: 0.000329s | total: 0.329309s DENGS per-iter: 0.062039s | total: 62.038953s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 1.31x (≈ 31% faster) Speedup (per-iter): 2.30x (≈ 130% faster) Energy Savings: 56.55% ROLV vs cuSPARSE -> Speedup (per-iter): 2.30x | total: 1.31x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "ef3c072370841e3130690e4f6793ea35e3e0c704fce673efdbae340a03091d07", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "3954f5e832c294f5f63bb74e0179a360d788ef7079faeb84f69713613cc4ea79", "CSR_norm_hash": "c4ae745d2e96aa98d3b351f897423007074eb39904008bc22b91b9d13c04799d", "ROLF_norm_hash": "043a6ad806a02a4dbfd07edb33a20bdebb586c17d7abfcd015ef28379fdadb27", "DENGS_norm_hash": "3954f5e832c294f5f63bb74e0179a360d788ef7079faeb84f69713613cc4ea79", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "cffe32a36e15addac2c5b2fef97a992d6d9c8731c108477cdc00f463498f0e00", "CSR_qhash_d6": "180dbad63e572a632a9d39d744e0c9bac459e0543e212631f2e5db2574764101", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.062047, "pilot_csr_per_iter_s": 0.002319, "rolv_build_s": 0.764816, "rolv_iter_s": 0.001008, "dense_iter_s": 0.002319, "csr_iter_s": 0.002319, "rolv_total_s": 1.772417, "baseline_total_s": 2.319079, "speedup_total_vs_selected_x": 1.308, "speedup_iter_vs_selected_x": 2.302, "rolv_vs_vendor_sparse_iter_x": 2.302, "rolv_vs_vendor_sparse_total_x": 1.308, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 21:40:40] Seed: 123456 | Pattern: random | Zeros: 95% A_hash: c926d3fc034ec0adbed3fa6ecc74c1e0c4191486cd48fd095fa3c179c6ef96db | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062050s | CSR: 0.039153s Selected baseline: Sparse CSR (cuSPARSE) ROLV load time (operator build): 1.026166 s ROLV per-iter: 0.001006s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 84c978b53770c26147d968a662799d1d5eaca1367970637c91636a457c6bf9ce (Sparse CSR (cuSPARSE)) CSR_norm_hash: 84c978b53770c26147d968a662799d1d5eaca1367970637c91636a457c6bf9ce ROLF_norm_hash: 438099e42e0c75acf4c952287040709c2d6b5b9b38bb0103770643c2625437f3 DENGS_norm_hash: f5570aa0c2e30ca57e4bfba6db1ba2ed338b13cf6c3a3861cd9efa7d20b15d71 ROLF per-iter: 0.000329s | total: 0.329368s DENGS per-iter: 0.062034s | total: 62.034379s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 19.27x (≈ 1827% faster) Speedup (per-iter): 38.93x (≈ 3793% faster) Energy Savings: 97.43% ROLV vs cuSPARSE -> Speedup (per-iter): 38.93x | total: 19.27x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "c926d3fc034ec0adbed3fa6ecc74c1e0c4191486cd48fd095fa3c179c6ef96db", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "f5570aa0c2e30ca57e4bfba6db1ba2ed338b13cf6c3a3861cd9efa7d20b15d71", "CSR_norm_hash": "84c978b53770c26147d968a662799d1d5eaca1367970637c91636a457c6bf9ce", "ROLF_norm_hash": "438099e42e0c75acf4c952287040709c2d6b5b9b38bb0103770643c2625437f3", "DENGS_norm_hash": "f5570aa0c2e30ca57e4bfba6db1ba2ed338b13cf6c3a3861cd9efa7d20b15d71", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "c341e71c1d6aaaff215d32bf3ce2823faac0512f379a99a19bf0274553f15740", "CSR_qhash_d6": "6b3aff356388fd8668afb15d42c185cffc2b0fe1b37f7cbfe87adcfceea52e09", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.06205, "pilot_csr_per_iter_s": 0.039153, "rolv_build_s": 1.026166, "rolv_iter_s": 0.001006, "dense_iter_s": 0.039158, "csr_iter_s": 0.039159, "rolv_total_s": 2.031953, "baseline_total_s": 39.157895, "speedup_total_vs_selected_x": 19.271, "speedup_iter_vs_selected_x": 38.933, "rolv_vs_vendor_sparse_iter_x": 38.933, "rolv_vs_vendor_sparse_total_x": 19.271, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 21:43:24] Seed: 123456 | Pattern: power_law | Zeros: 95% A_hash: 6417a2a60f09c4389956722addb9e641d9618bcfe0eae0e987dfe602fdefb429 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062042s | CSR: 0.036607s Selected baseline: Sparse CSR (cuSPARSE) ROLV load time (operator build): 0.803289 s ROLV per-iter: 0.001005s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: a5fc0843afe21a83f95d37dc38ca03d5f235fe6631e96b7e7da8039c6a7e42fb (Sparse CSR (cuSPARSE)) CSR_norm_hash: a5fc0843afe21a83f95d37dc38ca03d5f235fe6631e96b7e7da8039c6a7e42fb ROLF_norm_hash: d02f72a24d8562f072ac59b90a05de2bb21f3e093177220577b1c675fdd1e16e DENGS_norm_hash: e1a90360ba8f3a6ce55dfff4d1515996f6b04b522f024ed9ab8124c98d9cb2cf ROLF per-iter: 0.000329s | total: 0.329757s DENGS per-iter: 0.062035s | total: 62.035113s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 20.24x (≈ 1924% faster) Speedup (per-iter): 36.42x (≈ 3542% faster) Energy Savings: 97.25% ROLV vs cuSPARSE -> Speedup (per-iter): 36.42x | total: 20.24x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "6417a2a60f09c4389956722addb9e641d9618bcfe0eae0e987dfe602fdefb429", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "e1a90360ba8f3a6ce55dfff4d1515996f6b04b522f024ed9ab8124c98d9cb2cf", "CSR_norm_hash": "a5fc0843afe21a83f95d37dc38ca03d5f235fe6631e96b7e7da8039c6a7e42fb", "ROLF_norm_hash": "d02f72a24d8562f072ac59b90a05de2bb21f3e093177220577b1c675fdd1e16e", "DENGS_norm_hash": "e1a90360ba8f3a6ce55dfff4d1515996f6b04b522f024ed9ab8124c98d9cb2cf", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "100de79f25f9eba2174c313cdd119f1094c254144c65bddf7539fe46bd2b2afb", "CSR_qhash_d6": "e0a84e8dbc5db086506f4f617ada421bb273ac7d8e29e1427a7cd8c813274d6d", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.062042, "pilot_csr_per_iter_s": 0.036607, "rolv_build_s": 0.803289, "rolv_iter_s": 0.001005, "dense_iter_s": 0.036612, "csr_iter_s": 0.036614, "rolv_total_s": 1.80865, "baseline_total_s": 36.611973, "speedup_total_vs_selected_x": 20.243, "speedup_iter_vs_selected_x": 36.417, "rolv_vs_vendor_sparse_iter_x": 36.419, "rolv_vs_vendor_sparse_total_x": 20.244, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 21:46:43] Seed: 123456 | Pattern: banded | Zeros: 95% A_hash: f9841b629a96caca12ae5093b69047a66277d824418f1f09df0d2ec6bec61381 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062066s | CSR: 0.001913s Selected baseline: Sparse CSR (cuSPARSE) ROLV load time (operator build): 0.838468 s ROLV per-iter: 0.001006s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: d09ca089bda61758eb0c66d1bc2c974dbfe8ca89dfb645769b1dab01237c008d (Sparse CSR (cuSPARSE)) CSR_norm_hash: d09ca089bda61758eb0c66d1bc2c974dbfe8ca89dfb645769b1dab01237c008d ROLF_norm_hash: da09351c75b5339ca035bd5d494f359e4c85eb4b75b0e3f8b5462fc1a53761ba DENGS_norm_hash: a00aed09a9ca1a80f3acaafa376dc0e568949aeb81fcc547c28bc98683803a08 ROLF per-iter: 0.000329s | total: 0.329490s DENGS per-iter: 0.062049s | total: 62.048598s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 1.04x (≈ 4% faster) Speedup (per-iter): 1.90x (≈ 90% faster) Energy Savings: 47.33% ROLV vs cuSPARSE -> Speedup (per-iter): 1.90x | total: 1.04x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "f9841b629a96caca12ae5093b69047a66277d824418f1f09df0d2ec6bec61381", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "a00aed09a9ca1a80f3acaafa376dc0e568949aeb81fcc547c28bc98683803a08", "CSR_norm_hash": "d09ca089bda61758eb0c66d1bc2c974dbfe8ca89dfb645769b1dab01237c008d", "ROLF_norm_hash": "da09351c75b5339ca035bd5d494f359e4c85eb4b75b0e3f8b5462fc1a53761ba", "DENGS_norm_hash": "a00aed09a9ca1a80f3acaafa376dc0e568949aeb81fcc547c28bc98683803a08", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "9c378462296ec1cacef0a364ddaeebca041f1cfda7d5705308d0e32cbdf1f369", "CSR_qhash_d6": "dad36af733c02218be6dbd4a63576ba741c9082474ae89376d4f8b5ecd697ef9", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.062066, "pilot_csr_per_iter_s": 0.001913, "rolv_build_s": 0.838468, "rolv_iter_s": 0.001006, "dense_iter_s": 0.001911, "csr_iter_s": 0.001911, "rolv_total_s": 1.844749, "baseline_total_s": 1.910616, "speedup_total_vs_selected_x": 1.036, "speedup_iter_vs_selected_x": 1.899, "rolv_vs_vendor_sparse_iter_x": 1.899, "rolv_vs_vendor_sparse_total_x": 1.036, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 21:48:03] Seed: 123456 | Pattern: block_diagonal | Zeros: 95% A_hash: 743ed1c8dc03a5de5d0b131edc508c8c9e30dc02e5406aeb9cb6e8c0ce493874 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062048s | CSR: 0.001374s Selected baseline: Sparse CSR (cuSPARSE) ROLV load time (operator build): 1.265122 s ROLV per-iter: 0.001007s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 1e98b0627fac47ec4499c4248603bb5eee16d26c6e711f20b1fdcc9234339f89 (Sparse CSR (cuSPARSE)) CSR_norm_hash: 1e98b0627fac47ec4499c4248603bb5eee16d26c6e711f20b1fdcc9234339f89 ROLF_norm_hash: a23f52d144ae759bed22286b940fe6340572d8562f2e0cdd90f665c8a5c99fa6 DENGS_norm_hash: cb31498379c907686529a18f196926f32e7b1052704f4ed5aaebf0f0e0ed14b1 ROLF per-iter: 0.000329s | total: 0.329715s DENGS per-iter: 0.062040s | total: 62.040117s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 0.61x (≈ -39% faster) Speedup (per-iter): 1.37x (≈ 37% faster) Energy Savings: 26.79% ROLV vs cuSPARSE -> Speedup (per-iter): 1.37x | total: 0.61x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "743ed1c8dc03a5de5d0b131edc508c8c9e30dc02e5406aeb9cb6e8c0ce493874", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "cb31498379c907686529a18f196926f32e7b1052704f4ed5aaebf0f0e0ed14b1", "CSR_norm_hash": "1e98b0627fac47ec4499c4248603bb5eee16d26c6e711f20b1fdcc9234339f89", "ROLF_norm_hash": "a23f52d144ae759bed22286b940fe6340572d8562f2e0cdd90f665c8a5c99fa6", "DENGS_norm_hash": "cb31498379c907686529a18f196926f32e7b1052704f4ed5aaebf0f0e0ed14b1", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "efbe0ca36ca8f3c6721275268db28beadf0ad8a4b2a7aa76452f596348100e65", "CSR_qhash_d6": "ba8ebf59af7aabae5524cf716079f7d04963926d63bc6640c7289ec4a949ab21", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.062048, "pilot_csr_per_iter_s": 0.001374, "rolv_build_s": 1.265122, "rolv_iter_s": 0.001007, "dense_iter_s": 0.001376, "csr_iter_s": 0.001376, "rolv_total_s": 2.272611, "baseline_total_s": 1.376224, "speedup_total_vs_selected_x": 0.606, "speedup_iter_vs_selected_x": 1.366, "rolv_vs_vendor_sparse_iter_x": 1.366, "rolv_vs_vendor_sparse_total_x": 0.606, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 21:49:34] Seed: 123456 | Pattern: random | Zeros: 99% A_hash: 9fde8b5d279f5d4d8297c2b0a4f006d0bf2475b62e6dabc7da09b547c8edbc8a | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062032s | CSR: 0.008145s Selected baseline: Sparse CSR (cuSPARSE) ROLV load time (operator build): 1.365414 s ROLV per-iter: 0.001008s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: b38cc6a6e76a63577fb73a93c840bff8b41c832affd07a36708b747859a51ca0 (Sparse CSR (cuSPARSE)) CSR_norm_hash: b38cc6a6e76a63577fb73a93c840bff8b41c832affd07a36708b747859a51ca0 ROLF_norm_hash: cd16ca2eb816f4e3a4324002312ca0268a38b732796aab7888293ef7f24633b7 DENGS_norm_hash: cc8c3cc839a9d0c5930d0938d01d9bd2a7f5d66f47fcce4e53dec39e0dc7faa9 ROLF per-iter: 0.000329s | total: 0.329465s DENGS per-iter: 0.062037s | total: 62.037168s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 3.44x (≈ 244% faster) Speedup (per-iter): 8.11x (≈ 711% faster) Energy Savings: 87.67% ROLV vs cuSPARSE -> Speedup (per-iter): 8.10x | total: 3.44x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "9fde8b5d279f5d4d8297c2b0a4f006d0bf2475b62e6dabc7da09b547c8edbc8a", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "cc8c3cc839a9d0c5930d0938d01d9bd2a7f5d66f47fcce4e53dec39e0dc7faa9", "CSR_norm_hash": "b38cc6a6e76a63577fb73a93c840bff8b41c832affd07a36708b747859a51ca0", "ROLF_norm_hash": "cd16ca2eb816f4e3a4324002312ca0268a38b732796aab7888293ef7f24633b7", "DENGS_norm_hash": "cc8c3cc839a9d0c5930d0938d01d9bd2a7f5d66f47fcce4e53dec39e0dc7faa9", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "18a55821ec1e53d19620c8b3fdf1bd7dc27837e4d3f0bc4b8bb33ab2e24117eb", "CSR_qhash_d6": "71756758bc815bc77db696fcfc029cf368aa25187dabe5cf998aeea59acfb5ce", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.062032, "pilot_csr_per_iter_s": 0.008145, "rolv_build_s": 1.365414, "rolv_iter_s": 0.001008, "dense_iter_s": 0.008173, "csr_iter_s": 0.00816, "rolv_total_s": 2.372968, "baseline_total_s": 8.172776, "speedup_total_vs_selected_x": 3.444, "speedup_iter_vs_selected_x": 8.112, "rolv_vs_vendor_sparse_iter_x": 8.099, "rolv_vs_vendor_sparse_total_x": 3.439, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 21:51:23] Seed: 123456 | Pattern: power_law | Zeros: 99% A_hash: 3884cba828aa7a1488fc132da5edbcb037e4d5cda60d2548cbb05d1438117888 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062051s | CSR: 0.007630s Selected baseline: Sparse CSR (cuSPARSE) ROLV load time (operator build): 0.832143 s ROLV per-iter: 0.001007s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 24e2c42f91ccf1e6e12fd28ef32ae939c322d0f124b961f1d307c4d34f6e198f (Sparse CSR (cuSPARSE)) CSR_norm_hash: 24e2c42f91ccf1e6e12fd28ef32ae939c322d0f124b961f1d307c4d34f6e198f ROLF_norm_hash: 4727304f2130748ccc072bc431f0db2c0cbe41049c05e3323613e00e8f930359 DENGS_norm_hash: c3e9496be8a189fc75d306823ff6359f8fa3b5aa5557bbc72bae19d4a7ed7d5d ROLF per-iter: 0.000329s | total: 0.329341s DENGS per-iter: 0.062037s | total: 62.037355s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 4.16x (≈ 316% faster) Speedup (per-iter): 7.60x (≈ 660% faster) Energy Savings: 86.85% ROLV vs cuSPARSE -> Speedup (per-iter): 7.60x | total: 4.16x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "3884cba828aa7a1488fc132da5edbcb037e4d5cda60d2548cbb05d1438117888", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "c3e9496be8a189fc75d306823ff6359f8fa3b5aa5557bbc72bae19d4a7ed7d5d", "CSR_norm_hash": "24e2c42f91ccf1e6e12fd28ef32ae939c322d0f124b961f1d307c4d34f6e198f", "ROLF_norm_hash": "4727304f2130748ccc072bc431f0db2c0cbe41049c05e3323613e00e8f930359", "DENGS_norm_hash": "c3e9496be8a189fc75d306823ff6359f8fa3b5aa5557bbc72bae19d4a7ed7d5d", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "570879d26c1763504bd05013868ba03d6e5c343019d405fbb6664799c3446a0a", "CSR_qhash_d6": "accd3dd4bbfc190043334b65cc33688934cc529738d5462481a4c7e632fbdda4", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.062051, "pilot_csr_per_iter_s": 0.00763, "rolv_build_s": 0.832143, "rolv_iter_s": 0.001007, "dense_iter_s": 0.007653, "csr_iter_s": 0.007656, "rolv_total_s": 1.838941, "baseline_total_s": 7.653371, "speedup_total_vs_selected_x": 4.162, "speedup_iter_vs_selected_x": 7.602, "rolv_vs_vendor_sparse_iter_x": 7.604, "rolv_vs_vendor_sparse_total_x": 4.163, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 21:54:00] Seed: 123456 | Pattern: banded | Zeros: 99% A_hash: 1b643fe5ac4811868b9b5bfee7d7ed4d02a612b4add98ac2d0f399d014599b67 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062062s | CSR: 0.000696s Selected baseline: Sparse CSR (cuSPARSE) ROLV load time (operator build): 0.832876 s ROLV per-iter: 0.001006s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 3fe8ac1fab005fbd1e63982dee89afe3d065e979ad01a731d3ebf842a3648e4e (Sparse CSR (cuSPARSE)) CSR_norm_hash: 3fe8ac1fab005fbd1e63982dee89afe3d065e979ad01a731d3ebf842a3648e4e ROLF_norm_hash: 832a9e85bddee020a7655ba9a16031e464caa87f1ceb92d8567b75780ccb7c6d DENGS_norm_hash: 2974cb2e57d28e2c52f6b64f51fb5c2c10ab828f6e7279b652ec7d4c7fa407ad ROLF per-iter: 0.000329s | total: 0.329290s DENGS per-iter: 0.062052s | total: 62.052086s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 0.38x (≈ -62% faster) Speedup (per-iter): 0.69x (≈ -31% faster) Energy Savings: -44.45% ROLV vs cuSPARSE -> Speedup (per-iter): 0.69x | total: 0.38x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "1b643fe5ac4811868b9b5bfee7d7ed4d02a612b4add98ac2d0f399d014599b67", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "2974cb2e57d28e2c52f6b64f51fb5c2c10ab828f6e7279b652ec7d4c7fa407ad", "CSR_norm_hash": "3fe8ac1fab005fbd1e63982dee89afe3d065e979ad01a731d3ebf842a3648e4e", "ROLF_norm_hash": "832a9e85bddee020a7655ba9a16031e464caa87f1ceb92d8567b75780ccb7c6d", "DENGS_norm_hash": "2974cb2e57d28e2c52f6b64f51fb5c2c10ab828f6e7279b652ec7d4c7fa407ad", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "36fabc0207e13e7ae08a8c6db45526167c20c66464e356ba7cd4969570c1cb52", "CSR_qhash_d6": "ef059811d87737b49d1022a11bffb21a5179c83ffdaf83784cd83d02836065b0", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.062062, "pilot_csr_per_iter_s": 0.000696, "rolv_build_s": 0.832876, "rolv_iter_s": 0.001006, "dense_iter_s": 0.000697, "csr_iter_s": 0.000696, "rolv_total_s": 1.838995, "baseline_total_s": 0.696515, "speedup_total_vs_selected_x": 0.379, "speedup_iter_vs_selected_x": 0.692, "rolv_vs_vendor_sparse_iter_x": 0.692, "rolv_vs_vendor_sparse_total_x": 0.379, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

