-
Notifications
You must be signed in to change notification settings - Fork 1.2k
Pull requests: deepseek-ai/DeepGEMM
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
[Bug] Fix bench_kineto returning zero on the first call in a process
#426
opened Sep 2, 2026 by
DiegoCao
Loading…
docs: note Mega MoE requires PyTorch >= 2.9 in Requirements
#424
opened Aug 28, 2026 by
pengfaming
Loading…
fix: make pack_ue8m0_to_int CUDA-graph capture safe (#414)
#421
opened Aug 26, 2026 by
XFDG
Loading…
[SM12x] Restore pure-fp8 1d1d path (removed in 8b1392b); fix kernel TU preamble
#419
opened Aug 26, 2026 by
maci0
Loading…
[SM100][MegaMoE] Fuse and publish Kimi K3's BF16 shared expert
#416
opened Aug 24, 2026 by
gcanlin
Loading…
Fix SM90 k-grouped GEMM hang and silent corruption when TMA multicast clusters straddle group boundaries
#415
opened Aug 24, 2026 by
shoemoney
Loading…
fix(layout): validate SF tensors are CUDA; add MXFP4 (1, 32) SF transform test
#407
opened Aug 14, 2026 by
qqtang-code
Loading…
fix: SM90 paged mqa
prefix_sum out of range (CUDBG_EXCEPTION_WARP_OUT_OF_RANGE_ADDRESS)
#399
opened Aug 5, 2026 by
cjackal
Loading…
Make the JIT cache key independent of the install include path
#398
opened Aug 5, 2026 by
matteso1
Loading…
Don't enable the assertion check for release build
#390
opened Jul 22, 2026 by
zhangqingshan373
Loading…
Optimize sm100 version with token-aware BM240 tiling, native UE8M0, and direct combine
#386
opened Jul 17, 2026 by
qinqinwo
Loading…
Add B200-Calibrated Sparse-Routing Adaptive Wave Sizing
#381
opened Jul 15, 2026 by
qinqinwo
Loading…
Previous Next
ProTip!
Find all pull requests that aren't related to any open issues with -linked:issue.