dylanyunlong

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-07 15:46:18 +08:00

e0344b1730 fix(critical): disable thinking for tool_call requests — fixes d03_tool_call FAIL

812c374f7a fix(critical): sync baseline.muh max_model_len=100000 gpu_mem=0.90 — match computility-run.yaml

19879ccaae docs: pipeline ground truth — CCCL parity audit, SM100 benchmark data, injection status

611c491b0f audit: CCCL vs muh parity check — scale_mem_bound 11/11 PASS, SM100 benchmark values extracted

95d03147e7 fix(deploy): reduce max-model-len 131072→100000, remove chunked-prefill, single-seq — prevent OOM crash

Compare 7 commits »

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-07 15:10:41 +08:00

cbd1f08a3e fix(protocol): clamp max_tokens to available context — fix t3_max_tokens_max HTTP 400

16f0b30d2e fix(critical): 3 bugs causing 0.0 score — OOM crash + thinking param + multimodal

539fe7745b fix(protocol): accept max_completion_tokens + thinking + tool_calls messages

dd077e1272 [ENGINE] CCCL dispatch_common.cuh use_default pattern: accept max_completion_tokens + thinking + content=None

8002900af0 [ENGINE] CCCL kernel_scan.cuh dual-algorithm dispatch: v1 (online norm) vs v2 (deferred norm)

Compare 6 commits »

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-07 14:22:01 +08:00

b86a121d6d fix(critical): patch_ops.sh cwd bug — all cp/deploy ./paths resolved against Dockerfile WORKDIR=/workspace/ instead of script dir

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-07 13:33:14 +08:00

36e67c00b0 fix(critical): deploy patches to BOTH lib and lib64 vllm paths

e9eaad0592 perf: increase prefix attention tile budget 96MB→256MB

025059d78e fix(critical): raise decode threshold to prevent service crash

Compare 3 commits »

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-07 12:39:04 +08:00

ca3697f4b0 [ENGINE] CCCL SmemResource pattern: pre-allocate staging buffers in V2 paged attention

bf1cccb750 refactor(moe): apply CCCL GridEvenShare + dispatch_batch_memcpy to BLOCK_SIZE_M

0ba2221d9c docs: CCCL → vllm kernel pattern mapping for competition

9907c9b8ee docs: comprehensive CCCL vs muh gap analysis with competition priority

68d500c960 test: add scan tuning verification — union SMEM model from agent_scan.cuh

Compare 48 commits »

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-06 11:14:46 +08:00

bf5d19991c [FIX] qwen3_5.py: replace solve_triangular with manual forward substitution

b4803c3259 [BASE] qwen3_6_scripts/sampler.py: CCCL topk unsorted output optimization

f59d30dcb2 [BASE] qwen3_6_scripts/paged_attn.py: CCCL shifted_output defensive init

8056641f08 [BASE] qwen3_6_scripts/xformers.py: CCCL block_load_to_shared pre-alloc pattern

Compare 4 commits »

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-06 10:48:50 +08:00

2d1588d261 [BASE] qwen3_6_scripts/sampler.py: CCCL dispatch_topk DoubleBuffer pattern

e784910d47 [ENGINE] Pattern 7: CCCL C API JIT → Triton autotune mapping

da553227e9 [BASE] qwen3_6_scripts/verify_functional.py: add CCCL-derived boundary tests

6c472d640f [ENGINE] CCCL system-level patterns → BI-V100 engine module

6148e03bc7 [BASE] vllm/core/evictor_v2.py: CCCL bucket_sort2d design pattern annotation

Compare 9 commits »

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-06 10:25:32 +08:00

11cbc00cf2 [DOCS] CCCL benchmark reference: 199 annotations from 27 tuning files extracted

dedf08166a [CCCL] Add missing CCCL components: c2h, nvbench_helper, cmake, cudax, AGENTS.md

b0d597363a [BUGFIX] qwen3_6_scripts/model_runner.py: fix max_decode_seq_len (deployment version)

dac9aa46f5 [BUGFIX] vllm/worker/model_runner.py: fix max_decode_seq_len passed as max_encoder_seq_len

29f119c094 [ENGINE] vllm/attention/ops/paged_attn.py: CCCL block_reduce_raking V1/V2 dispatch

Compare 175 commits »

dylanyunlong created repository dylanyunlong/project_6

2026-08-06 09:12:49 +08:00