[2025-12-08 21:55:17] Seed: 123456 | Pattern: block_diagonal | Zeros: 99% A_hash: d78e202117fb1b5ee60605254db62aa72b0d2b72a9d6ceec1a84ad78c44df368 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.062061s | CSR: 0.000621s Selected baseline: Sparse CSR (cuSPARSE) ROLV load time (operator build): 1.015146 s ROLV per-iter: 0.001007s ROLV_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: a87d97bfb7c2c30e63f5a68be7b57390d7f240f5ad68a6c8ba553bbb6b913de8 (Sparse CSR (cuSPARSE)) CSR_norm_hash: a87d97bfb7c2c30e63f5a68be7b57390d7f240f5ad68a6c8ba553bbb6b913de8 ROLF_norm_hash: aedd4cf9d03fcc98a2d4b33b9c56bc89ad6ca996f94ce0f83609d26a240cfcd3 DENGS_norm_hash: 81df416748e59ff8fb7b3e31c4a3a74db121c9fc011e70c1604e496e3b107c2b ROLF per-iter: 0.000337s | total: 0.336894s DENGS per-iter: 0.062042s | total: 62.042180s Correctness vs Selected Baseline: Verified | vs CSR: Verified Speedup (total): 0.31x (≈ -69% faster) Speedup (per-iter): 0.62x (≈ -38% faster) Energy Savings: -61.81% ROLV vs cuSPARSE -> Speedup (per-iter): 0.62x | total: 0.31x {"platform": "CUDA", "device": "NVIDIA B200", "adapted_batch": false, "effective_batch": 5000, "dense_label": "cuBLAS", "sparse_label": "cuSPARSE", "input_hash_A": "d78e202117fb1b5ee60605254db62aa72b0d2b72a9d6ceec1a84ad78c44df368", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "81df416748e59ff8fb7b3e31c4a3a74db121c9fc011e70c1604e496e3b107c2b", "CSR_norm_hash": "a87d97bfb7c2c30e63f5a68be7b57390d7f240f5ad68a6c8ba553bbb6b913de8", "ROLF_norm_hash": "aedd4cf9d03fcc98a2d4b33b9c56bc89ad6ca996f94ce0f83609d26a240cfcd3", "DENGS_norm_hash": "81df416748e59ff8fb7b3e31c4a3a74db121c9fc011e70c1604e496e3b107c2b", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "72ae2e8b1b11989b240bd407be5eae8d1ffdbf75beeacce50c056cdb544c412f", "CSR_qhash_d6": "2f7bfb9127eae8277e2d1d70e892b6d1a1c2ada0aa940a4bba915f09e9f01c96", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.062061, "pilot_csr_per_iter_s": 0.000621, "rolv_build_s": 1.015146, "rolv_iter_s": 0.001007, "dense_iter_s": 0.000622, "csr_iter_s": 0.000623, "rolv_total_s": 2.022349, "baseline_total_s": 0.622465, "speedup_total_vs_selected_x": 0.308, "speedup_iter_vs_selected_x": 0.618, "rolv_vs_vendor_sparse_iter_x": 0.618, "rolv_vs_vendor_sparse_total_x": 0.308, "energy_iter_adaptive_telemetry": null, "telemetry_samples": 0, "correct_norm": "OK"}

=== FOOTER REPORT (CUDA) ===

Aggregate speedup (total vs selected): 15.51x (≈ 1451% faster)
Aggregate speedup (per-iter vs selected): 29.46x (≈ 2846% faster)
Aggregate energy savings (proxy vs selected): 79.0%
Verification: TF32 off, deterministic algorithms, CSR canonicalization, CPU-fp64 normalization and SHA-256 hashing. {"platform": "CUDA", "device": "NVIDIA B200", "aggregate_speedup_total_vs_selected_x": 15.509, "aggregate_speedup_iter_vs_selected_x": 29.456, "aggregate_energy_savings_pct": 78.973, "verification": "TF32 off, deterministic algorithms, CSR canonicalization, CPU-fp64 normalization, SHA-256 hashing"}

=== Timing & Energy Measurement Explanation ===

Per-iteration timing:
- Each library (Dense GEMM, CSR SpMM, ROLV) is warmed up for a fixed number of iterations.
- Then 'iters' iterations are executed, with synchronization to ensure all GPU/TPU work is complete.
- The average time per iteration is reported as _iter_s.
Build/setup time:
- For ROLV, operator construction (tiling, quantization, surrogate build) is timed separately as rolv_build_s.
- Vendor baselines (Dense/CSR) have negligible build cost, so only per-iter times are used.
Total time:
- For each library, total runtime = build/setup time + (per-iter time × number of iterations).
- Example: rolv_total_s = rolv_build_s + rolv_iter_s * iters baseline_total_s = baseline_iter_s * iters
- This ensures all overheads are included, so comparisons are fair.
Speedup calculation:
- Speedup (per-iter) = baseline_iter_s / rolv_iter_s
- Speedup (total) = baseline_total_s / rolv_total_s
- Both metrics are reported to show raw kernel efficiency and end-to-end cost.
Energy measurement:
- Proxy energy savings are computed from per-iter times: energy_savings_pct = 100 × (1 - rolv_iter_s / baseline_iter_s)
- If telemetry is enabled (NVML/ROCm SMI), instantaneous power samples (W) are integrated over time to yield Joules (trapz).
- Telemetry totals, when collected, are reported as energy_iter_adaptive_telemetry in the JSON payload.
Fairness guarantee:
- All libraries run the same matrix/vector inputs (identical seeds, identical input hashes).
- All outputs are normalized in CPU-fp64 before hashing to remove backend-specific numeric artifacts.
- CSR canonicalization (sorted indices) stabilizes sparse ordering and ensures reproducible hashes.
- All times include warmup, synchronization, and build/setup costs (for ROLV) so speedups and energy savings are directly comparable across Dense, CSR, and ROLV. “Imagination is the only Limitation to Innovation” Rolv E. Heggenhougen ================================================

Intel Benchmarks 12/10/2025Xeon CPU === RUN SUITE (CPU) on CPU === [2025-12-10 11:12:47] Platform: CPU | Seed: 123456 | Pattern: random | Zeros: 40% A_hash: 4c73788bc9a9e39550de335aa10ba05e9ed71a3aa8ce12d27fa93a2ce83f123e | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 /tmp/ipython-input-251883314.py:374: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /pytorch/aten/src/ATen/SparseCsrTensorImpl.cpp:53.) A_csr_raw = A_dense.to_sparse_csr() Baseline pilots per-iter -> Dense: 0.062780s | CSR: 0.133077s | COO: 0.780264s | ELL: 5.819332s Selected baseline: Dense rolv load time (operator build): 1.856413 s rolv per-iter: 0.001885s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 21aecb3e10ac2069591fe79012b1257f2f05d05a8e6ed94f988b403d1b955ed6 (Dense) CSR_norm_hash: 103dfd8771d09a1c35f87435c8913cd880a3100da1f21e7eb08cd882de55f72d COO_norm_hash: 103dfd8771d09a1c35f87435c8913cd880a3100da1f21e7eb08cd882de55f72d ELL_norm_hash: fe5fb0c3127784ae85850d6b6b7facd558b797b2a96b5cc29dd3c1ec5bb2c68f ROLF_norm_hash: e750a7854e4262786dc9e374fb5cc7ce53ffd636aa476d44146ec6c4ed55d086 DENGS_norm_hash: 21aecb3e10ac2069591fe79012b1257f2f05d05a8e6ed94f988b403d1b955ed6 COO per-iter: 0.766074s | total: 766.074429s CSR per-iter: 0.136301s | total: 136.301073s ELL per-iter: 5.819026s | total: 5819.425355s ROLF per-iter: 0.001161s | total: 1.163469s DENGS per-iter: 0.037825s | total: 37.825388s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 10.27x (≈ 927% faster) Speedup (per-iter): 20.38x (≈ 1938% faster) Energy Savings (proxy): 95.09% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 72.29x | total: 36.43x rolv vs COO: Speedup (per-iter): 406.32x | total: 204.73x rolv vs ELL: Speedup (per-iter): 3086.37x | total: 1555.24x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "4c73788bc9a9e39550de335aa10ba05e9ed71a3aa8ce12d27fa93a2ce83f123e", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "21aecb3e10ac2069591fe79012b1257f2f05d05a8e6ed94f988b403d1b955ed6", "CSR_norm_hash": "103dfd8771d09a1c35f87435c8913cd880a3100da1f21e7eb08cd882de55f72d", "COO_norm_hash": "103dfd8771d09a1c35f87435c8913cd880a3100da1f21e7eb08cd882de55f72d", "ELL_norm_hash": "fe5fb0c3127784ae85850d6b6b7facd558b797b2a96b5cc29dd3c1ec5bb2c68f", "ROLF_norm_hash": "e750a7854e4262786dc9e374fb5cc7ce53ffd636aa476d44146ec6c4ed55d086", "DENGS_norm_hash": "21aecb3e10ac2069591fe79012b1257f2f05d05a8e6ed94f988b403d1b955ed6", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "7602756637c1a50a635821ff9922240f7b5fd83878a4c557db058a2fa10d82a2", "CSR_qhash_d6": "a6f169a1022de5b553014a33cd9c097e7d030f6316d65b2826faff3240c91618", "COO_qhash_d6": "a6f169a1022de5b553014a33cd9c097e7d030f6316d65b2826faff3240c91618", "ELL_qhash_d6": "691ae3980682df696b43055d50641211b68d7c682aeab4732aee07704e083644", "path_selected": "Dense", "pilot_dense_per_iter_s": 0.06278, "pilot_csr_per_iter_s": 0.133077, "pilot_coo_per_iter_s": 0.780264, "pilot_ell_per_iter_s": 5.819332, "rolv_build_s": 1.856413, "rolv_iter_s": 0.001885, "dense_iter_s": 0.038417, "csr_iter_s": 0.136301, "coo_iter_s": 0.766074, "ell_iter_s": 5.819026, "rolv_total_s": 3.741808, "baseline_total_s": 38.416785, "speedup_total_vs_selected_x": 10.267, "speedup_iter_vs_selected_x": 20.376, "rolv_vs_vendor_sparse_iter_x": 72.293, "rolv_vs_vendor_sparse_total_x": 36.427, "rolv_vs_coo_iter_x": 406.32, "rolv_vs_coo_total_x": 204.734, "rolv_vs_ell_iter_x": 3086.369, "rolv_vs_ell_total_x": 1555.244, "correct_norm": "OK"}

[2025-12-10 13:08:18] Platform: CPU | Seed: 123456 | Pattern: power_law | Zeros: 40% A_hash: c9336e8fbf29443952f70d5003372b73093ff69b9619649c6f2df5830b1a4056 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.036021s | CSR: 0.130226s | COO: 0.759276s | ELL: 5.769723s Selected baseline: Dense rolv load time (operator build): 1.800192 s rolv per-iter: 0.001490s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: e95feed32a85a487845c9cfcb851cccf64428b5e1c11959ce35fef4b536b6b80 (Dense) CSR_norm_hash: e8c3ae84fb9d818161dfc58925e7e745b3f956e3abb58b9d1330ebd3c40da596 COO_norm_hash: e8c3ae84fb9d818161dfc58925e7e745b3f956e3abb58b9d1330ebd3c40da596 ELL_norm_hash: a9f2c09e5526cda1c2ad904364c03deee38b6136a77c39916c977de97bfc8079 ROLF_norm_hash: 3f022921b021d9e6ca34b8b0f4849294b7dce11a7aa0e7199fea14bdc90f4622 DENGS_norm_hash: e95feed32a85a487845c9cfcb851cccf64428b5e1c11959ce35fef4b536b6b80 COO per-iter: 0.767699s | total: 767.699001s CSR per-iter: 0.135485s | total: 135.484682s ELL per-iter: 5.814724s | total: 5815.038002s ROLF per-iter: 0.001137s | total: 1.139562s DENGS per-iter: 0.038237s | total: 38.236677s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 11.39x (≈ 1039% faster) Speedup (per-iter): 25.15x (≈ 2415% faster) Energy Savings (proxy): 96.02% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 90.93x | total: 41.18x rolv vs COO: Speedup (per-iter): 515.26x | total: 233.34x rolv vs ELL: Speedup (per-iter): 3902.72x | total: 1767.43x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "c9336e8fbf29443952f70d5003372b73093ff69b9619649c6f2df5830b1a4056", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "e95feed32a85a487845c9cfcb851cccf64428b5e1c11959ce35fef4b536b6b80", "CSR_norm_hash": "e8c3ae84fb9d818161dfc58925e7e745b3f956e3abb58b9d1330ebd3c40da596", "COO_norm_hash": "e8c3ae84fb9d818161dfc58925e7e745b3f956e3abb58b9d1330ebd3c40da596", "ELL_norm_hash": "a9f2c09e5526cda1c2ad904364c03deee38b6136a77c39916c977de97bfc8079", "ROLF_norm_hash": "3f022921b021d9e6ca34b8b0f4849294b7dce11a7aa0e7199fea14bdc90f4622", "DENGS_norm_hash": "e95feed32a85a487845c9cfcb851cccf64428b5e1c11959ce35fef4b536b6b80", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "93f115f2931ad246272cdb457d18130510ef0fbf6b0c1a7c98c85edb453a2320", "CSR_qhash_d6": "8e00b085a0b6122ee091dabfb91f9c22e98b7834ab7610e716cb47d33c125621", "COO_qhash_d6": "8e00b085a0b6122ee091dabfb91f9c22e98b7834ab7610e716cb47d33c125621", "ELL_qhash_d6": "83f586f076a980d63b931c65ba19bd8ff6ef501fd447ba4b0d9e9d79bfd2a9a3", "path_selected": "Dense", "pilot_dense_per_iter_s": 0.036021, "pilot_csr_per_iter_s": 0.130226, "pilot_coo_per_iter_s": 0.759276, "pilot_ell_per_iter_s": 5.769723, "rolv_build_s": 1.800192, "rolv_iter_s": 0.00149, "dense_iter_s": 0.037479, "csr_iter_s": 0.135485, "coo_iter_s": 0.767699, "ell_iter_s": 5.814724, "rolv_total_s": 3.290109, "baseline_total_s": 37.478561, "speedup_total_vs_selected_x": 11.391, "speedup_iter_vs_selected_x": 25.155, "rolv_vs_vendor_sparse_iter_x": 90.934, "rolv_vs_vendor_sparse_total_x": 41.179, "rolv_vs_coo_iter_x": 515.263, "rolv_vs_coo_total_x": 233.335, "rolv_vs_ell_iter_x": 3902.716, "rolv_vs_ell_total_x": 1767.43, "correct_norm": "OK"}

[2025-12-10 15:03:40] Platform: CPU | Seed: 123456 | Pattern: banded | Zeros: 40% A_hash: 3bde9c3b2fd3b2a62bc0a638d1d9c6503de9bdbd46c306c75a6bbee0c590a2f9 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.044849s | CSR: 0.005211s | COO: 0.027409s | ELL: 0.306448s Selected baseline: CSR rolv load time (operator build): 1.847498 s rolv per-iter: 0.001687s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 69f4a5c913a3b11ebf15ed7b55579e84daef021d7f71e15669de763e3a48d055 (CSR) CSR_norm_hash: 69f4a5c913a3b11ebf15ed7b55579e84daef021d7f71e15669de763e3a48d055 COO_norm_hash: 69f4a5c913a3b11ebf15ed7b55579e84daef021d7f71e15669de763e3a48d055 ELL_norm_hash: 8298b1d6073e898adac62f34b2554501079c87340a1c6c8b2ea19f4031f1fab2 ROLF_norm_hash: c3959404c2f4afcb6a707b0a4c758ea51a71f95fd4415bc1b6d1dc9df43527ba DENGS_norm_hash: d3b1b922715343c5ea72bfcfcba3272fca34cc51bc421a3f18a23e45a649ef85 COO per-iter: 0.027901s | total: 27.900765s CSR per-iter: 0.004219s | total: 4.218945s ELL per-iter: 0.319621s | total: 319.736728s ROLF per-iter: 0.000860s | total: 0.862356s DENGS per-iter: 0.037666s | total: 37.665588s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 0.96x (≈ -4% faster) Speedup (per-iter): 2.00x (≈ 100% faster) Energy Savings (proxy): 50.02% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 2.50x | total: 1.19x rolv vs COO: Speedup (per-iter): 16.54x | total: 7.89x rolv vs ELL: Speedup (per-iter): 189.47x | total: 90.46x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "3bde9c3b2fd3b2a62bc0a638d1d9c6503de9bdbd46c306c75a6bbee0c590a2f9", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "d3b1b922715343c5ea72bfcfcba3272fca34cc51bc421a3f18a23e45a649ef85", "CSR_norm_hash": "69f4a5c913a3b11ebf15ed7b55579e84daef021d7f71e15669de763e3a48d055", "COO_norm_hash": "69f4a5c913a3b11ebf15ed7b55579e84daef021d7f71e15669de763e3a48d055", "ELL_norm_hash": "8298b1d6073e898adac62f34b2554501079c87340a1c6c8b2ea19f4031f1fab2", "ROLF_norm_hash": "c3959404c2f4afcb6a707b0a4c758ea51a71f95fd4415bc1b6d1dc9df43527ba", "DENGS_norm_hash": "d3b1b922715343c5ea72bfcfcba3272fca34cc51bc421a3f18a23e45a649ef85", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "d2170ce654981ea8c4ffb3747fe45f52b06d5487f4e8f20038dc897f4171b843", "CSR_qhash_d6": "13b0ce1aa75e18d63d63a16049f62649b93ea11335fc8869a362dc98c84f062c", "COO_qhash_d6": "13b0ce1aa75e18d63d63a16049f62649b93ea11335fc8869a362dc98c84f062c", "ELL_qhash_d6": "be7bb31beb804cee1db31cd24d6b143f864c205c2d582917e263e2231b3656a5", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.044849, "pilot_csr_per_iter_s": 0.005211, "pilot_coo_per_iter_s": 0.027409, "pilot_ell_per_iter_s": 0.306448, "rolv_build_s": 1.847498, "rolv_iter_s": 0.001687, "dense_iter_s": 0.003375, "csr_iter_s": 0.004219, "coo_iter_s": 0.027901, "ell_iter_s": 0.319621, "rolv_total_s": 3.534379, "baseline_total_s": 3.375409, "speedup_total_vs_selected_x": 0.955, "speedup_iter_vs_selected_x": 2.001, "rolv_vs_vendor_sparse_iter_x": 2.501, "rolv_vs_vendor_sparse_total_x": 1.194, "rolv_vs_coo_iter_x": 16.54, "rolv_vs_coo_total_x": 7.894, "rolv_vs_ell_iter_x": 189.475, "rolv_vs_ell_total_x": 90.465, "correct_norm": "OK"}

