[QDP] Fix latency benchmark to use batch encoding

[guan404ming]

Purpose of PR

• Added encode_batch() method to Python bindings accepting NumPy 2D array directly (zero-copy)
• Updated run_mahout() to process entire batch in single GPU kernel call

Related Issues or PRs

Changes Made

• Bug fix
• New feature
• Refactoring
• Documentation
• Test
• CI/CD pipeline
• Other

Breaking Changes

• Yes
• No

Checklist

• Added or updated unit tests for all changes
• Added or updated documentation for all changes
• Successfully built and ran all unit tests or manual tests locally
• PR title follows "MAHOUT-XXX: Brief Description" format (if related to an issue)
• Code follows ASF guidelines

[guan404ming]

(qdp-python) titan% uv run python benchmark/benchmark_latency.py --qubits 12 --batches 20 --batch-size 8 --prefetch 4
Uninstalled 1 package in 1ms
Installed 1 package in 5ms
Generating 160 samples of 12 qubits...
  Batch size   : 8
  Vector length: 4096
  Batches      : 20
  Prefetch     : 4
  Frameworks   : pennylane, qiskit-init, qiskit-statevector, mahout
  Generated 160 samples
  PennyLane/Qiskit format: 5.00 MB
  Mahout format: 5.00 MB

======================================================================
DATA-TO-STATE LATENCY BENCHMARK: 12 Qubits, 160 Samples
======================================================================

[PennyLane] Full Pipeline (DataLoader -> GPU)...
  Total Time: 0.0423 s (0.264 ms/vector)

[Qiskit Initialize] Full Pipeline (DataLoader -> GPU)...
  Total Time: 11.8223 s (73.890 ms/vector)

[Qiskit Statevector] Full Pipeline (DataLoader -> GPU)...
  Total Time: 0.0222 s (0.139 ms/vector)

[Mahout] Full Pipeline (DataLoader -> GPU)...
  Total Time: 0.0157 s (0.098 ms/vector)

======================================================================
LATENCY (Lower is Better)
Samples: 160, Qubits: 12
======================================================================
Mahout                  0.098 ms/vector
Qiskit Statevector      0.139 ms/vector
PennyLane               0.264 ms/vector
Qiskit Initialize      73.890 ms/vector
----------------------------------------------------------------------
Speedup vs PennyLane:            2.69x
Speedup vs Qiskit Init:         751.84x
Speedup vs Qiskit Statevec:       1.41x

[400Ping]

Thanks for the fix! I was referencing benchmark_throughput when I wrote the code so I didn't notice this issue. I also opened a issue to fix benchmark_throughput to batch encoding.

[guan404ming]

Thanks for the fix! I was referencing benchmark_throughput when I wrote the code so I didn't notice this issue. I also opened a issue to fix benchmark_throughput to batch encoding.

Sure, let's do it after this one.

[guan404ming]

cc @ryankert01 @rich7420

[ryankert01]

This is nice

[guan404ming]

Thanks!