Repository navigation
1bit comfy: run the checked binary; fetch the submodule when missing - #66
Conversation
…fetch the submodule when missing PR-Agent review on #65: the binary is opened once, checked through the descriptor (regular, executable, not writable by others) and run with fexecve, so no swap between check and exec; ONEBIT_COMFYUI must be absolute and relative PATH entries are skipped. CMake and build-comfyui.sh fetch third_party/comfyui.cpp themselves when it is not initialised. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Docs7 for 1bit-monster/engine
Commit |
PR Reviewer Guide 🔍(Review updated until commit 9a5ad41)Here are some key observations to aid the review process:
|
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Persistent review updated to latest commit 9a5ad41 |
…512 tokens, deterministic flash attention llama.cpp fork since cebcd70 (balanced mode, figures from each PR): - #66 MUL_MAT_ID at multiples of 32 plus a decode-loader stride fix; #67 ADD_ID and SWIGLU_OAI on HRX; #73 a placement guard for the CPU/HRX split bug (engine #286). gpt-oss-20b MXFP4: pp512 25.8 -> ~1000, tg128 12.6 -> ~35 tok/s, text correct, KLD vs CPU 0.029. - #69 TQ1_0/TQ2_0 on HRX: Ternary-Bonsai-1.7B KLD vs CPU 0.000523; pp512/tg128 3542/113 and 4100/156. - #70 MLA V transpose: GLM-4.7-Flash prompts of 512+ tokens gave garbage (PPL 315,664), now 5.916 (CPU 5.959). - #71 llama-hadamard folds qwen3next's ssm_ba and refuses unfoldable stamped files. - #72 masked flash-attention keys reach P*V as V = +0: identical requests now give identical logits (Qwen3-0.6B and Qwen3.8-27B bit-identical repeats); pp512 -4.7% on Qwen3-0.6B. Docs: docs/hrx.md "Our patches". Registry regenerated (no mapping changes). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…tic FA, attention sinks) (#295) * Pin llama.cpp 2bd7f58: gpt-oss on HRX, TQ1_0/TQ2_0, MLA prompts past 512 tokens, deterministic flash attention llama.cpp fork since cebcd70 (balanced mode, figures from each PR): - #66 MUL_MAT_ID at multiples of 32 plus a decode-loader stride fix; #67 ADD_ID and SWIGLU_OAI on HRX; #73 a placement guard for the CPU/HRX split bug (engine #286). gpt-oss-20b MXFP4: pp512 25.8 -> ~1000, tg128 12.6 -> ~35 tok/s, text correct, KLD vs CPU 0.029. - #69 TQ1_0/TQ2_0 on HRX: Ternary-Bonsai-1.7B KLD vs CPU 0.000523; pp512/tg128 3542/113 and 4100/156. - #70 MLA V transpose: GLM-4.7-Flash prompts of 512+ tokens gave garbage (PPL 315,664), now 5.916 (CPU 5.959). - #71 llama-hadamard folds qwen3next's ssm_ba and refuses unfoldable stamped files. - #72 masked flash-attention keys reach P*V as V = +0: identical requests now give identical logits (Qwen3-0.6B and Qwen3.8-27B bit-identical repeats); pp512 -4.7% on Qwen3-0.6B. Docs: docs/hrx.md "Our patches". Registry regenerated (no mapping changes). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * Pin llama.cpp 4485916: attention sinks on HRX (gpt-oss) #68 runs gpt-oss's sink logits on HRX as an exact post-correction of the flash-attention output. Docs: docs/hrx.md "Our patches". Registry pin updated. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * Pin llama.cpp f5b7f4a: PrismML tile bytes (PQ2_0 small-model decode) #74: the low-token SwiGLU read every PQ2_0 / PTQ1_0 row from row 0; Ternary-Bonsai-1.7B PQ2_0 now matches the CPU. Found by the release format matrix. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: bong-water-water-bong <bong-water-water-bong@1bit.gg> Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Follow-up to #65: fixes both things PR-Agent flagged there (#65 was merged before this commit went up).
Running the binary safely:
1bit comfyopens the binary once and checks it through that open file: a regular file, executable, and not writable by other users.fexecve, so nothing can swap it between the check and the run.ONEBIT_COMFYUImust be an absolute path, and relativePATHentries are skipped.Tested on strixhalo:
Submodule: CMake (
-DONEBIT_COMFYUI=ON) andscripts/build-comfyui.shfetchthird_party/comfyui.cppthemselves when it's missing. In a fresh clone with no submodules, configuring fetched it at the pinned4a08238.🤖 Generated with Claude Code