[2025-12-10 15:10:25] Platform: CPU | Seed: 123456 | Pattern: block_diagonal | Zeros: 40% A_hash: 5888e65e5a10fecbee0bea1d7d4aa35dcc2183a34aa4ecc900e012f1019cd48a | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.036458s | CSR: 0.002470s | COO: 0.017182s | ELL: 0.394667s Selected baseline: CSR rolv load time (operator build): 1.882139 s rolv per-iter: 0.001958s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: b01d8c5c1167b8370e99425b74918597e51ee3936430d62b001626d557061979 (CSR) CSR_norm_hash: b01d8c5c1167b8370e99425b74918597e51ee3936430d62b001626d557061979 COO_norm_hash: b01d8c5c1167b8370e99425b74918597e51ee3936430d62b001626d557061979 ELL_norm_hash: 029f17a40d4ee611a2ed0b56fa3f261fd37e52cacb0d907257d3b4562a45f123 ROLF_norm_hash: 5546266b253bb6404112934c7d925e2c174737febcec3060fc9e650ab51785bb DENGS_norm_hash: 8ff706976b5de74778c99eb8537c6b5e8d67492f0f6bf43d3dc1806736234878 COO per-iter: 0.017626s | total: 17.626145s CSR per-iter: 0.002262s | total: 2.262320s ELL per-iter: 0.424377s | total: 424.454144s ROLF per-iter: 0.000864s | total: 0.866015s DENGS per-iter: 0.037506s | total: 37.505597s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 0.59x (≈ -41% faster) Speedup (per-iter): 1.15x (≈ 15% faster) Energy Savings (proxy): 13.11% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 1.16x | total: 0.59x rolv vs COO: Speedup (per-iter): 9.00x | total: 4.59x rolv vs ELL: Speedup (per-iter): 216.72x | total: 110.53x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "5888e65e5a10fecbee0bea1d7d4aa35dcc2183a34aa4ecc900e012f1019cd48a", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "8ff706976b5de74778c99eb8537c6b5e8d67492f0f6bf43d3dc1806736234878", "CSR_norm_hash": "b01d8c5c1167b8370e99425b74918597e51ee3936430d62b001626d557061979", "COO_norm_hash": "b01d8c5c1167b8370e99425b74918597e51ee3936430d62b001626d557061979", "ELL_norm_hash": "029f17a40d4ee611a2ed0b56fa3f261fd37e52cacb0d907257d3b4562a45f123", "ROLF_norm_hash": "5546266b253bb6404112934c7d925e2c174737febcec3060fc9e650ab51785bb", "DENGS_norm_hash": "8ff706976b5de74778c99eb8537c6b5e8d67492f0f6bf43d3dc1806736234878", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "b9db1764459ea7faeddcb613ad791b6dc6918cb4c75269df783764b5d8479289", "CSR_qhash_d6": "33f2f2818226e282854dfee66f97c84d3645271b79c6093f75a5bbe5e122dd57", "COO_qhash_d6": "33f2f2818226e282854dfee66f97c84d3645271b79c6093f75a5bbe5e122dd57", "ELL_qhash_d6": "cd09f09b41e201371129e9616e3b01f427c9a35e468bf26ac4f578424dfe66b3", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.036458, "pilot_csr_per_iter_s": 0.00247, "pilot_coo_per_iter_s": 0.017182, "pilot_ell_per_iter_s": 0.394667, "rolv_build_s": 1.882139, "rolv_iter_s": 0.001958, "dense_iter_s": 0.002254, "csr_iter_s": 0.002262, "coo_iter_s": 0.017626, "ell_iter_s": 0.424377, "rolv_total_s": 3.840307, "baseline_total_s": 2.253549, "speedup_total_vs_selected_x": 0.587, "speedup_iter_vs_selected_x": 1.151, "rolv_vs_vendor_sparse_iter_x": 1.155, "rolv_vs_vendor_sparse_total_x": 0.589, "rolv_vs_coo_iter_x": 9.001, "rolv_vs_coo_total_x": 4.59, "rolv_vs_ell_iter_x": 216.721, "rolv_vs_ell_total_x": 110.526, "correct_norm": "OK"}

[2025-12-10 15:18:42] Platform: CPU | Seed: 123456 | Pattern: random | Zeros: 50% A_hash: 45a7c0782476630ef958d70423217f394bd5f7aa4846c79d82a5f94fc853f75a | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.036578s | CSR: 0.163289s | COO: 0.643147s | ELL: 4.841854s Selected baseline: Dense rolv load time (operator build): 1.835658 s rolv per-iter: 0.001444s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 5cc98ee834848ac3c8539436dc9be5e6a3ab8ddac943b6b8b16296b450545911 (Dense) CSR_norm_hash: a5b74a08926e2a6267e489167848a8461c5f457a4792d19f945be7a266106815 COO_norm_hash: a5b74a08926e2a6267e489167848a8461c5f457a4792d19f945be7a266106815 ELL_norm_hash: fc8ba16578a5790d6ca7ff2b4818e5026d6e187d57f4bb09ea5fcc57de64daf5 ROLF_norm_hash: 86f7cfc426e816a9ae6b7917e0f52828b2726d6fd0d061f84b15ab89e19ca7ed DENGS_norm_hash: 5cc98ee834848ac3c8539436dc9be5e6a3ab8ddac943b6b8b16296b450545911 COO per-iter: 0.669503s | total: 669.502773s CSR per-iter: 0.115684s | total: 115.684163s ELL per-iter: 4.946401s | total: 4946.703714s ROLF per-iter: 0.000843s | total: 0.845624s DENGS per-iter: 0.037636s | total: 37.636395s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 11.40x (≈ 1040% faster) Speedup (per-iter): 25.88x (≈ 2488% faster) Energy Savings (proxy): 96.14% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 80.11x | total: 35.27x rolv vs COO: Speedup (per-iter): 463.63x | total: 204.14x rolv vs ELL: Speedup (per-iter): 3425.40x | total: 1508.28x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "45a7c0782476630ef958d70423217f394bd5f7aa4846c79d82a5f94fc853f75a", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "5cc98ee834848ac3c8539436dc9be5e6a3ab8ddac943b6b8b16296b450545911", "CSR_norm_hash": "a5b74a08926e2a6267e489167848a8461c5f457a4792d19f945be7a266106815", "COO_norm_hash": "a5b74a08926e2a6267e489167848a8461c5f457a4792d19f945be7a266106815", "ELL_norm_hash": "fc8ba16578a5790d6ca7ff2b4818e5026d6e187d57f4bb09ea5fcc57de64daf5", "ROLF_norm_hash": "86f7cfc426e816a9ae6b7917e0f52828b2726d6fd0d061f84b15ab89e19ca7ed", "DENGS_norm_hash": "5cc98ee834848ac3c8539436dc9be5e6a3ab8ddac943b6b8b16296b450545911", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "3d9123de68e8902207b7eb890239d9d8c8181cf222b826286b082aa8926423ee", "CSR_qhash_d6": "b24c665dec415532ef7548d393dbe46dbfe7cc3448c401a0f206cc140a51b240", "COO_qhash_d6": "b24c665dec415532ef7548d393dbe46dbfe7cc3448c401a0f206cc140a51b240", "ELL_qhash_d6": "67dc5245807264fddc3a8224375c61ff1b82e1c317988993a545021773a25365", "path_selected": "Dense", "pilot_dense_per_iter_s": 0.036578, "pilot_csr_per_iter_s": 0.163289, "pilot_coo_per_iter_s": 0.643147, "pilot_ell_per_iter_s": 4.841854, "rolv_build_s": 1.835658, "rolv_iter_s": 0.001444, "dense_iter_s": 0.037374, "csr_iter_s": 0.115684, "coo_iter_s": 0.669503, "ell_iter_s": 4.946401, "rolv_total_s": 3.279694, "baseline_total_s": 37.374486, "speedup_total_vs_selected_x": 11.396, "speedup_iter_vs_selected_x": 25.882, "rolv_vs_vendor_sparse_iter_x": 80.112, "rolv_vs_vendor_sparse_total_x": 35.273, "rolv_vs_coo_iter_x": 463.633, "rolv_vs_coo_total_x": 204.136, "rolv_vs_ell_iter_x": 3425.399, "rolv_vs_ell_total_x": 1508.282, "correct_norm": "OK"}

[2025-12-10 16:57:26] Platform: CPU | Seed: 123456 | Pattern: power_law | Zeros: 50% A_hash: 965f7f7d80a5d6059fae47aebe00cc05a484b022cb40c545b585ae1efba484d7 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.034543s | CSR: 0.108467s | COO: 0.634112s | ELL: 4.926911s Selected baseline: Dense rolv load time (operator build): 1.872530 s rolv per-iter: 0.001623s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 269d89aca368824b1c994853e753aa7d76df5d8b2cee2297bd6b544b61c73364 (Dense) CSR_norm_hash: 9e5d70eb9b9e931548a70fbfc139fb4359e0c83084d585301daa3bfd855090a6 COO_norm_hash: 9e5d70eb9b9e931548a70fbfc139fb4359e0c83084d585301daa3bfd855090a6 ELL_norm_hash: 8ad313d83978421dfc91bf5743841eada784cbbaa205ad887a7ffe8692cc49df ROLF_norm_hash: 9374300fdbac7745cef2ceddacdeac548bb2b329581e2301aa2487a5ecd28626 DENGS_norm_hash: 269d89aca368824b1c994853e753aa7d76df5d8b2cee2297bd6b544b61c73364 COO per-iter: 0.664123s | total: 664.122625s CSR per-iter: 0.121576s | total: 121.575840s ELL per-iter: 4.971765s | total: 4972.067197s ROLF per-iter: 0.000845s | total: 0.848003s DENGS per-iter: 0.038173s | total: 38.172818s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 11.09x (≈ 1009% faster) Speedup (per-iter): 23.89x (≈ 2289% faster) Energy Savings (proxy): 95.81% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 74.91x | total: 34.78x rolv vs COO: Speedup (per-iter): 409.20x | total: 189.99x rolv vs ELL: Speedup (per-iter): 3063.35x | total: 1422.41x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "965f7f7d80a5d6059fae47aebe00cc05a484b022cb40c545b585ae1efba484d7", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "269d89aca368824b1c994853e753aa7d76df5d8b2cee2297bd6b544b61c73364", "CSR_norm_hash": "9e5d70eb9b9e931548a70fbfc139fb4359e0c83084d585301daa3bfd855090a6", "COO_norm_hash": "9e5d70eb9b9e931548a70fbfc139fb4359e0c83084d585301daa3bfd855090a6", "ELL_norm_hash": "8ad313d83978421dfc91bf5743841eada784cbbaa205ad887a7ffe8692cc49df", "ROLF_norm_hash": "9374300fdbac7745cef2ceddacdeac548bb2b329581e2301aa2487a5ecd28626", "DENGS_norm_hash": "269d89aca368824b1c994853e753aa7d76df5d8b2cee2297bd6b544b61c73364", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "f26e652ee8cac77ade50066fe44ef48bc7990abb9e4089de905ff98bfee4da90", "CSR_qhash_d6": "adb34d574efbcee26bc8c45f076961f2fd055f564b4cfda55657c9e1799a18be", "COO_qhash_d6": "adb34d574efbcee26bc8c45f076961f2fd055f564b4cfda55657c9e1799a18be", "ELL_qhash_d6": "c7ef3fbd57ce909a88d94a3d5e87990263cf1549ccec885df32652ac175a09fb", "path_selected": "Dense", "pilot_dense_per_iter_s": 0.034543, "pilot_csr_per_iter_s": 0.108467, "pilot_coo_per_iter_s": 0.634112, "pilot_ell_per_iter_s": 4.926911, "rolv_build_s": 1.87253, "rolv_iter_s": 0.001623, "dense_iter_s": 0.038769, "csr_iter_s": 0.121576, "coo_iter_s": 0.664123, "ell_iter_s": 4.971765, "rolv_total_s": 3.495512, "baseline_total_s": 38.769365, "speedup_total_vs_selected_x": 11.091, "speedup_iter_vs_selected_x": 23.888, "rolv_vs_vendor_sparse_iter_x": 74.909, "rolv_vs_vendor_sparse_total_x": 34.781, "rolv_vs_coo_iter_x": 409.199, "rolv_vs_coo_total_x": 189.993, "rolv_vs_ell_iter_x": 3063.353, "rolv_vs_ell_total_x": 1422.415, "correct_norm": "OK"}

[2025-12-10 18:36:33] Platform: CPU | Seed: 123456 | Pattern: banded | Zeros: 50% A_hash: 1ddcb1159ccd802ded709bd9c232540d9c69faa3ab3409ce540e72058690f33d | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.060922s | CSR: 0.005281s | COO: 0.026698s | ELL: 0.263479s Selected baseline: CSR rolv load time (operator build): 1.890682 s rolv per-iter: 0.001916s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 4f8f518cf334a549ce31497a04e41b716cd0092130966bb22617ce6128dae6aa (CSR) CSR_norm_hash: 4f8f518cf334a549ce31497a04e41b716cd0092130966bb22617ce6128dae6aa COO_norm_hash: 4f8f518cf334a549ce31497a04e41b716cd0092130966bb22617ce6128dae6aa ELL_norm_hash: 1ee4a7c534d4ae87a60dbbf1ec5a72d57c32b7a967df2e415ba2d4595c2125f8 ROLF_norm_hash: eca9f80707f863983a67e01b0dc9d5afc00193b18f0188d22ffc7a904b2cc05e DENGS_norm_hash: 47d6388010582e71bba7d4a996b61e12f4fd83f8f457ecbfa5298c55514a687c COO per-iter: 0.023901s | total: 23.900965s CSR per-iter: 0.003819s | total: 3.818614s ELL per-iter: 0.275169s | total: 275.297473s ROLF per-iter: 0.000854s | total: 0.857086s DENGS per-iter: 0.038712s | total: 38.711778s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 0.79x (≈ -21% faster) Speedup (per-iter): 1.56x (≈ 56% faster) Energy Savings (proxy): 36.01% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 1.99x | total: 1.00x rolv vs COO: Speedup (per-iter): 12.48x | total: 6.28x rolv vs ELL: Speedup (per-iter): 143.64x | total: 72.33x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "1ddcb1159ccd802ded709bd9c232540d9c69faa3ab3409ce540e72058690f33d", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "47d6388010582e71bba7d4a996b61e12f4fd83f8f457ecbfa5298c55514a687c", "CSR_norm_hash": "4f8f518cf334a549ce31497a04e41b716cd0092130966bb22617ce6128dae6aa", "COO_norm_hash": "4f8f518cf334a549ce31497a04e41b716cd0092130966bb22617ce6128dae6aa", "ELL_norm_hash": "1ee4a7c534d4ae87a60dbbf1ec5a72d57c32b7a967df2e415ba2d4595c2125f8", "ROLF_norm_hash": "eca9f80707f863983a67e01b0dc9d5afc00193b18f0188d22ffc7a904b2cc05e", "DENGS_norm_hash": "47d6388010582e71bba7d4a996b61e12f4fd83f8f457ecbfa5298c55514a687c", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "7830c6c8243dda152717aa90122d1ae0971d03f5795695071347590daa7391ca", "CSR_qhash_d6": "745c69200754034d81f6fc7b16fad0f54225eec8ac8e194e2df3b3d341c835ad", "COO_qhash_d6": "745c69200754034d81f6fc7b16fad0f54225eec8ac8e194e2df3b3d341c835ad", "ELL_qhash_d6": "da56587d84f96a443eee8808745d905f5a94c0184458f34d0ce4862c038d2aaf", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.060922, "pilot_csr_per_iter_s": 0.005281, "pilot_coo_per_iter_s": 0.026698, "pilot_ell_per_iter_s": 0.263479, "rolv_build_s": 1.890682, "rolv_iter_s": 0.001916, "dense_iter_s": 0.002994, "csr_iter_s": 0.003819, "coo_iter_s": 0.023901, "ell_iter_s": 0.275169, "rolv_total_s": 3.806369, "baseline_total_s": 2.993617, "speedup_total_vs_selected_x": 0.786, "speedup_iter_vs_selected_x": 1.563, "rolv_vs_vendor_sparse_iter_x": 1.993, "rolv_vs_vendor_sparse_total_x": 1.003, "rolv_vs_coo_iter_x": 12.476, "rolv_vs_coo_total_x": 6.279, "rolv_vs_ell_iter_x": 143.64, "rolv_vs_ell_total_x": 72.325, "correct_norm": "OK"}

[2025-12-10 18:42:30] Platform: CPU | Seed: 123456 | Pattern: block_diagonal | Zeros: 50% A_hash: 0ed5a234c555d3a7c4fa71f3c9fdc858d280adee5455e26fdcd17393449e3a59 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.035816s | CSR: 0.001996s | COO: 0.014732s | ELL: 0.331788s Selected baseline: CSR rolv load time (operator build): 2.155547 s rolv per-iter: 0.001794s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 9e8b39c815ec55fa1b0d218f4fc2a471d9fb1b16e0f21495b30532c1620053a5 (CSR) CSR_norm_hash: 9e8b39c815ec55fa1b0d218f4fc2a471d9fb1b16e0f21495b30532c1620053a5 COO_norm_hash: 9e8b39c815ec55fa1b0d218f4fc2a471d9fb1b16e0f21495b30532c1620053a5 ELL_norm_hash: 47f4992d97d5d3228507b0361b39c2050f25bb189441f6cd34644fdd6b51ade9 ROLF_norm_hash: 7472a05704d3591781355599cd00c5666383557823564e3e3a74b47672e199ad DENGS_norm_hash: 1b22ac89b669da8c14f66c253c5d785e938e7c59321ad3f8e92686e4bb1007c2 COO per-iter: 0.015209s | total: 15.209294s CSR per-iter: 0.002001s | total: 2.001323s ELL per-iter: 0.381257s | total: 381.339563s ROLF per-iter: 0.000862s | total: 0.864758s DENGS per-iter: 0.038252s | total: 38.252093s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 0.53x (≈ -47% faster) Speedup (per-iter): 1.16x (≈ 16% faster) Energy Savings (proxy): 13.68% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 1.12x | total: 0.51x rolv vs COO: Speedup (per-iter): 8.48x | total: 3.85x rolv vs ELL: Speedup (per-iter): 212.52x | total: 96.55x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "0ed5a234c555d3a7c4fa71f3c9fdc858d280adee5455e26fdcd17393449e3a59", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "1b22ac89b669da8c14f66c253c5d785e938e7c59321ad3f8e92686e4bb1007c2", "CSR_norm_hash": "9e8b39c815ec55fa1b0d218f4fc2a471d9fb1b16e0f21495b30532c1620053a5", "COO_norm_hash": "9e8b39c815ec55fa1b0d218f4fc2a471d9fb1b16e0f21495b30532c1620053a5", "ELL_norm_hash": "47f4992d97d5d3228507b0361b39c2050f25bb189441f6cd34644fdd6b51ade9", "ROLF_norm_hash": "7472a05704d3591781355599cd00c5666383557823564e3e3a74b47672e199ad", "DENGS_norm_hash": "1b22ac89b669da8c14f66c253c5d785e938e7c59321ad3f8e92686e4bb1007c2", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "e3b18e6bea6e42d89d8613adb468e13b135d235a34f4fc8edb31ad66aec5cc97", "CSR_qhash_d6": "cf8e452b1291baaa994c2812e778b6cd5abb7a821a7b88c19b9d68b70fa5d056", "COO_qhash_d6": "cf8e452b1291baaa994c2812e778b6cd5abb7a821a7b88c19b9d68b70fa5d056", "ELL_qhash_d6": "f1711a0d389bfdea2b12be82f727bcdb05b1257c264e3ea3e3a0796e34e1bf28", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.035816, "pilot_csr_per_iter_s": 0.001996, "pilot_coo_per_iter_s": 0.014732, "pilot_ell_per_iter_s": 0.331788, "rolv_build_s": 2.155547, "rolv_iter_s": 0.001794, "dense_iter_s": 0.002078, "csr_iter_s": 0.002001, "coo_iter_s": 0.015209, "ell_iter_s": 0.381257, "rolv_total_s": 3.949535, "baseline_total_s": 2.078255, "speedup_total_vs_selected_x": 0.526, "speedup_iter_vs_selected_x": 1.158, "rolv_vs_vendor_sparse_iter_x": 1.116, "rolv_vs_vendor_sparse_total_x": 0.507, "rolv_vs_coo_iter_x": 8.478, "rolv_vs_coo_total_x": 3.851, "rolv_vs_ell_iter_x": 212.519, "rolv_vs_ell_total_x": 96.553, "correct_norm": "OK"}

