dylanyunlong

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-14 15:10:54 +08:00

d32822c5d2 prebuilt: corex_moe_index_combine.so + ix_full_bridge.so (BI-V100 compiled)

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-14 15:02:54 +08:00

d9ffc5159d fix(bridge): rewrite ix_full_bridge.cpp for real ixformer_torch_ext namespace

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-14 14:56:50 +08:00

9e3157b444 fix(P0): extra=allow + topk_softmax fallback + deploy_local.sh + SO chain verify

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-14 14:54:01 +08:00

101db8774c diag: probe_bi100.py — run on real machine, paste output back

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-14 14:14:28 +08:00

07fad2cce7 fix: max-num-seqs=1 + pure Q-tiling — match a3c45d3b exactly

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-14 11:59:50 +08:00

1f69311375 fix: flash_attn import path ixformer.contrib → ixformer.functions

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-14 11:52:57 +08:00

4c1adc11db fix: restore libcccl_allocator.so + minimal .dockerignore + clean pycache

dfaaae988e diag: test_dlopen_chain.py — verify all .so and ixformer dispatch on BI-V100

336f3349ca fix(submit): restore flash_attn prefill + all 38eca5c2 improvements

67a5639c3c fix(submit): restore a3c45d3b-proven config + safe improvements

Compare 4 commits »

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-14 11:08:31 +08:00

bce79e44be perf(prefill): restore flash_attn_varlen_func with profiling safety guard

74ce61712b fix(build): update prebuilt CoreX artifact count 13→14 — SHA256SUMS has 14 entries after corex_gdn_chunk_recurrent.so was added

38eca5c26a revert: restore a3c45d3b yaml + Q-tiling + remove all OOM hacks

Compare 3 commits »

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-14 10:28:51 +08:00

eb57eb7d1c clean: remove 444 .pyc files + libcccl_allocator.so from git tracking

dylanyunlong created repository dylanyunlong/8215_linux

2026-08-14 10:24:19 +08:00

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-14 10:14:20 +08:00

41b51382fd prebuilt: corex_gdn_chunk_recurrent.so + remove __pycache__ from tracking

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-14 10:10:56 +08:00

9a52f05783 prebuilt: add corex_gdn_chunk_recurrent.so with fixed pybind kwargs

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-14 10:07:20 +08:00

bb0de83d45 fix(pybind): c10::optional<Tensor>() default — c10::nullopt not registered in pybind

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-14 10:03:46 +08:00

768d89c31a fix(pybind): add py::arg + defaults to corex_gdn_chunk_recurrent

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-14 09:51:47 +08:00

a4d16d36b8 fix(build): std::nullopt → c10::nullopt for CoreX c10::optional

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-14 09:49:54 +08:00

56fe58ada3 fix(build): c10::optional for CoreX CUDA 10.2 — std::optional incompatible

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-14 09:40:33 +08:00

1d9b620416 fix(crash): disable corex_gdn_chunk_recurrent — pybind signature mismatch

716034bdd0 fix(OOM): lower blocks cap 5000→3000 — flash_attn needs ~4GB temp buffer

Compare 2 commits »

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-14 09:26:17 +08:00

c8a982c4e8 feat: ix_full_bridge.so — dlopen bridge for ixformer::infer C++ API

872be0effa fix(build): strip \r\n from all .py files — CRLF breaks patch_ops.sh text matching

Compare 2 commits »

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-14 09:00:59 +08:00

aa4b4992d1 fix(build): hardcode blocks cap 5000 in .py — remove yaml env var

456380eed0 fix(OOM): cap GPU blocks at 5000 via BI100_MAX_GPU_BLOCKS env var

c6aa1b9c62 fix(P0): protocol.py extra=allow — recover 180 rejected replay requests

20aac5b212 fix(OOM): return zeros during profiling — skip both flash_attn AND Q-tiling

048302bd4a Revert "fix: remove --enable-chunked-prefill — conflicts with small max_num_batched_tokens"

Compare 6 commits »

dylanyunlong pushed to main at dylanyunlong/project_6

2026-08-14 00:14:21 +08:00

e31bd69779 fix(OOM): max-num-batched-tokens 4096→256 — profiling Q-tiling also OOMs