docs: MPS/Apple-GPU acceleration pathways (#72) - #73
Merged
neuromechanist merged 1 commit intoJul 8, 2026
Merged
Conversation
Research into accelerating AMICATorchNG on Apple Silicon. Governing fact: Apple GPUs have no FP64 hardware, and neither PyTorch MPS nor MLX supports float64 on GPU (CPU-only, confirmed mid-2026). So the float64 parity path can never run on an Apple GPU, and every MPS pathway is gated on a stable float32 AMICA. Pathways: (A) float32 stabilization via Kahan/compensated summation + mixed precision -- the enabler, with a concrete technique the #70 attempt lacked; (B) exploit the large-tensor regime where MPS's fixed dispatch overhead amortizes (sweep channels/blocks to find the CPU crossover); (C) MLX port (2-3x over PyTorch MPS, full rewrite, v2); (D) FP64 emulation -- dead end (custom-Metal-only, no transcendentals). Cited throughout.
neuromechanist
deleted the
72-mps-apple-silicon-gpu-acceleration-pathway-gated-on-float32-stability
branch
July 8, 2026 02:40
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes #72. Research deliverable:
.context/mps_pathways.md.Governing finding
Apple-Silicon GPUs have no FP64 hardware; PyTorch MPS and MLX both support float64 only on CPU (confirmed mid-2026 via PyTorch/Apple-Developer/MLX sources). AMICATorchNG runs float64 for parity, so the parity path can never run on an Apple GPU -- every MPS pathway is gated on a numerically stable float32 AMICA.
Pathways (cited)
Recommendation
Gated on A. Sequence A -> B -> C. Until then float64-CUDA (4.5x, bit-safe) is the production GPU path. Docs-only.