[2025-12-10 18:50:01] Platform: CPU | Seed: 123456 | Pattern: random | Zeros: 60% A_hash: c0a8dc0e049ce454cdd7dd179da68c3fa2332cd8aa1a11fdd0271576247bd6cd | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.035751s | CSR: 0.099384s | COO: 0.542260s | ELL: 3.976969s Selected baseline: Dense rolv load time (operator build): 1.862188 s rolv per-iter: 0.001887s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 3749d9ea285b55bd550110f1547a67b55929e8f58b1480847f3da4ac6d43eb07 (Dense) CSR_norm_hash: 13da6976b8b3ac1175a4c7bdcc69b36b86afa4ca3e6f79d3a86335ec9a44483f COO_norm_hash: 13da6976b8b3ac1175a4c7bdcc69b36b86afa4ca3e6f79d3a86335ec9a44483f ELL_norm_hash: c746ebd67bd797406180b1941ec8289db2203c931393c93b5e2377064773b94c ROLF_norm_hash: 8f493c2fd4d54f1f1a08baa0352ca67949dbbbea0d8b7b8eeae06289c7292cb5 DENGS_norm_hash: 3749d9ea285b55bd550110f1547a67b55929e8f58b1480847f3da4ac6d43eb07 COO per-iter: 0.551911s | total: 551.911313s CSR per-iter: 0.095802s | total: 95.802105s ELL per-iter: 3.997985s | total: 3998.248330s ROLF per-iter: 0.000991s | total: 0.993757s DENGS per-iter: 0.038410s | total: 38.410063s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 10.27x (≈ 927% faster) Speedup (per-iter): 20.40x (≈ 1940% faster) Energy Savings (proxy): 95.10% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 50.76x | total: 25.55x rolv vs COO: Speedup (per-iter): 292.45x | total: 147.20x rolv vs ELL: Speedup (per-iter): 2118.47x | total: 1066.37x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "c0a8dc0e049ce454cdd7dd179da68c3fa2332cd8aa1a11fdd0271576247bd6cd", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "3749d9ea285b55bd550110f1547a67b55929e8f58b1480847f3da4ac6d43eb07", "CSR_norm_hash": "13da6976b8b3ac1175a4c7bdcc69b36b86afa4ca3e6f79d3a86335ec9a44483f", "COO_norm_hash": "13da6976b8b3ac1175a4c7bdcc69b36b86afa4ca3e6f79d3a86335ec9a44483f", "ELL_norm_hash": "c746ebd67bd797406180b1941ec8289db2203c931393c93b5e2377064773b94c", "ROLF_norm_hash": "8f493c2fd4d54f1f1a08baa0352ca67949dbbbea0d8b7b8eeae06289c7292cb5", "DENGS_norm_hash": "3749d9ea285b55bd550110f1547a67b55929e8f58b1480847f3da4ac6d43eb07", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "c897957c5a4f9990e556dfd615c0ad2affe25039498bdac3b40a261387c8e248", "CSR_qhash_d6": "ab1f327b4885d6e9bfaa0c306e0a694006d5f1d81b4e736efbf2437003654856", "COO_qhash_d6": "ab1f327b4885d6e9bfaa0c306e0a694006d5f1d81b4e736efbf2437003654856", "ELL_qhash_d6": "362301821f1ff8d8a5728ed27752aff1d8bc531cce97156a5163e76872755905", "path_selected": "Dense", "pilot_dense_per_iter_s": 0.035751, "pilot_csr_per_iter_s": 0.099384, "pilot_coo_per_iter_s": 0.54226, "pilot_ell_per_iter_s": 3.976969, "rolv_build_s": 1.862188, "rolv_iter_s": 0.001887, "dense_iter_s": 0.038499, "csr_iter_s": 0.095802, "coo_iter_s": 0.551911, "ell_iter_s": 3.997985, "rolv_total_s": 3.749394, "baseline_total_s": 38.498602, "speedup_total_vs_selected_x": 10.268, "speedup_iter_vs_selected_x": 20.4, "rolv_vs_vendor_sparse_iter_x": 50.764, "rolv_vs_vendor_sparse_total_x": 25.551, "rolv_vs_coo_iter_x": 292.449, "rolv_vs_coo_total_x": 147.2, "rolv_vs_ell_iter_x": 2118.468, "rolv_vs_ell_total_x": 1066.372, "correct_norm": "OK"}

[2025-12-10 20:10:21] Platform: CPU | Seed: 123456 | Pattern: power_law | Zeros: 60% A_hash: 069ee261d27a00145238997c362ebb98b924c75a6f91a5c1b8d06db556292af7 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.063603s | CSR: 0.122710s | COO: 0.531372s | ELL: 4.019755s Selected baseline: Dense rolv load time (operator build): 1.835632 s rolv per-iter: 0.001856s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: ec5754b649d7e679672ead308376fdbdc6623a435dbfee6f417827ddf82e5f5b (Dense) CSR_norm_hash: 04a3cb70dd35cb9e685f05bf2c3cd97b1cd694fc776c870cb134eb52aebaa381 COO_norm_hash: 04a3cb70dd35cb9e685f05bf2c3cd97b1cd694fc776c870cb134eb52aebaa381 ELL_norm_hash: 980815ffbec56a269ccc1c76c8dda03b5510065b603aa58f0b3070334c5214f3 ROLF_norm_hash: 63538076ad4204607732e77e3268f724f9b88231c51f4dc3e93d691a7485c2b6 DENGS_norm_hash: ec5754b649d7e679672ead308376fdbdc6623a435dbfee6f417827ddf82e5f5b COO per-iter: 0.540496s | total: 540.496012s CSR per-iter: 0.095307s | total: 95.306986s ELL per-iter: 3.954836s | total: 3955.081461s ROLF per-iter: 0.000844s | total: 0.846742s DENGS per-iter: 0.038834s | total: 38.834420s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 10.51x (≈ 951% faster) Speedup (per-iter): 20.90x (≈ 1990% faster) Energy Savings (proxy): 95.21% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 51.36x | total: 25.82x rolv vs COO: Speedup (per-iter): 291.25x | total: 146.42x rolv vs ELL: Speedup (per-iter): 2131.06x | total: 1071.42x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "069ee261d27a00145238997c362ebb98b924c75a6f91a5c1b8d06db556292af7", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "ec5754b649d7e679672ead308376fdbdc6623a435dbfee6f417827ddf82e5f5b", "CSR_norm_hash": "04a3cb70dd35cb9e685f05bf2c3cd97b1cd694fc776c870cb134eb52aebaa381", "COO_norm_hash": "04a3cb70dd35cb9e685f05bf2c3cd97b1cd694fc776c870cb134eb52aebaa381", "ELL_norm_hash": "980815ffbec56a269ccc1c76c8dda03b5510065b603aa58f0b3070334c5214f3", "ROLF_norm_hash": "63538076ad4204607732e77e3268f724f9b88231c51f4dc3e93d691a7485c2b6", "DENGS_norm_hash": "ec5754b649d7e679672ead308376fdbdc6623a435dbfee6f417827ddf82e5f5b", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "89090f9d1cd5d6ead841956c1f4a0707ece1853ff3bb202c46b7224ae01a435e", "CSR_qhash_d6": "2ebf593bd3ea44e6595255cb0864f4a71874fc658678a382ae68e73b4801d464", "COO_qhash_d6": "2ebf593bd3ea44e6595255cb0864f4a71874fc658678a382ae68e73b4801d464", "ELL_qhash_d6": "88b930814ce942dc6070cf089174f92c0c9a3235ab7d5e3017b63e07add0c554", "path_selected": "Dense", "pilot_dense_per_iter_s": 0.063603, "pilot_csr_per_iter_s": 0.12271, "pilot_coo_per_iter_s": 0.531372, "pilot_ell_per_iter_s": 4.019755, "rolv_build_s": 1.835632, "rolv_iter_s": 0.001856, "dense_iter_s": 0.038783, "csr_iter_s": 0.095307, "coo_iter_s": 0.540496, "ell_iter_s": 3.954836, "rolv_total_s": 3.69144, "baseline_total_s": 38.782508, "speedup_total_vs_selected_x": 10.506, "speedup_iter_vs_selected_x": 20.898, "rolv_vs_vendor_sparse_iter_x": 51.356, "rolv_vs_vendor_sparse_total_x": 25.818, "rolv_vs_coo_iter_x": 291.246, "rolv_vs_coo_total_x": 146.419, "rolv_vs_ell_iter_x": 2131.059, "rolv_vs_ell_total_x": 1071.42, "correct_norm": "OK"}

[2025-12-10 21:29:42] Platform: CPU | Seed: 123456 | Pattern: banded | Zeros: 60% A_hash: 847932963bf1f1ebfafcdc1d4cd3fb3f897c2d6a403ca0a1c9c501f65d1a1449 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.035300s | CSR: 0.002257s | COO: 0.018720s | ELL: 0.225630s Selected baseline: CSR rolv load time (operator build): 1.869394 s rolv per-iter: 0.001805s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 12e8f730ffa3781b5845fd07558092b58d3ea067156f27ff56280a35de3feca4 (CSR) CSR_norm_hash: 12e8f730ffa3781b5845fd07558092b58d3ea067156f27ff56280a35de3feca4 COO_norm_hash: 12e8f730ffa3781b5845fd07558092b58d3ea067156f27ff56280a35de3feca4 ELL_norm_hash: a02949f6d30b20a712b619e6ce933b3b9f50cb6c8f2c0a41b9ee5a14adef2d65 ROLF_norm_hash: 7a1914416353064289aa06aeb77ec2afa7f71a0e7914dfb2f7155a4b2c50ebd2 DENGS_norm_hash: 8d7f510f86bef203d62040c1554cae338757d4f6196fbb15591f11b27303e707 COO per-iter: 0.019112s | total: 19.112360s CSR per-iter: 0.002620s | total: 2.620247s ELL per-iter: 0.232898s | total: 233.016401s ROLF per-iter: 0.000836s | total: 0.838455s DENGS per-iter: 0.038965s | total: 38.965297s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 0.71x (≈ -29% faster) Speedup (per-iter): 1.44x (≈ 44% faster) Energy Savings (proxy): 30.38% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 1.45x | total: 0.71x rolv vs COO: Speedup (per-iter): 10.59x | total: 5.20x rolv vs ELL: Speedup (per-iter): 129.02x | total: 63.41x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "847932963bf1f1ebfafcdc1d4cd3fb3f897c2d6a403ca0a1c9c501f65d1a1449", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "8d7f510f86bef203d62040c1554cae338757d4f6196fbb15591f11b27303e707", "CSR_norm_hash": "12e8f730ffa3781b5845fd07558092b58d3ea067156f27ff56280a35de3feca4", "COO_norm_hash": "12e8f730ffa3781b5845fd07558092b58d3ea067156f27ff56280a35de3feca4", "ELL_norm_hash": "a02949f6d30b20a712b619e6ce933b3b9f50cb6c8f2c0a41b9ee5a14adef2d65", "ROLF_norm_hash": "7a1914416353064289aa06aeb77ec2afa7f71a0e7914dfb2f7155a4b2c50ebd2", "DENGS_norm_hash": "8d7f510f86bef203d62040c1554cae338757d4f6196fbb15591f11b27303e707", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "012ae712a04f19272c6833ba507261a3ee4677c5efc58db5b2680d1fd46ca320", "CSR_qhash_d6": "44f792ba162f1d197ec18a519718b87043950d855ad5ae315f9c1e37706fed12", "COO_qhash_d6": "44f792ba162f1d197ec18a519718b87043950d855ad5ae315f9c1e37706fed12", "ELL_qhash_d6": "de81f8c5c71bbb54ba6f11b9e2c028dce63724b414fd6826845ab3e0867e76ab", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.0353, "pilot_csr_per_iter_s": 0.002257, "pilot_coo_per_iter_s": 0.01872, "pilot_ell_per_iter_s": 0.22563, "rolv_build_s": 1.869394, "rolv_iter_s": 0.001805, "dense_iter_s": 0.002593, "csr_iter_s": 0.00262, "coo_iter_s": 0.019112, "ell_iter_s": 0.232898, "rolv_total_s": 3.674591, "baseline_total_s": 2.5931, "speedup_total_vs_selected_x": 0.706, "speedup_iter_vs_selected_x": 1.436, "rolv_vs_vendor_sparse_iter_x": 1.452, "rolv_vs_vendor_sparse_total_x": 0.713, "rolv_vs_coo_iter_x": 10.587, "rolv_vs_coo_total_x": 5.201, "rolv_vs_ell_iter_x": 129.016, "rolv_vs_ell_total_x": 63.413, "correct_norm": "OK"}

[2025-12-10 21:34:49] Platform: CPU | Seed: 123456 | Pattern: block_diagonal | Zeros: 60% A_hash: a966633c7116eae1259550f37e6baa481735947ed8261c6c68623e72275589aa | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.034786s | CSR: 0.001605s | COO: 0.012016s | ELL: 0.254058s Selected baseline: CSR rolv load time (operator build): 1.960221 s rolv per-iter: 0.001653s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 18fca23d79c5fb6dd1aba9013cd8c63f23d6e868e0b3414ea1eabf48577148e5 (CSR) CSR_norm_hash: 18fca23d79c5fb6dd1aba9013cd8c63f23d6e868e0b3414ea1eabf48577148e5 COO_norm_hash: 18fca23d79c5fb6dd1aba9013cd8c63f23d6e868e0b3414ea1eabf48577148e5 ELL_norm_hash: dc1e78f806d59bef5944260df784c532bc34161c0031a8b01f60bb9f02d17667 ROLF_norm_hash: 428de573a9c4292d0afa25b086ee188f970200e4dc2420185a0a985800b5e843 DENGS_norm_hash: 9469cf477afa522430730a61a60afcdfa296cce23971aa28fa49dbe642db4def COO per-iter: 0.012182s | total: 12.181873s CSR per-iter: 0.002096s | total: 2.096218s ELL per-iter: 0.262926s | total: 263.011034s ROLF per-iter: 0.000842s | total: 0.844178s DENGS per-iter: 0.037945s | total: 37.944546s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 0.53x (≈ -47% faster) Speedup (per-iter): 1.16x (≈ 16% faster) Energy Savings (proxy): 14.02% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 1.27x | total: 0.58x rolv vs COO: Speedup (per-iter): 7.37x | total: 3.37x rolv vs ELL: Speedup (per-iter): 159.08x | total: 72.80x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "a966633c7116eae1259550f37e6baa481735947ed8261c6c68623e72275589aa", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "9469cf477afa522430730a61a60afcdfa296cce23971aa28fa49dbe642db4def", "CSR_norm_hash": "18fca23d79c5fb6dd1aba9013cd8c63f23d6e868e0b3414ea1eabf48577148e5", "COO_norm_hash": "18fca23d79c5fb6dd1aba9013cd8c63f23d6e868e0b3414ea1eabf48577148e5", "ELL_norm_hash": "dc1e78f806d59bef5944260df784c532bc34161c0031a8b01f60bb9f02d17667", "ROLF_norm_hash": "428de573a9c4292d0afa25b086ee188f970200e4dc2420185a0a985800b5e843", "DENGS_norm_hash": "9469cf477afa522430730a61a60afcdfa296cce23971aa28fa49dbe642db4def", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "7c29acc9777ffc7f4223a17f3f346617023f23866cbe5e9d1eeb992942c01b1c", "CSR_qhash_d6": "ff92f8f322a1a2d7453b7701fc2ad8ca4dc072576d26893dbd88d3634e380930", "COO_qhash_d6": "ff92f8f322a1a2d7453b7701fc2ad8ca4dc072576d26893dbd88d3634e380930", "ELL_qhash_d6": "cf0819b4494f006861b992b60bee5d5728696919d3bd481e120fdfcd8c9b818c", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.034786, "pilot_csr_per_iter_s": 0.001605, "pilot_coo_per_iter_s": 0.012016, "pilot_ell_per_iter_s": 0.254058, "rolv_build_s": 1.960221, "rolv_iter_s": 0.001653, "dense_iter_s": 0.001922, "csr_iter_s": 0.002096, "coo_iter_s": 0.012182, "ell_iter_s": 0.262926, "rolv_total_s": 3.613027, "baseline_total_s": 1.922236, "speedup_total_vs_selected_x": 0.532, "speedup_iter_vs_selected_x": 1.163, "rolv_vs_vendor_sparse_iter_x": 1.268, "rolv_vs_vendor_sparse_total_x": 0.58, "rolv_vs_coo_iter_x": 7.37, "rolv_vs_coo_total_x": 3.372, "rolv_vs_ell_iter_x": 159.079, "rolv_vs_ell_total_x": 72.795, "correct_norm": "OK"}

[2025-12-10 21:40:16] Platform: CPU | Seed: 123456 | Pattern: random | Zeros: 70% A_hash: edabea164fed9c1e1178aa1e0293e4bcc84954c36ab7990906ee640d4bb92d50 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.036138s | CSR: 0.065529s | COO: 0.398159s | ELL: 3.007983s Selected baseline: Dense rolv load time (operator build): 1.862905 s rolv per-iter: 0.001724s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 81eceb55ae8bfb75e7b118c979ef60e1cadb59cd2eb9e233310bbe0ba3092278 (Dense) CSR_norm_hash: 9dc20e644f27964e1935c42c8de92d89843117b411f6b362b2561d5c451d13c4 COO_norm_hash: 9dc20e644f27964e1935c42c8de92d89843117b411f6b362b2561d5c451d13c4 ELL_norm_hash: 6a538bb5297391078e1ef3431a60a4f19a19945a4c2fe9c0bfdd69ee2188716b ROLF_norm_hash: 808d5eb8b746a43e7efaa1d92fb834853dd22e73ce29f8ae6822338d7b0d8eb1 DENGS_norm_hash: 81eceb55ae8bfb75e7b118c979ef60e1cadb59cd2eb9e233310bbe0ba3092278 COO per-iter: 0.398705s | total: 398.705114s CSR per-iter: 0.069941s | total: 69.941014s ELL per-iter: 3.044799s | total: 3044.977996s ROLF per-iter: 0.000856s | total: 0.858799s DENGS per-iter: 0.037957s | total: 37.956752s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 10.61x (≈ 961% faster) Speedup (per-iter): 22.07x (≈ 2107% faster) Energy Savings (proxy): 95.47% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 40.57x | total: 19.50x rolv vs COO: Speedup (per-iter): 231.26x | total: 111.15x rolv vs ELL: Speedup (per-iter): 1766.08x | total: 848.91x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "edabea164fed9c1e1178aa1e0293e4bcc84954c36ab7990906ee640d4bb92d50", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "81eceb55ae8bfb75e7b118c979ef60e1cadb59cd2eb9e233310bbe0ba3092278", "CSR_norm_hash": "9dc20e644f27964e1935c42c8de92d89843117b411f6b362b2561d5c451d13c4", "COO_norm_hash": "9dc20e644f27964e1935c42c8de92d89843117b411f6b362b2561d5c451d13c4", "ELL_norm_hash": "6a538bb5297391078e1ef3431a60a4f19a19945a4c2fe9c0bfdd69ee2188716b", "ROLF_norm_hash": "808d5eb8b746a43e7efaa1d92fb834853dd22e73ce29f8ae6822338d7b0d8eb1", "DENGS_norm_hash": "81eceb55ae8bfb75e7b118c979ef60e1cadb59cd2eb9e233310bbe0ba3092278", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "3a59023a2d2ba7726bca0ac8cb2b9e7353edcc7076c3e43252c5396c09126d0b", "CSR_qhash_d6": "74636cc22761db600f6cc0bd37ef1387d79fc63d55eaf06ef0e3ef35878badda", "COO_qhash_d6": "74636cc22761db600f6cc0bd37ef1387d79fc63d55eaf06ef0e3ef35878badda", "ELL_qhash_d6": "16d9568cc1df56ed73fe2f5769c182732bfbe7131cdbc998f49bb3b71eeacd5b", "path_selected": "Dense", "pilot_dense_per_iter_s": 0.036138, "pilot_csr_per_iter_s": 0.065529, "pilot_coo_per_iter_s": 0.398159, "pilot_ell_per_iter_s": 3.007983, "rolv_build_s": 1.862905, "rolv_iter_s": 0.001724, "dense_iter_s": 0.038047, "csr_iter_s": 0.069941, "coo_iter_s": 0.398705, "ell_iter_s": 3.044799, "rolv_total_s": 3.586944, "baseline_total_s": 38.046631, "speedup_total_vs_selected_x": 10.607, "speedup_iter_vs_selected_x": 22.068, "rolv_vs_vendor_sparse_iter_x": 40.568, "rolv_vs_vendor_sparse_total_x": 19.499, "rolv_vs_coo_iter_x": 231.262, "rolv_vs_coo_total_x": 111.155, "rolv_vs_ell_iter_x": 1766.084, "rolv_vs_ell_total_x": 848.906, "correct_norm": "OK"}

