migrate to rustc's FxHasher#16
Merged
Merged
Conversation
Human9000-bit
force-pushed
the
faster-hashing
branch
from
f9d7c4e to
5b71f06
Compare
Collaborator
Author
|
/bench |
ONNX CPU benchmarkGitHub-hosted runner signal, advisory only. Base and head are compared within the same architecture job.
Head op/kernel summary
Missing paired result(s): aarch64/gemm_relu_256x512x512, aarch64/small_conv_relu_64, aarch64/small_dw_pw_64, aarch64/small_gemm_relu_1x128, aarch64/small_residual_conv_64, aarch64/winograd_3x3_20x20_c256, aarch64/winograd_3x3_80x80_c64, aarch64/yolo11n_bus_640, aarch64/yolov8n_bus_640 Raw JSON artifacts include per-run min/p50/avg/p95/p99, optimized node counts, yscv CPU dispatch report, and per-node runner profile summaries. |
Human9000-bit
force-pushed
the
faster-hashing
branch
2 times, most recently
from
24c8ad8 to
aabfde4
Compare
ONNX CPU benchmarkGitHub-hosted runner signal, advisory only. Base and head are compared within the same architecture job.
Head op/kernel summary
Missing paired result(s): aarch64/gemm_relu_256x512x512, aarch64/small_conv_relu_64, aarch64/small_dw_pw_64, aarch64/small_gemm_relu_1x128, aarch64/small_residual_conv_64, aarch64/winograd_3x3_20x20_c256, aarch64/winograd_3x3_80x80_c64, aarch64/yolo11n_bus_640, aarch64/yolov8n_bus_640 Raw JSON artifacts include per-run min/p50/avg/p95/p99, optimized node counts, yscv CPU dispatch report, and per-node runner profile summaries. |
Human9000-bit
force-pushed
the
faster-hashing
branch
3 times, most recently
from
b343b1b to
cbd1c0c
Compare
Human9000-bit
force-pushed
the
faster-hashing
branch
from
cbd1c0c to
7775595
Compare
enthropy7
added a commit
that referenced
this pull request
Jul 11, 2026
Revert e3cffe0 — put FxHashMap back across the public API of yscv-model/pipeline/video/onnx and restore rustc-hash as a dependency of the three crates, for one consistent map type across the codebase (the original #16 design). Speed is unchanged: the reverted maps were all cold boundaries (once-per-inference outputs, per-epoch metrics, parsed-once configs), so this is purely a consistency choice. Also migrate the metal-backend examples that still built std HashMap inputs (yolo_vis, debug_metal_yolo, bench_metal_yolo, bench_metal_vball, bench_mpsgraph) to FxHashMap with the proper import. Several were already broken under the macOS metal-backend CI job (E0433 missing import / E0308 once run_onnx_model's parameter is a concrete FxHashMap again).
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
2 participants
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.
std::collectionshashling algo is cryptographic and randomised; we do not need this. Let's migrate to rustc'sFxHasher, since it is cheaper in terms of compute