refactor: remove dead-end GPU experiments and clean up exec plans#115
Merged
ChrisLundquist merged 1 commit intomasterfrom Mar 10, 2026
Merged
refactor: remove dead-end GPU experiments and clean up exec plans#115ChrisLundquist merged 1 commit intomasterfrom
ChrisLundquist merged 1 commit intomasterfrom
Conversation
Remove confirmed dead-end code that wastes future agent time: - Delete lz77_hash.wgsl GPU hash-table kernel and all Rust references (fundamental SIMT atomic ordering problem, 6.25% match ratio vs 99.61% CPU) - Delete parlz module, pipeline variant, WebGPU integration, and CLI flag (37.6% ratio gap vs serial greedy, NO-GO for fully-parallel GPU LZ) - Disable gpu_fused_span() routing entropy to GPU for Lzr/LzSeqR (GPU rANS is 0.54-0.77x CPU after 29+ optimization iterations) Update exec plan statuses: - PLAN-p0a-gpu-rans-vertical-slice: CLOSED (perf gate failed) - PLAN-interleaved-rans: PARKED (Phase D cancelled, GPU rANS dead end) - PLAN-unified-scheduler-north-star: PARKED (Phases 2+5 blocked) - TODO-huffman-sync-decode: PARKED (zero implementation progress) - agent-harness-implementation: PARKED (Phase 1 done, rest deferred) Add benchmark caveat to CLAUDE.md documenting ~18x gap between bench.sh (per-process, includes wgpu init) and Criterion (pre-allocated). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
ChrisLundquist
force-pushed
the
claude/happy-hodgkin
branch
from
d6f1e7b to
345e190
Compare
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
gpu_fused_span()routing entropy to GPU for Lzr/LzSeqR — GPU rANS is 0.54–0.77x CPU after 29+ optimization iterations29 files changed, +53 / -1,596 lines. All removed code was confirmed dead-end with documented performance data.
Test plan
cargo fmt --checkpasses cleancargo clippy --all-targetspasses clean (zero warnings)cargo test --no-default-features— 571 passed, 0 failed🤖 Generated with Claude Code