[2025-12-10 22:41:12] Platform: CPU | Seed: 123456 | Pattern: power_law | Zeros: 70% A_hash: a4914aa6a0fd4d548a01e89fee20c5e0a228e5cbc020b46e550e7bb587fccfdb | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.037707s | CSR: 0.066399s | COO: 0.391915s | ELL: 3.017664s Selected baseline: Dense rolv load time (operator build): 1.812175 s rolv per-iter: 0.001947s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 1b2e15c51a0e7ab8e482773c1ec50cb87a2f70eb299d1d75bbccb3c3cd4cfd04 (Dense) CSR_norm_hash: c7faf12efecb0973b2573884ea7ecb8a551f28c30c57076ad8130df62c62486d COO_norm_hash: c7faf12efecb0973b2573884ea7ecb8a551f28c30c57076ad8130df62c62486d ELL_norm_hash: c9466c5154fc30292548e58667379c7a88d5104a16157352b0bd4ac0455f25cc ROLF_norm_hash: 1683e5ec7652bc89c167ebb3daa3fda66c4d71f35bc686fc1eaf2a27e7309ef8 DENGS_norm_hash: 1b2e15c51a0e7ab8e482773c1ec50cb87a2f70eb299d1d75bbccb3c3cd4cfd04 COO per-iter: 0.394688s | total: 394.688206s CSR per-iter: 0.069304s | total: 69.303541s ELL per-iter: 3.041862s | total: 3042.074367s ROLF per-iter: 0.001009s | total: 1.011253s DENGS per-iter: 0.040543s | total: 40.543496s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 10.54x (≈ 954% faster) Speedup (per-iter): 20.36x (≈ 1936% faster) Energy Savings (proxy): 95.09% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 35.60x | total: 18.44x rolv vs COO: Speedup (per-iter): 202.73x | total: 105.00x rolv vs ELL: Speedup (per-iter): 1562.43x | total: 809.27x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "a4914aa6a0fd4d548a01e89fee20c5e0a228e5cbc020b46e550e7bb587fccfdb", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "1b2e15c51a0e7ab8e482773c1ec50cb87a2f70eb299d1d75bbccb3c3cd4cfd04", "CSR_norm_hash": "c7faf12efecb0973b2573884ea7ecb8a551f28c30c57076ad8130df62c62486d", "COO_norm_hash": "c7faf12efecb0973b2573884ea7ecb8a551f28c30c57076ad8130df62c62486d", "ELL_norm_hash": "c9466c5154fc30292548e58667379c7a88d5104a16157352b0bd4ac0455f25cc", "ROLF_norm_hash": "1683e5ec7652bc89c167ebb3daa3fda66c4d71f35bc686fc1eaf2a27e7309ef8", "DENGS_norm_hash": "1b2e15c51a0e7ab8e482773c1ec50cb87a2f70eb299d1d75bbccb3c3cd4cfd04", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "79d1a70edaac5d6ef6f73feffd9425f525509ce79ac60372330604c75a4afaff", "CSR_qhash_d6": "bc6888810068bff881791707aeafee6db4c61cf4c4a09867ecda8a488006c6f9", "COO_qhash_d6": "bc6888810068bff881791707aeafee6db4c61cf4c4a09867ecda8a488006c6f9", "ELL_qhash_d6": "a286636e67061d3372a4eb9c64e38b94c43540180012e34128fca85adb27d840", "path_selected": "Dense", "pilot_dense_per_iter_s": 0.037707, "pilot_csr_per_iter_s": 0.066399, "pilot_coo_per_iter_s": 0.391915, "pilot_ell_per_iter_s": 3.017664, "rolv_build_s": 1.812175, "rolv_iter_s": 0.001947, "dense_iter_s": 0.039636, "csr_iter_s": 0.069304, "coo_iter_s": 0.394688, "ell_iter_s": 3.041862, "rolv_total_s": 3.759055, "baseline_total_s": 39.635781, "speedup_total_vs_selected_x": 10.544, "speedup_iter_vs_selected_x": 20.359, "rolv_vs_vendor_sparse_iter_x": 35.597, "rolv_vs_vendor_sparse_total_x": 18.436, "rolv_vs_coo_iter_x": 202.729, "rolv_vs_coo_total_x": 104.997, "rolv_vs_ell_iter_x": 1562.429, "rolv_vs_ell_total_x": 809.266, "correct_norm": "OK"}

[2025-12-10 23:42:06] Platform: CPU | Seed: 123456 | Pattern: banded | Zeros: 70% A_hash: 0142a2e7b61a825de01179219be78340bef67ab3e22843242724b58a667c3d7f | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.035884s | CSR: 0.002106s | COO: 0.014624s | ELL: 0.185143s Selected baseline: CSR rolv load time (operator build): 1.907379 s rolv per-iter: 0.001867s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 1a4edf8f25b3dd81e1e214d3690886bf82cb359e1db1f06a41be1392543d3ea9 (CSR) CSR_norm_hash: 1a4edf8f25b3dd81e1e214d3690886bf82cb359e1db1f06a41be1392543d3ea9 COO_norm_hash: 1a4edf8f25b3dd81e1e214d3690886bf82cb359e1db1f06a41be1392543d3ea9 ELL_norm_hash: 63a24442baef8f0d1a4164efe8160591bb92101337acbb48d6babf7b58df3569 ROLF_norm_hash: 233c3eadf2af88bcc9fda48121ccb506f37e063352521220816b001b5b1214fb DENGS_norm_hash: d66d537ba293b7dac5833c5c0926bd25ea79efc417fe54826ba3dc1dcaa0d74e COO per-iter: 0.014683s | total: 14.682682s CSR per-iter: 0.002168s | total: 2.168262s ELL per-iter: 0.193311s | total: 193.421761s ROLF per-iter: 0.000841s | total: 0.844324s DENGS per-iter: 0.038640s | total: 38.639931s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 0.66x (≈ -34% faster) Speedup (per-iter): 1.33x (≈ 33% faster) Energy Savings (proxy): 24.93% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 1.16x | total: 0.57x rolv vs COO: Speedup (per-iter): 7.87x | total: 3.89x rolv vs ELL: Speedup (per-iter): 103.56x | total: 51.25x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "0142a2e7b61a825de01179219be78340bef67ab3e22843242724b58a667c3d7f", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "d66d537ba293b7dac5833c5c0926bd25ea79efc417fe54826ba3dc1dcaa0d74e", "CSR_norm_hash": "1a4edf8f25b3dd81e1e214d3690886bf82cb359e1db1f06a41be1392543d3ea9", "COO_norm_hash": "1a4edf8f25b3dd81e1e214d3690886bf82cb359e1db1f06a41be1392543d3ea9", "ELL_norm_hash": "63a24442baef8f0d1a4164efe8160591bb92101337acbb48d6babf7b58df3569", "ROLF_norm_hash": "233c3eadf2af88bcc9fda48121ccb506f37e063352521220816b001b5b1214fb", "DENGS_norm_hash": "d66d537ba293b7dac5833c5c0926bd25ea79efc417fe54826ba3dc1dcaa0d74e", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "842fd33e8b1c9ddafe0ecd64b06375d9ed44df1054132e4d9eb85ba67a390caf", "CSR_qhash_d6": "e0fdfdfe9a7980992a5d95b68df283dca5850133880ba2b789cb0d57a86784d3", "COO_qhash_d6": "e0fdfdfe9a7980992a5d95b68df283dca5850133880ba2b789cb0d57a86784d3", "ELL_qhash_d6": "c778020152b6b1968044f165d1d7768a10098508c1a59eee96f2c605f97efeca", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.035884, "pilot_csr_per_iter_s": 0.002106, "pilot_coo_per_iter_s": 0.014624, "pilot_ell_per_iter_s": 0.185143, "rolv_build_s": 1.907379, "rolv_iter_s": 0.001867, "dense_iter_s": 0.002487, "csr_iter_s": 0.002168, "coo_iter_s": 0.014683, "ell_iter_s": 0.193311, "rolv_total_s": 3.77402, "baseline_total_s": 2.486593, "speedup_total_vs_selected_x": 0.659, "speedup_iter_vs_selected_x": 1.332, "rolv_vs_vendor_sparse_iter_x": 1.162, "rolv_vs_vendor_sparse_total_x": 0.575, "rolv_vs_coo_iter_x": 7.866, "rolv_vs_coo_total_x": 3.89, "rolv_vs_ell_iter_x": 103.561, "rolv_vs_ell_total_x": 51.251, "correct_norm": "OK"}

[2025-12-10 23:46:27] Platform: CPU | Seed: 123456 | Pattern: block_diagonal | Zeros: 70% A_hash: cef347a1932b5f8840dbe11528f254d128cfbf946eff13f63d476f616562814f | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.036889s | CSR: 0.001489s | COO: 0.009333s | ELL: 0.208148s Selected baseline: CSR rolv load time (operator build): 2.103125 s rolv per-iter: 0.001993s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 31a81000986fcbef203106563eebfcd421a73b54e425d1209e4eb7fcd22c7ad7 (CSR) CSR_norm_hash: 31a81000986fcbef203106563eebfcd421a73b54e425d1209e4eb7fcd22c7ad7 COO_norm_hash: 31a81000986fcbef203106563eebfcd421a73b54e425d1209e4eb7fcd22c7ad7 ELL_norm_hash: d34042bcaaf38c5c0fe6de1324e9746b4865a262f32cc1e90a02bcb93edc8be1 ROLF_norm_hash: cdc776a6e7bb392ba1b8bba930b1be11f481bf8b6edd6a912a383227e348f08e DENGS_norm_hash: cd653bd2e004650b100d9d467530441799b3da430170d0f492b6da7cf45e4a56 COO per-iter: 0.009339s | total: 9.338607s CSR per-iter: 0.001552s | total: 1.552308s ELL per-iter: 0.215756s | total: 215.835280s ROLF per-iter: 0.001023s | total: 1.025416s DENGS per-iter: 0.038369s | total: 38.368925s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 0.33x (≈ -67% faster) Speedup (per-iter): 0.67x (≈ -33% faster) Energy Savings (proxy): -48.47% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 0.78x | total: 0.38x rolv vs COO: Speedup (per-iter): 4.69x | total: 2.28x rolv vs ELL: Speedup (per-iter): 108.27x | total: 52.69x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "cef347a1932b5f8840dbe11528f254d128cfbf946eff13f63d476f616562814f", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "cd653bd2e004650b100d9d467530441799b3da430170d0f492b6da7cf45e4a56", "CSR_norm_hash": "31a81000986fcbef203106563eebfcd421a73b54e425d1209e4eb7fcd22c7ad7", "COO_norm_hash": "31a81000986fcbef203106563eebfcd421a73b54e425d1209e4eb7fcd22c7ad7", "ELL_norm_hash": "d34042bcaaf38c5c0fe6de1324e9746b4865a262f32cc1e90a02bcb93edc8be1", "ROLF_norm_hash": "cdc776a6e7bb392ba1b8bba930b1be11f481bf8b6edd6a912a383227e348f08e", "DENGS_norm_hash": "cd653bd2e004650b100d9d467530441799b3da430170d0f492b6da7cf45e4a56", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "9bb27c5182fa13d42b2537be16bef031ce38c9bfd32341ff91d3612d6c7b6cda", "CSR_qhash_d6": "1b8e30d0b60f1ad7b2c21d8a32ce36eef0b5957ed8ec3553553e2939a70d13db", "COO_qhash_d6": "1b8e30d0b60f1ad7b2c21d8a32ce36eef0b5957ed8ec3553553e2939a70d13db", "ELL_qhash_d6": "e555f99d5862c321c3039efe9e8930994e00daeef24a73b62d22d21891ccd3ee", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.036889, "pilot_csr_per_iter_s": 0.001489, "pilot_coo_per_iter_s": 0.009333, "pilot_ell_per_iter_s": 0.208148, "rolv_build_s": 2.103125, "rolv_iter_s": 0.001993, "dense_iter_s": 0.001342, "csr_iter_s": 0.001552, "coo_iter_s": 0.009339, "ell_iter_s": 0.215756, "rolv_total_s": 4.095952, "baseline_total_s": 1.34226, "speedup_total_vs_selected_x": 0.328, "speedup_iter_vs_selected_x": 0.674, "rolv_vs_vendor_sparse_iter_x": 0.779, "rolv_vs_vendor_sparse_total_x": 0.379, "rolv_vs_coo_iter_x": 4.686, "rolv_vs_coo_total_x": 2.28, "rolv_vs_ell_iter_x": 108.266, "rolv_vs_ell_total_x": 52.695, "correct_norm": "OK"}

[2025-12-10 23:51:04] Platform: CPU | Seed: 123456 | Pattern: random | Zeros: 80% A_hash: 64d493c0c3be5476d6605162bb5937a91fb2c1c64fa183deec0963643619e481 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.034608s | CSR: 0.042507s | COO: 0.264706s | ELL: 2.054948s Selected baseline: Dense rolv load time (operator build): 1.810927 s rolv per-iter: 0.001604s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 049b150f107b3625d10d60482afe96242869aa4dcccbc26c9111e924b6280b9b (Dense) CSR_norm_hash: 8f6b12d7ba925e7ef265fadecfd402c4dfce7716bfae19a8fd2b673ea63d48a4 COO_norm_hash: 8f6b12d7ba925e7ef265fadecfd402c4dfce7716bfae19a8fd2b673ea63d48a4 ELL_norm_hash: 9598958dbbe77164ee5c407dc547358b3f6620fc72413cf346c621b96b01fd64 ROLF_norm_hash: 96769f6eed3bdf83cd4dbc5f43e95096e26005cf029666713f027b9618bf3905 DENGS_norm_hash: 049b150f107b3625d10d60482afe96242869aa4dcccbc26c9111e924b6280b9b COO per-iter: 0.269996s | total: 269.995572s CSR per-iter: 0.046445s | total: 46.444693s ELL per-iter: 2.117732s | total: 2117.896411s ROLF per-iter: 0.001290s | total: 1.291801s DENGS per-iter: 0.038077s | total: 38.077365s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 12.18x (≈ 1118% faster) Speedup (per-iter): 25.94x (≈ 2494% faster) Energy Savings (proxy): 96.15% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 28.96x | total: 13.60x rolv vs COO: Speedup (per-iter): 168.37x | total: 79.07x rolv vs ELL: Speedup (per-iter): 1320.62x | total: 620.26x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "64d493c0c3be5476d6605162bb5937a91fb2c1c64fa183deec0963643619e481", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "049b150f107b3625d10d60482afe96242869aa4dcccbc26c9111e924b6280b9b", "CSR_norm_hash": "8f6b12d7ba925e7ef265fadecfd402c4dfce7716bfae19a8fd2b673ea63d48a4", "COO_norm_hash": "8f6b12d7ba925e7ef265fadecfd402c4dfce7716bfae19a8fd2b673ea63d48a4", "ELL_norm_hash": "9598958dbbe77164ee5c407dc547358b3f6620fc72413cf346c621b96b01fd64", "ROLF_norm_hash": "96769f6eed3bdf83cd4dbc5f43e95096e26005cf029666713f027b9618bf3905", "DENGS_norm_hash": "049b150f107b3625d10d60482afe96242869aa4dcccbc26c9111e924b6280b9b", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "be6ed676254440a932e5913581e3f60e08a8b88d8f6e290bff13fca2c1296ddd", "CSR_qhash_d6": "746e670f3286caeaf15cae83924effa66d6207ec49d476376de0a642b651c075", "COO_qhash_d6": "746e670f3286caeaf15cae83924effa66d6207ec49d476376de0a642b651c075", "ELL_qhash_d6": "606d92dc7f0acd8cdfb209102e656770d3cab9199142e8a8cf65898c7d766388", "path_selected": "Dense", "pilot_dense_per_iter_s": 0.034608, "pilot_csr_per_iter_s": 0.042507, "pilot_coo_per_iter_s": 0.264706, "pilot_ell_per_iter_s": 2.054948, "rolv_build_s": 1.810927, "rolv_iter_s": 0.001604, "dense_iter_s": 0.041601, "csr_iter_s": 0.046445, "coo_iter_s": 0.269996, "ell_iter_s": 2.117732, "rolv_total_s": 3.414522, "baseline_total_s": 41.601057, "speedup_total_vs_selected_x": 12.184, "speedup_iter_vs_selected_x": 25.942, "rolv_vs_vendor_sparse_iter_x": 28.963, "rolv_vs_vendor_sparse_total_x": 13.602, "rolv_vs_coo_iter_x": 168.369, "rolv_vs_coo_total_x": 79.073, "rolv_vs_ell_iter_x": 1320.615, "rolv_vs_ell_total_x": 620.261, "correct_norm": "OK"}

[2025-12-11 00:33:46] Platform: CPU | Seed: 123456 | Pattern: power_law | Zeros: 80% A_hash: 42d035943092f2f2847c69c7c20811f6a924622ae69da2990a7898aaa8812ad3 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.035103s | CSR: 0.046372s | COO: 0.271033s | ELL: 2.083610s Selected baseline: Dense rolv load time (operator build): 1.967291 s rolv per-iter: 0.001925s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 0b23c42b75a0a35a60734f4306d24f3c39615dd397c2a8e1ec4154b6547e597b (Dense) CSR_norm_hash: 4ee0186456af87a860f602bce3ee03add2bdf4b6bb3938714a8a88094dbb7796 COO_norm_hash: 4ee0186456af87a860f602bce3ee03add2bdf4b6bb3938714a8a88094dbb7796 ELL_norm_hash: 845c905f1c9b985da0966d3c86b859328c528f9dbb6e86234de48466b246454b ROLF_norm_hash: 085beb21928e66e7fb48102473f274dfe13f3ecf58657eda4c20510b737c66c8 DENGS_norm_hash: 0b23c42b75a0a35a60734f4306d24f3c39615dd397c2a8e1ec4154b6547e597b COO per-iter: 0.270358s | total: 270.357748s CSR per-iter: 0.046625s | total: 46.625181s ELL per-iter: 2.144520s | total: 2144.689446s ROLF per-iter: 0.000904s | total: 0.906390s DENGS per-iter: 0.039872s | total: 39.872430s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 10.09x (≈ 909% faster) Speedup (per-iter): 20.41x (≈ 1941% faster) Energy Savings (proxy): 95.10% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 24.22x | total: 11.98x rolv vs COO: Speedup (per-iter): 140.45x | total: 69.46x rolv vs ELL: Speedup (per-iter): 1114.06x | total: 551.02x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "42d035943092f2f2847c69c7c20811f6a924622ae69da2990a7898aaa8812ad3", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "0b23c42b75a0a35a60734f4306d24f3c39615dd397c2a8e1ec4154b6547e597b", "CSR_norm_hash": "4ee0186456af87a860f602bce3ee03add2bdf4b6bb3938714a8a88094dbb7796", "COO_norm_hash": "4ee0186456af87a860f602bce3ee03add2bdf4b6bb3938714a8a88094dbb7796", "ELL_norm_hash": "845c905f1c9b985da0966d3c86b859328c528f9dbb6e86234de48466b246454b", "ROLF_norm_hash": "085beb21928e66e7fb48102473f274dfe13f3ecf58657eda4c20510b737c66c8", "DENGS_norm_hash": "0b23c42b75a0a35a60734f4306d24f3c39615dd397c2a8e1ec4154b6547e597b", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "e8097442ba22651f533a2e3f6d86f12b0a33993afaac63a1611755e2903f4ffa", "CSR_qhash_d6": "5aeca76a01894acf5ef8e3f20e7ae624429159cc7505010dda7934475acc9ed9", "COO_qhash_d6": "5aeca76a01894acf5ef8e3f20e7ae624429159cc7505010dda7934475acc9ed9", "ELL_qhash_d6": "cde2ad186dd00c47b17021b935538d889c9a53c7404dca74642c843ab1723547", "path_selected": "Dense", "pilot_dense_per_iter_s": 0.035103, "pilot_csr_per_iter_s": 0.046372, "pilot_coo_per_iter_s": 0.271033, "pilot_ell_per_iter_s": 2.08361, "rolv_build_s": 1.967291, "rolv_iter_s": 0.001925, "dense_iter_s": 0.039282, "csr_iter_s": 0.046625, "coo_iter_s": 0.270358, "ell_iter_s": 2.14452, "rolv_total_s": 3.89225, "baseline_total_s": 39.281855, "speedup_total_vs_selected_x": 10.092, "speedup_iter_vs_selected_x": 20.407, "rolv_vs_vendor_sparse_iter_x": 24.221, "rolv_vs_vendor_sparse_total_x": 11.979, "rolv_vs_coo_iter_x": 140.449, "rolv_vs_coo_total_x": 69.461, "rolv_vs_ell_iter_x": 1114.06, "rolv_vs_ell_total_x": 551.015, "correct_norm": "OK"}

[2025-12-11 01:16:56] Platform: CPU | Seed: 123456 | Pattern: banded | Zeros: 80% A_hash: 67876d2abc872eda99ab2583102936d618e98bfb5f5623abfaed19da8d34f428 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.038339s | CSR: 0.001505s | COO: 0.010380s | ELL: 0.143286s Selected baseline: CSR rolv load time (operator build): 1.903425 s rolv per-iter: 0.001926s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: d69b3d693eeebf5754efcefa03d3c077c36867fdc0f690a5dfb83caf283399d3 (CSR) CSR_norm_hash: d69b3d693eeebf5754efcefa03d3c077c36867fdc0f690a5dfb83caf283399d3 COO_norm_hash: d69b3d693eeebf5754efcefa03d3c077c36867fdc0f690a5dfb83caf283399d3 ELL_norm_hash: 30cc5fb7c2c8809df850efe659d5f3087e74ddb4d099adc98fd6aa66e1700553 ROLF_norm_hash: 76495886fe7922a17f5e48d146499330a23ea30ea85bafc1411390f5f82eb477 DENGS_norm_hash: 7c7f6bd471e797752c4bc8a7fc6826c85288d946730037f63c526dcfef5e7825 COO per-iter: 0.010446s | total: 10.446254s CSR per-iter: 0.001715s | total: 1.714760s ELL per-iter: 0.147828s | total: 147.936351s ROLF per-iter: 0.000891s | total: 0.894225s DENGS per-iter: 0.040074s | total: 40.073549s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 0.55x (≈ -45% faster) Speedup (per-iter): 1.10x (≈ 10% faster) Energy Savings (proxy): 9.01% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 0.89x | total: 0.45x rolv vs COO: Speedup (per-iter): 5.42x | total: 2.73x rolv vs ELL: Speedup (per-iter): 76.76x | total: 38.63x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "67876d2abc872eda99ab2583102936d618e98bfb5f5623abfaed19da8d34f428", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "7c7f6bd471e797752c4bc8a7fc6826c85288d946730037f63c526dcfef5e7825", "CSR_norm_hash": "d69b3d693eeebf5754efcefa03d3c077c36867fdc0f690a5dfb83caf283399d3", "COO_norm_hash": "d69b3d693eeebf5754efcefa03d3c077c36867fdc0f690a5dfb83caf283399d3", "ELL_norm_hash": "30cc5fb7c2c8809df850efe659d5f3087e74ddb4d099adc98fd6aa66e1700553", "ROLF_norm_hash": "76495886fe7922a17f5e48d146499330a23ea30ea85bafc1411390f5f82eb477", "DENGS_norm_hash": "7c7f6bd471e797752c4bc8a7fc6826c85288d946730037f63c526dcfef5e7825", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "aeac2c2f7ad80148d3e91e02d26b98b9c36098a11f5f199a6848db5957c8bf7f", "CSR_qhash_d6": "fefdf97b41809eef28a1f532481d4687c6c3c2548edede4ea7f67b2731546087", "COO_qhash_d6": "fefdf97b41809eef28a1f532481d4687c6c3c2548edede4ea7f67b2731546087", "ELL_qhash_d6": "7409a17dc5a5e059ffa4b47400be0f54418d5554284d3eb519d8ddd478f74b38", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.038339, "pilot_csr_per_iter_s": 0.001505, "pilot_coo_per_iter_s": 0.01038, "pilot_ell_per_iter_s": 0.143286, "rolv_build_s": 1.903425, "rolv_iter_s": 0.001926, "dense_iter_s": 0.002117, "csr_iter_s": 0.001715, "coo_iter_s": 0.010446, "ell_iter_s": 0.147828, "rolv_total_s": 3.829318, "baseline_total_s": 2.116628, "speedup_total_vs_selected_x": 0.553, "speedup_iter_vs_selected_x": 1.099, "rolv_vs_vendor_sparse_iter_x": 0.89, "rolv_vs_vendor_sparse_total_x": 0.448, "rolv_vs_coo_iter_x": 5.424, "rolv_vs_coo_total_x": 2.728, "rolv_vs_ell_iter_x": 76.758, "rolv_vs_ell_total_x": 38.633, "correct_norm": "OK"}

[2025-12-11 01:20:27] Platform: CPU | Seed: 123456 | Pattern: block_diagonal | Zeros: 80% A_hash: ecc44a84722a2b4a5ffee2c3bbc876333726f132e5bf939507803944fd1eec24 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.058435s | CSR: 0.001807s | COO: 0.007146s | ELL: 0.166938s Selected baseline: CSR rolv load time (operator build): 2.272878 s rolv per-iter: 0.002375s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: c5bac0f924edc6e06a1df9c0bca371ded85a5061eb55ffbb16f1abebd7ada66b (CSR) CSR_norm_hash: c5bac0f924edc6e06a1df9c0bca371ded85a5061eb55ffbb16f1abebd7ada66b COO_norm_hash: c5bac0f924edc6e06a1df9c0bca371ded85a5061eb55ffbb16f1abebd7ada66b ELL_norm_hash: 8d33815dac7b7ddeb2373e0737b2c8050b80884834f7ba69d17d941e4063c6aa ROLF_norm_hash: 57fb5eb0d85d5f30a4401e4a0dbe30473ee15299477e383b2462059a5c34c4ee DENGS_norm_hash: e6a781e6e87ded4657d1b5126e01e30cfc78064d923336aeb09da4464205a830 COO per-iter: 0.006767s | total: 6.766916s CSR per-iter: 0.001598s | total: 1.598362s ELL per-iter: 0.176573s | total: 176.679314s ROLF per-iter: 0.000869s | total: 0.872087s DENGS per-iter: 0.038935s | total: 38.934685s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 0.33x (≈ -67% faster) Speedup (per-iter): 0.65x (≈ -35% faster) Energy Savings (proxy): -52.70% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 0.67x | total: 0.34x rolv vs COO: Speedup (per-iter): 2.85x | total: 1.46x rolv vs ELL: Speedup (per-iter): 74.36x | total: 38.02x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "ecc44a84722a2b4a5ffee2c3bbc876333726f132e5bf939507803944fd1eec24", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "e6a781e6e87ded4657d1b5126e01e30cfc78064d923336aeb09da4464205a830", "CSR_norm_hash": "c5bac0f924edc6e06a1df9c0bca371ded85a5061eb55ffbb16f1abebd7ada66b", "COO_norm_hash": "c5bac0f924edc6e06a1df9c0bca371ded85a5061eb55ffbb16f1abebd7ada66b", "ELL_norm_hash": "8d33815dac7b7ddeb2373e0737b2c8050b80884834f7ba69d17d941e4063c6aa", "ROLF_norm_hash": "57fb5eb0d85d5f30a4401e4a0dbe30473ee15299477e383b2462059a5c34c4ee", "DENGS_norm_hash": "e6a781e6e87ded4657d1b5126e01e30cfc78064d923336aeb09da4464205a830", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "68b8aae6924856cd57f64c756f75e860b191fff4ac395fdadbb82996fa1d9162", "CSR_qhash_d6": "e702e186b655acbfac9d2bfdd8cda8e3cf2e633fd652b3b046a54b11e848aab6", "COO_qhash_d6": "e702e186b655acbfac9d2bfdd8cda8e3cf2e633fd652b3b046a54b11e848aab6", "ELL_qhash_d6": "9048cd282379a77b983678ad7cd0b5096cfc7074e8bd86a860fac25d94de2b38", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.058435, "pilot_csr_per_iter_s": 0.001807, "pilot_coo_per_iter_s": 0.007146, "pilot_ell_per_iter_s": 0.166938, "rolv_build_s": 2.272878, "rolv_iter_s": 0.002375, "dense_iter_s": 0.001555, "csr_iter_s": 0.001598, "coo_iter_s": 0.006767, "ell_iter_s": 0.176573, "rolv_total_s": 4.647525, "baseline_total_s": 1.555073, "speedup_total_vs_selected_x": 0.335, "speedup_iter_vs_selected_x": 0.655, "rolv_vs_vendor_sparse_iter_x": 0.673, "rolv_vs_vendor_sparse_total_x": 0.344, "rolv_vs_coo_iter_x": 2.85, "rolv_vs_coo_total_x": 1.456, "rolv_vs_ell_iter_x": 74.358, "rolv_vs_ell_total_x": 38.016, "correct_norm": "OK"}

[2025-12-11 01:24:22] Platform: CPU | Seed: 123456 | Pattern: random | Zeros: 90% A_hash: 7c426951bcc4d163ce47e51853f129c1cf264a69a73f0ec5fc2ffa57e17c857e | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.038684s | CSR: 0.024451s | COO: 0.144540s | ELL: 1.111894s Selected baseline: CSR rolv load time (operator build): 2.281104 s rolv per-iter: 0.001813s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 0851215af1500ecf790b8c2e7998e41b57d629c3ad3d9b645190db8d3a7a98a9 (CSR) CSR_norm_hash: 0851215af1500ecf790b8c2e7998e41b57d629c3ad3d9b645190db8d3a7a98a9 COO_norm_hash: 0851215af1500ecf790b8c2e7998e41b57d629c3ad3d9b645190db8d3a7a98a9 ELL_norm_hash: e54f1a15e80bd104e81fc35db860065b6a84bfd93fe9a2bab8f9d1098e099b99 ROLF_norm_hash: f87e2a755b73c93caac91e924863a7faba34ee5295260c08a6f80152def35b23 DENGS_norm_hash: e8f1df1bb63d5ee4c0fd781fcabd9157504f8bb2b2b5e3fac4e6f6b9e0fa5878 COO per-iter: 0.141165s | total: 141.164571s CSR per-iter: 0.024928s | total: 24.927500s ELL per-iter: 1.245623s | total: 1245.777753s ROLF per-iter: 0.001394s | total: 1.397057s DENGS per-iter: 0.038490s | total: 38.490441s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 5.92x (≈ 492% faster) Speedup (per-iter): 13.36x (≈ 1236% faster) Energy Savings (proxy): 92.51% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 13.75x | total: 6.09x rolv vs COO: Speedup (per-iter): 77.86x | total: 34.48x rolv vs ELL: Speedup (per-iter): 687.01x | total: 304.28x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "7c426951bcc4d163ce47e51853f129c1cf264a69a73f0ec5fc2ffa57e17c857e", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "e8f1df1bb63d5ee4c0fd781fcabd9157504f8bb2b2b5e3fac4e6f6b9e0fa5878", "CSR_norm_hash": "0851215af1500ecf790b8c2e7998e41b57d629c3ad3d9b645190db8d3a7a98a9", "COO_norm_hash": "0851215af1500ecf790b8c2e7998e41b57d629c3ad3d9b645190db8d3a7a98a9", "ELL_norm_hash": "e54f1a15e80bd104e81fc35db860065b6a84bfd93fe9a2bab8f9d1098e099b99", "ROLF_norm_hash": "f87e2a755b73c93caac91e924863a7faba34ee5295260c08a6f80152def35b23", "DENGS_norm_hash": "e8f1df1bb63d5ee4c0fd781fcabd9157504f8bb2b2b5e3fac4e6f6b9e0fa5878", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "b8d073b69eb67461aa5ab484f345d60cbc83e2776a3aa46df32029cfe8bdebc0", "CSR_qhash_d6": "755e41997772a5bab04dd7c9716d65801ff6420b484d5068a6a7c4b204b357eb", "COO_qhash_d6": "755e41997772a5bab04dd7c9716d65801ff6420b484d5068a6a7c4b204b357eb", "ELL_qhash_d6": "24cee67ccb3a12ec1922b7645928e757e6402a3484caeeedfb9eec042ec4e781", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.038684, "pilot_csr_per_iter_s": 0.024451, "pilot_coo_per_iter_s": 0.14454, "pilot_ell_per_iter_s": 1.111894, "rolv_build_s": 2.281104, "rolv_iter_s": 0.001813, "dense_iter_s": 0.024223, "csr_iter_s": 0.024928, "coo_iter_s": 0.141165, "ell_iter_s": 1.245623, "rolv_total_s": 4.09422, "baseline_total_s": 24.222525, "speedup_total_vs_selected_x": 5.916, "speedup_iter_vs_selected_x": 13.36, "rolv_vs_vendor_sparse_iter_x": 13.748, "rolv_vs_vendor_sparse_total_x": 6.088, "rolv_vs_coo_iter_x": 77.857, "rolv_vs_coo_total_x": 34.479, "rolv_vs_ell_iter_x": 687.007, "rolv_vs_ell_total_x": 304.277, "correct_norm": "OK"}

[2025-12-11 01:49:27] Platform: CPU | Seed: 123456 | Pattern: power_law | Zeros: 90% A_hash: 363159a29709d2a96c15d393f4309b3c8e48e6d6869a6e822955a0d1b739efad | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.037001s | CSR: 0.022104s | COO: 0.137292s | ELL: 1.103616s Selected baseline: CSR rolv load time (operator build): 2.668994 s rolv per-iter: 0.001889s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 89236392d28844847058334c61d6aa770b904f0d2cdbde32b4fcf5c323fb8e08 (CSR) CSR_norm_hash: 89236392d28844847058334c61d6aa770b904f0d2cdbde32b4fcf5c323fb8e08 COO_norm_hash: 89236392d28844847058334c61d6aa770b904f0d2cdbde32b4fcf5c323fb8e08 ELL_norm_hash: e43c9f9b77ea5c541fcad3923f331f4737d616ba91cf6ab34b0380f78e92b1e3 ROLF_norm_hash: 206f1e291bf73aeab6a39808ad5eaf6bd8b79540d2553c10f41e0b799ed3825b DENGS_norm_hash: dfcbf6576639308276e4d5b0f1bca06d2eb87a61c6803b6ae8170c43c2fad0f4 COO per-iter: 0.143378s | total: 143.377937s CSR per-iter: 0.024033s | total: 24.032559s ELL per-iter: 1.253734s | total: 1253.882358s ROLF per-iter: 0.001436s | total: 1.439145s DENGS per-iter: 0.039700s | total: 39.699768s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 5.12x (≈ 412% faster) Speedup (per-iter): 12.36x (≈ 1136% faster) Energy Savings (proxy): 91.91% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 12.72x | total: 5.27x rolv vs COO: Speedup (per-iter): 75.91x | total: 31.46x rolv vs ELL: Speedup (per-iter): 663.77x | total: 275.11x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "363159a29709d2a96c15d393f4309b3c8e48e6d6869a6e822955a0d1b739efad", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "dfcbf6576639308276e4d5b0f1bca06d2eb87a61c6803b6ae8170c43c2fad0f4", "CSR_norm_hash": "89236392d28844847058334c61d6aa770b904f0d2cdbde32b4fcf5c323fb8e08", "COO_norm_hash": "89236392d28844847058334c61d6aa770b904f0d2cdbde32b4fcf5c323fb8e08", "ELL_norm_hash": "e43c9f9b77ea5c541fcad3923f331f4737d616ba91cf6ab34b0380f78e92b1e3", "ROLF_norm_hash": "206f1e291bf73aeab6a39808ad5eaf6bd8b79540d2553c10f41e0b799ed3825b", "DENGS_norm_hash": "dfcbf6576639308276e4d5b0f1bca06d2eb87a61c6803b6ae8170c43c2fad0f4", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "16090ba9eb3639e0f954c928053bcd8cdeee3ce352a50e6b3c53c417cec8bda5", "CSR_qhash_d6": "586d86426ef6681a81c37a1e02ae1e5f5a13402384e2ce7b07573ed87773298a", "COO_qhash_d6": "586d86426ef6681a81c37a1e02ae1e5f5a13402384e2ce7b07573ed87773298a", "ELL_qhash_d6": "e17650c5d9caeb77b0614375f3b82109842ddf1fe9b6480359b501cc14118f6f", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.037001, "pilot_csr_per_iter_s": 0.022104, "pilot_coo_per_iter_s": 0.137292, "pilot_ell_per_iter_s": 1.103616, "rolv_build_s": 2.668994, "rolv_iter_s": 0.001889, "dense_iter_s": 0.023354, "csr_iter_s": 0.024033, "coo_iter_s": 0.143378, "ell_iter_s": 1.253734, "rolv_total_s": 4.557788, "baseline_total_s": 23.353931, "speedup_total_vs_selected_x": 5.124, "speedup_iter_vs_selected_x": 12.364, "rolv_vs_vendor_sparse_iter_x": 12.724, "rolv_vs_vendor_sparse_total_x": 5.273, "rolv_vs_coo_iter_x": 75.91, "rolv_vs_coo_total_x": 31.458, "rolv_vs_ell_iter_x": 663.775, "rolv_vs_ell_total_x": 275.108, "correct_norm": "OK"}

[2025-12-11 02:14:40] Platform: CPU | Seed: 123456 | Pattern: banded | Zeros: 90% A_hash: 3a4be2953f56037da686c67d65e1991ea56aeecb64913c6c3cc6e2a2c59a68fa | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.036174s | CSR: 0.000909s | COO: 0.005629s | ELL: 0.080269s Selected baseline: CSR rolv load time (operator build): 2.672456 s rolv per-iter: 0.002085s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 47bddbd72e0af6309562fdeab35c98810b297c3f96495a20734d771cd0072e1c (CSR) CSR_norm_hash: 47bddbd72e0af6309562fdeab35c98810b297c3f96495a20734d771cd0072e1c COO_norm_hash: 47bddbd72e0af6309562fdeab35c98810b297c3f96495a20734d771cd0072e1c ELL_norm_hash: 7ef92ddca8b2b3a2852dbbdca6df6e45f4923229b40dc27a2e56415156b0777a ROLF_norm_hash: 529cfcbc569e7a92f217f055813268750768fc5c0353bc7766370e59eb728ae4 DENGS_norm_hash: 7b84fcd5f1f2a0485061b0f70bb9058986a215693541d4cdfe7994d5e5c400d2 COO per-iter: 0.005766s | total: 5.766394s CSR per-iter: 0.000967s | total: 0.967298s ELL per-iter: 0.095081s | total: 95.186601s ROLF per-iter: 0.000840s | total: 0.843675s DENGS per-iter: 0.040098s | total: 40.097726s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 0.25x (≈ -75% faster) Speedup (per-iter): 0.57x (≈ -43% faster) Energy Savings (proxy): -76.07% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 0.46x | total: 0.20x rolv vs COO: Speedup (per-iter): 2.77x | total: 1.21x rolv vs ELL: Speedup (per-iter): 45.60x | total: 20.01x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "3a4be2953f56037da686c67d65e1991ea56aeecb64913c6c3cc6e2a2c59a68fa", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "7b84fcd5f1f2a0485061b0f70bb9058986a215693541d4cdfe7994d5e5c400d2", "CSR_norm_hash": "47bddbd72e0af6309562fdeab35c98810b297c3f96495a20734d771cd0072e1c", "COO_norm_hash": "47bddbd72e0af6309562fdeab35c98810b297c3f96495a20734d771cd0072e1c", "ELL_norm_hash": "7ef92ddca8b2b3a2852dbbdca6df6e45f4923229b40dc27a2e56415156b0777a", "ROLF_norm_hash": "529cfcbc569e7a92f217f055813268750768fc5c0353bc7766370e59eb728ae4", "DENGS_norm_hash": "7b84fcd5f1f2a0485061b0f70bb9058986a215693541d4cdfe7994d5e5c400d2", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "de59863adf4626c23120f18433a82f13813c44b36d4ee26f4b2005b728f8b18a", "CSR_qhash_d6": "f43230edf62f64d3a8c98d76d6ecbb1c93d1886e36838686a8c37e605dfbfd88", "COO_qhash_d6": "f43230edf62f64d3a8c98d76d6ecbb1c93d1886e36838686a8c37e605dfbfd88", "ELL_qhash_d6": "56cf24faaf29e426ae0b51e13ff574fea3404d899110973a60462faac26170c2", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.036174, "pilot_csr_per_iter_s": 0.000909, "pilot_coo_per_iter_s": 0.005629, "pilot_ell_per_iter_s": 0.080269, "rolv_build_s": 2.672456, "rolv_iter_s": 0.002085, "dense_iter_s": 0.001184, "csr_iter_s": 0.000967, "coo_iter_s": 0.005766, "ell_iter_s": 0.095081, "rolv_total_s": 4.757562, "baseline_total_s": 1.18427, "speedup_total_vs_selected_x": 0.249, "speedup_iter_vs_selected_x": 0.568, "rolv_vs_vendor_sparse_iter_x": 0.464, "rolv_vs_vendor_sparse_total_x": 0.203, "rolv_vs_coo_iter_x": 2.766, "rolv_vs_coo_total_x": 1.212, "rolv_vs_ell_iter_x": 45.6, "rolv_vs_ell_total_x": 20.007, "correct_norm": "OK"}

[2025-12-11 02:17:12] Platform: CPU | Seed: 123456 | Pattern: block_diagonal | Zeros: 90% A_hash: 0171048603d1bc057942ef21a0ca65c5bfa3cce5516b26a2e7c8a97b4605b9d5 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.068906s | CSR: 0.001686s | COO: 0.004624s | ELL: 0.100975s Selected baseline: CSR rolv load time (operator build): 2.114612 s rolv per-iter: 0.002074s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: db4d6d5f892425f5b1a9fb6404df67a67145723b6b2413286069a03de72c9ec8 (CSR) CSR_norm_hash: db4d6d5f892425f5b1a9fb6404df67a67145723b6b2413286069a03de72c9ec8 COO_norm_hash: db4d6d5f892425f5b1a9fb6404df67a67145723b6b2413286069a03de72c9ec8 ELL_norm_hash: 00d6955588c1bd5df84ed3bee12ee54f20a45e2f70900b833bbeaaa1d24b408a ROLF_norm_hash: 8ed88523effeac3f818b45ad493b8510ed902a927c187e58593ccb88e5990120 DENGS_norm_hash: 84fe8696deb8cad6ea47af85a72b3f1a329393a02cd1871783547b8ef32893e1 COO per-iter: 0.004075s | total: 4.075007s CSR per-iter: 0.000963s | total: 0.963378s ELL per-iter: 0.104886s | total: 105.004027s ROLF per-iter: 0.000860s | total: 0.862514s DENGS per-iter: 0.039361s | total: 39.360587s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 0.23x (≈ -77% faster) Speedup (per-iter): 0.46x (≈ -54% faster) Energy Savings (proxy): -117.94% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 0.46x | total: 0.23x rolv vs COO: Speedup (per-iter): 1.96x | total: 0.97x rolv vs ELL: Speedup (per-iter): 50.56x | total: 25.07x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "0171048603d1bc057942ef21a0ca65c5bfa3cce5516b26a2e7c8a97b4605b9d5", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "84fe8696deb8cad6ea47af85a72b3f1a329393a02cd1871783547b8ef32893e1", "CSR_norm_hash": "db4d6d5f892425f5b1a9fb6404df67a67145723b6b2413286069a03de72c9ec8", "COO_norm_hash": "db4d6d5f892425f5b1a9fb6404df67a67145723b6b2413286069a03de72c9ec8", "ELL_norm_hash": "00d6955588c1bd5df84ed3bee12ee54f20a45e2f70900b833bbeaaa1d24b408a", "ROLF_norm_hash": "8ed88523effeac3f818b45ad493b8510ed902a927c187e58593ccb88e5990120", "DENGS_norm_hash": "84fe8696deb8cad6ea47af85a72b3f1a329393a02cd1871783547b8ef32893e1", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "e98a2d08c10f5a95e54f63ed06479febad5661e97c434d44acc9f88244c2e4c0", "CSR_qhash_d6": "1fbd2af9872415a272fabe42ec4881fd875453940f8f15b058c219f687a7a1b9", "COO_qhash_d6": "1fbd2af9872415a272fabe42ec4881fd875453940f8f15b058c219f687a7a1b9", "ELL_qhash_d6": "54dfe1d85aee227d6272ba2a5f8a5872009e3a6fcf67cb593d823beb20a02532", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.068906, "pilot_csr_per_iter_s": 0.001686, "pilot_coo_per_iter_s": 0.004624, "pilot_ell_per_iter_s": 0.100975, "rolv_build_s": 2.114612, "rolv_iter_s": 0.002074, "dense_iter_s": 0.000952, "csr_iter_s": 0.000963, "coo_iter_s": 0.004075, "ell_iter_s": 0.104886, "rolv_total_s": 4.188976, "baseline_total_s": 0.951809, "speedup_total_vs_selected_x": 0.227, "speedup_iter_vs_selected_x": 0.459, "rolv_vs_vendor_sparse_iter_x": 0.464, "rolv_vs_vendor_sparse_total_x": 0.23, "rolv_vs_coo_iter_x": 1.964, "rolv_vs_coo_total_x": 0.973, "rolv_vs_ell_iter_x": 50.563, "rolv_vs_ell_total_x": 25.067, "correct_norm": "OK"}

[2025-12-11 02:19:51] Platform: CPU | Seed: 123456 | Pattern: random | Zeros: 95% A_hash: e0613226f1838576cf6130cc6c9ea5c8b47359d694a00df5fb90e727f7a99782 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.035607s | CSR: 0.012056s | COO: 0.071207s | ELL: 0.640539s Selected baseline: CSR rolv load time (operator build): 2.332003 s rolv per-iter: 0.002171s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: dabb4146ff1d16a2de4369926706c93da6f53ef4739b87f126447fdd9207bc4a (CSR) CSR_norm_hash: dabb4146ff1d16a2de4369926706c93da6f53ef4739b87f126447fdd9207bc4a COO_norm_hash: dabb4146ff1d16a2de4369926706c93da6f53ef4739b87f126447fdd9207bc4a ELL_norm_hash: 832e1b07528a5c70cfe046b75d38abf1f8e8687c2542860283a7f5cdb4f65bdb ROLF_norm_hash: 2113a0575993d524efc054726b28d8f60390266246c73ad4d6d3fe8f13fa46cd DENGS_norm_hash: ee95b5c545a7dcf04fa45c25824c39f847ce7263c0a5c6729fcd9d6078f4f35b COO per-iter: 0.096412s | total: 96.411762s CSR per-iter: 0.013745s | total: 13.744711s ELL per-iter: 0.725244s | total: 725.372620s ROLF per-iter: 0.000870s | total: 0.872232s DENGS per-iter: 0.037909s | total: 37.908991s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 3.03x (≈ 203% faster) Speedup (per-iter): 6.29x (≈ 529% faster) Energy Savings (proxy): 84.10% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 6.33x | total: 3.05x rolv vs COO: Speedup (per-iter): 44.41x | total: 21.41x rolv vs ELL: Speedup (per-iter): 334.08x | total: 161.09x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "e0613226f1838576cf6130cc6c9ea5c8b47359d694a00df5fb90e727f7a99782", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "ee95b5c545a7dcf04fa45c25824c39f847ce7263c0a5c6729fcd9d6078f4f35b", "CSR_norm_hash": "dabb4146ff1d16a2de4369926706c93da6f53ef4739b87f126447fdd9207bc4a", "COO_norm_hash": "dabb4146ff1d16a2de4369926706c93da6f53ef4739b87f126447fdd9207bc4a", "ELL_norm_hash": "832e1b07528a5c70cfe046b75d38abf1f8e8687c2542860283a7f5cdb4f65bdb", "ROLF_norm_hash": "2113a0575993d524efc054726b28d8f60390266246c73ad4d6d3fe8f13fa46cd", "DENGS_norm_hash": "ee95b5c545a7dcf04fa45c25824c39f847ce7263c0a5c6729fcd9d6078f4f35b", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "734c412b8d3303e54a54ff61d9b5d2cc81c7dfcf350ffde11a88b454c2727f89", "CSR_qhash_d6": "b9b101604cc6cd1bf4c4a1c4cc66bc7439344eac973cf2914b60b0025dbc9c7c", "COO_qhash_d6": "b9b101604cc6cd1bf4c4a1c4cc66bc7439344eac973cf2914b60b0025dbc9c7c", "ELL_qhash_d6": "010700bba8e7f9f0e9acb48eec85442c993a68989f7e5b823ef599ffdf39d5e7", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.035607, "pilot_csr_per_iter_s": 0.012056, "pilot_coo_per_iter_s": 0.071207, "pilot_ell_per_iter_s": 0.640539, "rolv_build_s": 2.332003, "rolv_iter_s": 0.002171, "dense_iter_s": 0.013651, "csr_iter_s": 0.013745, "coo_iter_s": 0.096412, "ell_iter_s": 0.725244, "rolv_total_s": 4.502872, "baseline_total_s": 13.651, "speedup_total_vs_selected_x": 3.032, "speedup_iter_vs_selected_x": 6.288, "rolv_vs_vendor_sparse_iter_x": 6.331, "rolv_vs_vendor_sparse_total_x": 3.052, "rolv_vs_coo_iter_x": 44.412, "rolv_vs_coo_total_x": 21.411, "rolv_vs_ell_iter_x": 334.08, "rolv_vs_ell_total_x": 161.091, "correct_norm": "OK"}

[2025-12-11 02:35:03] Platform: CPU | Seed: 123456 | Pattern: power_law | Zeros: 95% A_hash: d83628ebede609a355b5f7d9a54635ad865f63954239ddff6d0d32e399ffeb79 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.034856s | CSR: 0.011004s | COO: 0.068926s | ELL: 0.665670s Selected baseline: CSR rolv load time (operator build): 3.087065 s rolv per-iter: 0.001927s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: ccf52a002c093338cefc9a078384844d920df0f89416ad39931f5e1ddbce4d57 (CSR) CSR_norm_hash: ccf52a002c093338cefc9a078384844d920df0f89416ad39931f5e1ddbce4d57 COO_norm_hash: ccf52a002c093338cefc9a078384844d920df0f89416ad39931f5e1ddbce4d57 ELL_norm_hash: 5ae59731aaff761b26ad73cf551bb96a04e6168bffcf235f8a9d4e07fee5dbdc ROLF_norm_hash: c7fb5eb9fcda9838dfdefb8d708b236273634d11d81587599641d17aa6997dcf DENGS_norm_hash: c1f5c798f2a280e5e4e3f3262abd4d7db701b5b7aa92b4f8efd70eb6f69e6246 COO per-iter: 0.075404s | total: 75.404105s CSR per-iter: 0.011974s | total: 11.974494s ELL per-iter: 0.706739s | total: 706.868850s ROLF per-iter: 0.000839s | total: 0.841779s DENGS per-iter: 0.037933s | total: 37.932655s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 2.39x (≈ 139% faster) Speedup (per-iter): 6.21x (≈ 521% faster) Energy Savings (proxy): 83.89% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 6.22x | total: 2.39x rolv vs COO: Speedup (per-iter): 39.14x | total: 15.04x rolv vs ELL: Speedup (per-iter): 366.84x | total: 140.99x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "d83628ebede609a355b5f7d9a54635ad865f63954239ddff6d0d32e399ffeb79", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "c1f5c798f2a280e5e4e3f3262abd4d7db701b5b7aa92b4f8efd70eb6f69e6246", "CSR_norm_hash": "ccf52a002c093338cefc9a078384844d920df0f89416ad39931f5e1ddbce4d57", "COO_norm_hash": "ccf52a002c093338cefc9a078384844d920df0f89416ad39931f5e1ddbce4d57", "ELL_norm_hash": "5ae59731aaff761b26ad73cf551bb96a04e6168bffcf235f8a9d4e07fee5dbdc", "ROLF_norm_hash": "c7fb5eb9fcda9838dfdefb8d708b236273634d11d81587599641d17aa6997dcf", "DENGS_norm_hash": "c1f5c798f2a280e5e4e3f3262abd4d7db701b5b7aa92b4f8efd70eb6f69e6246", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "baa0adc541688b9b739e6c0ee8a40ea4c932668338e17b3d26600af90f227472", "CSR_qhash_d6": "dda953919cdf47b8a3c775d3f9a8bbd82129b9f5b9647c7fb276a1dca05acf9a", "COO_qhash_d6": "dda953919cdf47b8a3c775d3f9a8bbd82129b9f5b9647c7fb276a1dca05acf9a", "ELL_qhash_d6": "826ad590c34dc93f5c7e19dc5fa56b685209598ee38d2f3023bd274bee57a496", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.034856, "pilot_csr_per_iter_s": 0.011004, "pilot_coo_per_iter_s": 0.068926, "pilot_ell_per_iter_s": 0.66567, "rolv_build_s": 3.087065, "rolv_iter_s": 0.001927, "dense_iter_s": 0.011958, "csr_iter_s": 0.011974, "coo_iter_s": 0.075404, "ell_iter_s": 0.706739, "rolv_total_s": 5.013603, "baseline_total_s": 11.958303, "speedup_total_vs_selected_x": 2.385, "speedup_iter_vs_selected_x": 6.207, "rolv_vs_vendor_sparse_iter_x": 6.216, "rolv_vs_vendor_sparse_total_x": 2.388, "rolv_vs_coo_iter_x": 39.14, "rolv_vs_coo_total_x": 15.04, "rolv_vs_ell_iter_x": 366.844, "rolv_vs_ell_total_x": 140.99, "correct_norm": "OK"}

[2025-12-11 02:49:26] Platform: CPU | Seed: 123456 | Pattern: banded | Zeros: 95% A_hash: 71668be72f1778fa4e03a508abcf5f066a1a5e61ca8e700fec3e9b8bd616b217 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.063245s | CSR: 0.001152s | COO: 0.003557s | ELL: 0.070612s Selected baseline: CSR rolv load time (operator build): 3.156332 s rolv per-iter: 0.002083s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: f686f8421c319a41341137c27b93293b615289094f27b94f8be2044d12a4d844 (CSR) CSR_norm_hash: f686f8421c319a41341137c27b93293b615289094f27b94f8be2044d12a4d844 COO_norm_hash: f686f8421c319a41341137c27b93293b615289094f27b94f8be2044d12a4d844 ELL_norm_hash: 8421cb62d01bd9abe6f209864e35c04907baba99dccddccaa5a6f60bacbb593e ROLF_norm_hash: 9997befbad9feea95635feee56536e5dbe238fc696c66994fca5516b44afebe5 DENGS_norm_hash: 402f095b0e2f0ba63a9075a8a73e2170675f9b7bd406b9a24d08379b49e2749d COO per-iter: 0.010000s | total: 9.999834s CSR per-iter: 0.000945s | total: 0.945185s ELL per-iter: 0.057557s | total: 57.685111s ROLF per-iter: 0.001319s | total: 1.321512s DENGS per-iter: 0.038183s | total: 38.183089s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 0.32x (≈ -68% faster) Speedup (per-iter): 0.81x (≈ -19% faster) Energy Savings (proxy): -24.14% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 0.45x | total: 0.18x rolv vs COO: Speedup (per-iter): 4.80x | total: 1.91x rolv vs ELL: Speedup (per-iter): 27.63x | total: 11.01x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "71668be72f1778fa4e03a508abcf5f066a1a5e61ca8e700fec3e9b8bd616b217", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "402f095b0e2f0ba63a9075a8a73e2170675f9b7bd406b9a24d08379b49e2749d", "CSR_norm_hash": "f686f8421c319a41341137c27b93293b615289094f27b94f8be2044d12a4d844", "COO_norm_hash": "f686f8421c319a41341137c27b93293b615289094f27b94f8be2044d12a4d844", "ELL_norm_hash": "8421cb62d01bd9abe6f209864e35c04907baba99dccddccaa5a6f60bacbb593e", "ROLF_norm_hash": "9997befbad9feea95635feee56536e5dbe238fc696c66994fca5516b44afebe5", "DENGS_norm_hash": "402f095b0e2f0ba63a9075a8a73e2170675f9b7bd406b9a24d08379b49e2749d", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "9005128f73ca8d887b59c68186d974c0b80369053e79dad5b655d05498d3bdc8", "CSR_qhash_d6": "d289db24c7e1ed447da5f9cdb1d9fed4dcb4f0bf7640da4361b3377979ebd2a5", "COO_qhash_d6": "d289db24c7e1ed447da5f9cdb1d9fed4dcb4f0bf7640da4361b3377979ebd2a5", "ELL_qhash_d6": "75dd4ea29aeec46836d7ac5e01d0139de6e9dadaf5a3897f94c7a316a6ee70dd", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.063245, "pilot_csr_per_iter_s": 0.001152, "pilot_coo_per_iter_s": 0.003557, "pilot_ell_per_iter_s": 0.070612, "rolv_build_s": 3.156332, "rolv_iter_s": 0.002083, "dense_iter_s": 0.001678, "csr_iter_s": 0.000945, "coo_iter_s": 0.01, "ell_iter_s": 0.057557, "rolv_total_s": 5.239682, "baseline_total_s": 1.678188, "speedup_total_vs_selected_x": 0.32, "speedup_iter_vs_selected_x": 0.806, "rolv_vs_vendor_sparse_iter_x": 0.454, "rolv_vs_vendor_sparse_total_x": 0.18, "rolv_vs_coo_iter_x": 4.8, "rolv_vs_coo_total_x": 1.908, "rolv_vs_ell_iter_x": 27.627, "rolv_vs_ell_total_x": 11.009, "correct_norm": "OK"}

[2025-12-11 02:51:24] Platform: CPU | Seed: 123456 | Pattern: block_diagonal | Zeros: 95% A_hash: f1403bc49489eefe4d4b3cc45cba7d9a09267cfb0c9396e42bcca0be270d5ef0 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.035083s | CSR: 0.000542s | COO: 0.002055s | ELL: 0.065174s Selected baseline: CSR rolv load time (operator build): 2.063174 s rolv per-iter: 0.001780s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 2bdad1e9ceadbffe60a88b4afa42f417f4752ab9d12a28fbbfc213075009e019 (CSR) CSR_norm_hash: 2bdad1e9ceadbffe60a88b4afa42f417f4752ab9d12a28fbbfc213075009e019 COO_norm_hash: 2bdad1e9ceadbffe60a88b4afa42f417f4752ab9d12a28fbbfc213075009e019 ELL_norm_hash: 8a30e12433d089976a9cd02c76b3c58b30d84307d728fbccfd057707eeece571 ROLF_norm_hash: b850a29467f2854f83dfd4d4bae49e0d61156d610f0fcee8dd6ef4edd9389fe0 DENGS_norm_hash: c2688f0d00d7da6e1cf92d160e5f5e6561593005a480663e8bd6328af2986d80 COO per-iter: 0.002378s | total: 2.377535s CSR per-iter: 0.000644s | total: 0.643688s ELL per-iter: 0.066813s | total: 66.890599s ROLF per-iter: 0.000843s | total: 0.845175s DENGS per-iter: 0.038949s | total: 38.949351s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 0.17x (≈ -83% faster) Speedup (per-iter): 0.36x (≈ -64% faster) Energy Savings (proxy): -175.06% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 0.36x | total: 0.17x rolv vs COO: Speedup (per-iter): 1.34x | total: 0.62x rolv vs ELL: Speedup (per-iter): 37.53x | total: 17.40x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "f1403bc49489eefe4d4b3cc45cba7d9a09267cfb0c9396e42bcca0be270d5ef0", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "c2688f0d00d7da6e1cf92d160e5f5e6561593005a480663e8bd6328af2986d80", "CSR_norm_hash": "2bdad1e9ceadbffe60a88b4afa42f417f4752ab9d12a28fbbfc213075009e019", "COO_norm_hash": "2bdad1e9ceadbffe60a88b4afa42f417f4752ab9d12a28fbbfc213075009e019", "ELL_norm_hash": "8a30e12433d089976a9cd02c76b3c58b30d84307d728fbccfd057707eeece571", "ROLF_norm_hash": "b850a29467f2854f83dfd4d4bae49e0d61156d610f0fcee8dd6ef4edd9389fe0", "DENGS_norm_hash": "c2688f0d00d7da6e1cf92d160e5f5e6561593005a480663e8bd6328af2986d80", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "19e1a972b2da9c94b2a18be4560ec3f8288e95043360844bea8c7eb519942efc", "CSR_qhash_d6": "ccc656fe810f0f5f0fb54ceb88b4e59e3449c30da8acc666842688855d851393", "COO_qhash_d6": "ccc656fe810f0f5f0fb54ceb88b4e59e3449c30da8acc666842688855d851393", "ELL_qhash_d6": "b6cdd4fdd41473e9c2c019eda893e114c089f4ac22b5c261158de89449085791", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.035083, "pilot_csr_per_iter_s": 0.000542, "pilot_coo_per_iter_s": 0.002055, "pilot_ell_per_iter_s": 0.065174, "rolv_build_s": 2.063174, "rolv_iter_s": 0.00178, "dense_iter_s": 0.000647, "csr_iter_s": 0.000644, "coo_iter_s": 0.002378, "ell_iter_s": 0.066813, "rolv_total_s": 3.843466, "baseline_total_s": 0.647228, "speedup_total_vs_selected_x": 0.168, "speedup_iter_vs_selected_x": 0.364, "rolv_vs_vendor_sparse_iter_x": 0.362, "rolv_vs_vendor_sparse_total_x": 0.167, "rolv_vs_coo_iter_x": 1.335, "rolv_vs_coo_total_x": 0.619, "rolv_vs_ell_iter_x": 37.529, "rolv_vs_ell_total_x": 17.404, "correct_norm": "OK"}

[2025-12-11 02:53:21] Platform: CPU | Seed: 123456 | Pattern: random | Zeros: 99% A_hash: 2e6915146bcd1a4f569ec8311a0001151255ab66dda033e775861a3e2fbf7394 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.058375s | CSR: 0.003943s | COO: 0.016139s | ELL: 0.165058s Selected baseline: CSR rolv load time (operator build): 1.925057 s rolv per-iter: 0.001752s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 439773943ecd707dc91bf3fd276ceac196571eeb43ed11cd323729e82e02d087 (CSR) CSR_norm_hash: 439773943ecd707dc91bf3fd276ceac196571eeb43ed11cd323729e82e02d087 COO_norm_hash: 439773943ecd707dc91bf3fd276ceac196571eeb43ed11cd323729e82e02d087 ELL_norm_hash: 9afddcb767aa0eb9c8eec04ec0ad2068d9684a17af990077ec0ee62c02aaf7b8 ROLF_norm_hash: 89d3acefa79af5b4dd14cbccbb144b08b5629e7134a64ccba89ac3ffea59d740 DENGS_norm_hash: 7e119c999677b6bcb9c85bd8aab75351043ab89eb3375262284f278aaebb5e40 COO per-iter: 0.015688s | total: 15.688462s CSR per-iter: 0.003284s | total: 3.283711s ELL per-iter: 0.229945s | total: 230.081358s ROLF per-iter: 0.000850s | total: 0.852165s DENGS per-iter: 0.038164s | total: 38.163997s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 0.76x (≈ -24% faster) Speedup (per-iter): 1.59x (≈ 59% faster) Energy Savings (proxy): 37.16% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 1.87x | total: 0.89x rolv vs COO: Speedup (per-iter): 8.95x | total: 4.27x rolv vs ELL: Speedup (per-iter): 131.25x | total: 62.57x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "2e6915146bcd1a4f569ec8311a0001151255ab66dda033e775861a3e2fbf7394", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "7e119c999677b6bcb9c85bd8aab75351043ab89eb3375262284f278aaebb5e40", "CSR_norm_hash": "439773943ecd707dc91bf3fd276ceac196571eeb43ed11cd323729e82e02d087", "COO_norm_hash": "439773943ecd707dc91bf3fd276ceac196571eeb43ed11cd323729e82e02d087", "ELL_norm_hash": "9afddcb767aa0eb9c8eec04ec0ad2068d9684a17af990077ec0ee62c02aaf7b8", "ROLF_norm_hash": "89d3acefa79af5b4dd14cbccbb144b08b5629e7134a64ccba89ac3ffea59d740", "DENGS_norm_hash": "7e119c999677b6bcb9c85bd8aab75351043ab89eb3375262284f278aaebb5e40", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "5964bfeeeb7693295c2ff996e8d59457657b83ad2dc173a946204da5edcd9a14", "CSR_qhash_d6": "d28e9b0ad1cfedd00fd9585bb57ced681b20a1be731afd36c1767637c1bae832", "COO_qhash_d6": "d28e9b0ad1cfedd00fd9585bb57ced681b20a1be731afd36c1767637c1bae832", "ELL_qhash_d6": "322a1ed976f63503b08042093e9e8329add0ac44c48636f32ca838fd24be0a41", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.058375, "pilot_csr_per_iter_s": 0.003943, "pilot_coo_per_iter_s": 0.016139, "pilot_ell_per_iter_s": 0.165058, "rolv_build_s": 1.925057, "rolv_iter_s": 0.001752, "dense_iter_s": 0.002788, "csr_iter_s": 0.003284, "coo_iter_s": 0.015688, "ell_iter_s": 0.229945, "rolv_total_s": 3.677025, "baseline_total_s": 2.787905, "speedup_total_vs_selected_x": 0.758, "speedup_iter_vs_selected_x": 1.591, "rolv_vs_vendor_sparse_iter_x": 1.874, "rolv_vs_vendor_sparse_total_x": 0.893, "rolv_vs_coo_iter_x": 8.955, "rolv_vs_coo_total_x": 4.267, "rolv_vs_ell_iter_x": 131.25, "rolv_vs_ell_total_x": 62.573, "correct_norm": "OK"}

[2025-12-11 02:58:20] Platform: CPU | Seed: 123456 | Pattern: power_law | Zeros: 99% A_hash: a1474886466033ec37cf6d6065e869d1d3f0aae5c7e9fda9f81307291f500937 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.058788s | CSR: 0.003674s | COO: 0.017346s | ELL: 0.159393s Selected baseline: CSR rolv load time (operator build): 3.770959 s rolv per-iter: 0.002005s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: ed603b9a20f12cf4161c8a7e082189571a1365fec0b7f3cebf296c6005eaabcb (CSR) CSR_norm_hash: ed603b9a20f12cf4161c8a7e082189571a1365fec0b7f3cebf296c6005eaabcb COO_norm_hash: ed603b9a20f12cf4161c8a7e082189571a1365fec0b7f3cebf296c6005eaabcb ELL_norm_hash: 7d293e8852b9ff05ee3eceaecf071f5a78161936b2e4468ee759ff63ef42cf6b ROLF_norm_hash: 82e4df55469edc827f0f4b09f7bd5028e0a50fe2d0e15e6c6ba9590d4f914565 DENGS_norm_hash: 68fb6189b088fa42e10e8faf70b264158fb1226c5c304a24812f305bbc527d37 COO per-iter: 0.015598s | total: 15.597968s CSR per-iter: 0.003330s | total: 3.330215s ELL per-iter: 0.170776s | total: 170.923476s ROLF per-iter: 0.001021s | total: 1.023496s DENGS per-iter: 0.041853s | total: 41.853096s Correctness vs Selected Baseline: Failed | vs CSR: Failed | vs COO: Failed | vs ELL: Failed Speedup (total): 0.47x (≈ -53% faster) Speedup (per-iter): 1.34x (≈ 34% faster) Energy Savings (proxy): 25.55% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 1.66x | total: 0.58x rolv vs COO: Speedup (per-iter): 7.78x | total: 2.70x rolv vs ELL: Speedup (per-iter): 85.19x | total: 29.59x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "a1474886466033ec37cf6d6065e869d1d3f0aae5c7e9fda9f81307291f500937", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "68fb6189b088fa42e10e8faf70b264158fb1226c5c304a24812f305bbc527d37", "CSR_norm_hash": "ed603b9a20f12cf4161c8a7e082189571a1365fec0b7f3cebf296c6005eaabcb", "COO_norm_hash": "ed603b9a20f12cf4161c8a7e082189571a1365fec0b7f3cebf296c6005eaabcb", "ELL_norm_hash": "7d293e8852b9ff05ee3eceaecf071f5a78161936b2e4468ee759ff63ef42cf6b", "ROLF_norm_hash": "82e4df55469edc827f0f4b09f7bd5028e0a50fe2d0e15e6c6ba9590d4f914565", "DENGS_norm_hash": "68fb6189b088fa42e10e8faf70b264158fb1226c5c304a24812f305bbc527d37", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "e12dd9e9a516bb13d9326d75a8c6590147d7cee5b593e4d5a7e06ccf50f17a50", "CSR_qhash_d6": "d77210277faf7e9379f698685f7c380aeafeb62067130d033d10cecd231aea37", "COO_qhash_d6": "d77210277faf7e9379f698685f7c380aeafeb62067130d033d10cecd231aea37", "ELL_qhash_d6": "08dd7345d533c59bf0e581ee2b7b0dd56c6ce33e70bc9e5888529a38789d4aa7", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.058788, "pilot_csr_per_iter_s": 0.003674, "pilot_coo_per_iter_s": 0.017346, "pilot_ell_per_iter_s": 0.159393, "rolv_build_s": 3.770959, "rolv_iter_s": 0.002005, "dense_iter_s": 0.002693, "csr_iter_s": 0.00333, "coo_iter_s": 0.015598, "ell_iter_s": 0.170776, "rolv_total_s": 5.775659, "baseline_total_s": 2.6925, "speedup_total_vs_selected_x": 0.466, "speedup_iter_vs_selected_x": 1.343, "rolv_vs_vendor_sparse_iter_x": 1.661, "rolv_vs_vendor_sparse_total_x": 0.577, "rolv_vs_coo_iter_x": 7.781, "rolv_vs_coo_total_x": 2.701, "rolv_vs_ell_iter_x": 85.188, "rolv_vs_ell_total_x": 29.594, "correct_norm": "FAIL"}

[2025-12-11 03:02:26] Platform: CPU | Seed: 123456 | Pattern: banded | Zeros: 99% A_hash: dcc586ce4a3ce72d269fcc671f30ddffd61a65d9394954523f1f6b56dcf096c3 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.037779s | CSR: 0.000656s | COO: 0.001655s | ELL: 0.029914s Selected baseline: CSR rolv load time (operator build): 2.167854 s rolv per-iter: 0.002145s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 64ae169ae73ae36192138d9f02273ea3745ce308c92512a444f733a19e0b8f76 (CSR) CSR_norm_hash: 64ae169ae73ae36192138d9f02273ea3745ce308c92512a444f733a19e0b8f76 COO_norm_hash: 64ae169ae73ae36192138d9f02273ea3745ce308c92512a444f733a19e0b8f76 ELL_norm_hash: d38fd4c1d7aef498481114debefc4632750d24db494be8e6fc066303eddd87c3 ROLF_norm_hash: 5d271c07b690baa7ebfc262bb366eb35cfc516dc9c27b54220ed9b04549149e8 DENGS_norm_hash: 75da0442b53c98b2d636a1820871e24bbda1edfb5e2aeee6cb352b6af8d7f8bd COO per-iter: 0.001333s | total: 1.333411s CSR per-iter: 0.000520s | total: 0.520355s ELL per-iter: 0.030322s | total: 30.427032s ROLF per-iter: 0.000805s | total: 0.807986s DENGS per-iter: 0.038300s | total: 38.300254s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 0.13x (≈ -87% faster) Speedup (per-iter): 0.25x (≈ -75% faster) Energy Savings (proxy): -292.56% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 0.24x | total: 0.12x rolv vs COO: Speedup (per-iter): 0.62x | total: 0.31x rolv vs ELL: Speedup (per-iter): 14.14x | total: 7.06x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "dcc586ce4a3ce72d269fcc671f30ddffd61a65d9394954523f1f6b56dcf096c3", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "75da0442b53c98b2d636a1820871e24bbda1edfb5e2aeee6cb352b6af8d7f8bd", "CSR_norm_hash": "64ae169ae73ae36192138d9f02273ea3745ce308c92512a444f733a19e0b8f76", "COO_norm_hash": "64ae169ae73ae36192138d9f02273ea3745ce308c92512a444f733a19e0b8f76", "ELL_norm_hash": "d38fd4c1d7aef498481114debefc4632750d24db494be8e6fc066303eddd87c3", "ROLF_norm_hash": "5d271c07b690baa7ebfc262bb366eb35cfc516dc9c27b54220ed9b04549149e8", "DENGS_norm_hash": "75da0442b53c98b2d636a1820871e24bbda1edfb5e2aeee6cb352b6af8d7f8bd", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "fe8c57e1f1f285635e77c417dbc5b809f7bccf4a4daaf19899bb6762922f35b9", "CSR_qhash_d6": "55659a0f91a90ccb4d4120560b0706c8c64e347df8ca52adae2d76bd3f542cdf", "COO_qhash_d6": "55659a0f91a90ccb4d4120560b0706c8c64e347df8ca52adae2d76bd3f542cdf", "ELL_qhash_d6": "6a179c02412231385f1d14adf6b16bfd39a02144585e1bcee53f0e5e854f546f", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.037779, "pilot_csr_per_iter_s": 0.000656, "pilot_coo_per_iter_s": 0.001655, "pilot_ell_per_iter_s": 0.029914, "rolv_build_s": 2.167854, "rolv_iter_s": 0.002145, "dense_iter_s": 0.000546, "csr_iter_s": 0.00052, "coo_iter_s": 0.001333, "ell_iter_s": 0.030322, "rolv_total_s": 4.312762, "baseline_total_s": 0.546389, "speedup_total_vs_selected_x": 0.127, "speedup_iter_vs_selected_x": 0.255, "rolv_vs_vendor_sparse_iter_x": 0.243, "rolv_vs_vendor_sparse_total_x": 0.121, "rolv_vs_coo_iter_x": 0.622, "rolv_vs_coo_total_x": 0.309, "rolv_vs_ell_iter_x": 14.137, "rolv_vs_ell_total_x": 7.055, "correct_norm": "OK"}

[2025-12-11 03:03:44] Platform: CPU | Seed: 123456 | Pattern: block_diagonal | Zeros: 99% A_hash: 9da76096a333c2891d4bec24c31e88cfcc1a607b373b260b9566521846a1fa49 | V_hash: 448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070 Baseline pilots per-iter -> Dense: 0.067664s | CSR: 0.000989s | COO: 0.001090s | ELL: 0.018720s Selected baseline: CSR rolv load time (operator build): 1.884291 s rolv per-iter: 0.001994s rolv_norm_hash: 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd | qhash(d=6): 8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd BASE_norm_hash: 42642c5296cbf598a54b7bf9a00f6ea8801f11e44e8d6f8620edd696f5c7f0ab (CSR) CSR_norm_hash: 42642c5296cbf598a54b7bf9a00f6ea8801f11e44e8d6f8620edd696f5c7f0ab COO_norm_hash: 42642c5296cbf598a54b7bf9a00f6ea8801f11e44e8d6f8620edd696f5c7f0ab ELL_norm_hash: a1d106ed49b5b83cf0f2dd18194b2074366915a87bca94e50a174686c7fc4414 ROLF_norm_hash: f97324bcdf2a80316f91e5a7e8f3f5ee30c796a74ddcac21b5d9360fedec1a57 DENGS_norm_hash: e86340871a374c44c261495df484d71e3455f443cb20c6bb9c1e42dd6efadbb3 COO per-iter: 0.000753s | total: 0.752986s CSR per-iter: 0.000486s | total: 0.486206s ELL per-iter: 0.019887s | total: 19.984992s ROLF per-iter: 0.000836s | total: 0.838725s DENGS per-iter: 0.038351s | total: 38.351180s Correctness vs Selected Baseline: Verified | vs CSR: Verified | vs COO: Verified | vs ELL: Verified Speedup (total): 0.12x (≈ -88% faster) Speedup (per-iter): 0.24x (≈ -76% faster) Energy Savings (proxy): -319.89% rolv vs CPU Sparse (CSR) -> Speedup (per-iter): 0.24x | total: 0.13x rolv vs COO: Speedup (per-iter): 0.38x | total: 0.19x rolv vs ELL: Speedup (per-iter): 9.97x | total: 5.15x {"platform": "CPU", "device": "CPU", "dense_label": "CPU Dense", "sparse_label": "CPU Sparse", "input_hash_A": "9da76096a333c2891d4bec24c31e88cfcc1a607b373b260b9566521846a1fa49", "input_hash_B": "448b453ff50675840e0e32980b9e77974b1188713089eb2c28e45b6d12701070", "ROLV_norm_hash": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_norm_hash": "e86340871a374c44c261495df484d71e3455f443cb20c6bb9c1e42dd6efadbb3", "CSR_norm_hash": "42642c5296cbf598a54b7bf9a00f6ea8801f11e44e8d6f8620edd696f5c7f0ab", "COO_norm_hash": "42642c5296cbf598a54b7bf9a00f6ea8801f11e44e8d6f8620edd696f5c7f0ab", "ELL_norm_hash": "a1d106ed49b5b83cf0f2dd18194b2074366915a87bca94e50a174686c7fc4414", "ROLF_norm_hash": "f97324bcdf2a80316f91e5a7e8f3f5ee30c796a74ddcac21b5d9360fedec1a57", "DENGS_norm_hash": "e86340871a374c44c261495df484d71e3455f443cb20c6bb9c1e42dd6efadbb3", "ROLV_qhash_d6": "8dbe5f139fd946d4cd84e8cc612cd9f68cbc87e394457884acc0c5dad56dd8dd", "DENSE_qhash_d6": "c46b9eeff102821df6ea4c337840534119d54ee25f6934da15fa5919a4458d44", "CSR_qhash_d6": "35232511a40e7656ed02e2960e271fd9ef4274ac6852c0e7f567f26648bfeaa4", "COO_qhash_d6": "35232511a40e7656ed02e2960e271fd9ef4274ac6852c0e7f567f26648bfeaa4", "ELL_qhash_d6": "ee8d3e58910f776fd058f4f2690b5f5d90aa3dfb589c28f079d7a7a08868937b", "path_selected": "CSR", "pilot_dense_per_iter_s": 0.067664, "pilot_csr_per_iter_s": 0.000989, "pilot_coo_per_iter_s": 0.00109, "pilot_ell_per_iter_s": 0.01872, "rolv_build_s": 1.884291, "rolv_iter_s": 0.001994, "dense_iter_s": 0.000475, "csr_iter_s": 0.000486, "coo_iter_s": 0.000753, "ell_iter_s": 0.019887, "rolv_total_s": 3.878486, "baseline_total_s": 0.474935, "speedup_total_vs_selected_x": 0.122, "speedup_iter_vs_selected_x": 0.238, "rolv_vs_vendor_sparse_iter_x": 0.244, "rolv_vs_vendor_sparse_total_x": 0.125, "rolv_vs_coo_iter_x": 0.378, "rolv_vs_coo_total_x": 0.194, "rolv_vs_ell_iter_x": 9.972, "rolv_vs_ell_total_x": 5.153, "correct_norm": "OK"}

=== FOOTER REPORT (CPU) ===

Aggregate speedup (total vs selected): 4.16x (≈ 316% faster)
Aggregate speedup (per-iter vs selected): 8.80x (≈ 780% faster)
Aggregate energy savings (proxy vs selected): 14.2%
Verification: deterministic algorithms, CSR/COO canonicalization, CPU-fp64 normalization and SHA-256 hashing. {"platform": "CPU", "device": "CPU", "aggregate_speedup_total_vs_selected_x": 4.163, "aggregate_speedup_iter_vs_selected_x": 8.795, "aggregate_energy_savings_pct": 14.207, "verification": "Deterministic algorithms, CSR/COO canonicalization, CPU-fp64 normalization, SHA-256 hashing"}

=== Timing Measurement Explanation (CPU) ===

Per-iteration timing:
- Each library (Dense GEMM, CSR/COO SpMM, ROLV, ROLF, DENGS) is warmed up for a fixed number of iterations.
- Then 'iters' iterations are executed. The average time per iteration is reported as _iter_s.
Build/setup time:
- For ROLV, operator construction (tiling, quantization, surrogate build) is timed separately as rolv_build_s.
- Vendor baselines (Dense/CSR/COO) have negligible build cost, so only per-iter times are used.
Total time:
- For each library, total runtime = build/setup time + (per-iter time × number of iterations).
Speedup calculation:
- Speedup (per-iter) = baseline_iter_s / rolv_iter_s
- Speedup (total) = baseline_total_s / rolv_total_s
Fairness guarantee:
- Identical input generation via fixed RNG seed; input hashes printed.
- Outputs normalized in CPU-fp64 before hashing to remove numeric artifacts.
- CSR/COO canonicalization (sorted indices) stabilizes sparse ordering and ensures reproducible hashes.

Imagination is the Only Limitation to Innovation

Name		Name	Last commit message	Last commit date
Latest commit History 8 Commits
README.md		README.md

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

GPU Validation Harness – Dense and CSR baselines (NVIDIA/AMD)

CPU Validation Harness (Dec 2025)

generate_matrix and generate_vectors as in previous CPU doc...

Parser and cases as before...

Apple Silicon Mac Validation Harness – Dense, CSR, COO baselines with timing

generate_matrix and generate_vectors from full Mac code (supports random, power_law, banded, block_diagonal)...

Rolv E. Heggenhougen

Rolv E. Heggenhougen

About

Uh oh!

Releases

Packages

Uh oh!

Contributors

Uh oh!

Folders and files

Latest commit

History

Repository files navigation

GPU Validation Harness – Dense and CSR baselines (NVIDIA/AMD)

CPU Validation Harness (Dec 2025)

generate_matrix and generate_vectors as in previous CPU doc...

Parser and cases as before...

Apple Silicon Mac Validation Harness – Dense, CSR, COO baselines with timing

generate_matrix and generate_vectors from full Mac code (supports random, power_law, banded, block_diagonal)...

Rolv E. Heggenhougen

Rolv E. Heggenhougen

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Packages