COMMITS
August 12, 2026
P
[CPU] Support atomic ops on CPU (#2941)
penguin_wwy committed
August 11, 2026
D
[Enhancement] Expose cluster_mask on T.tma_copy (#2932)
Diao Han committed
D
[BugFix][Hopper] Select the widest legal WGMMA N instead of gcd (#2931)
Diao Han committed
C
[CUDA] Add pre-SM80 fallback for bf16 atomic add (#2938)
Chennes committed
Y
[Feature] Warp specialization schedules and materialization (#2892)
Yongqi Zhuo committed
T
R
[BugFix][CUDA] Provide htan overloads for fp16/bf16 tangent (#2894)
Ruihan11 committed
C
[Refactor][BugFix] Refactor the loop vectorization plan with ConstraintKind (#2935)
Chaofan Lin committed
L
[JIT][FFI] Allocate kernel outputs through the packed API (#2937)
Lei Wang committed
August 10, 2026
E
C
[Cherry-Pick][BugFix] Fix get_parent_locals frame self-reference leak (#2934)
Chaofan Lin committed
L
[Debug][TIR][JIT] Unify pass instrumentation per compilation (#2923)
Lei Wang committed
R
G
[ROCm] Resolve versioned device properties in HIP stub (#2919)
guanjiawei committed
D
[CI]: Bump pypa/cibuildwheel from 4.1 to 4.2 (#2930)
dependabot[bot] committed
August 9, 2026
M
[Fix] Discover flat CUDA includes for NVRTC (#2829)
morluto committed
August 8, 2026
A
[BugFix][ROCm] Reject sub-wavefront block sizes instead of crashing (#2918)
andyluo7 committed
Y
[CuTeDSL] Port backend intrinsics to CUTLASS DSL primitives (#2871)
Yuxi Chi committed
P
[Enhancement] Show kernel name in pass timing report (#2905)
penguin_wwy committed
August 7, 2026
A
[BugFix][Carver] Parse lettered SM arch strings in check_sm_version (#2891)
Aditya Singh committed
C
[Testing] Run portable regression tests on auto targets (#2914)
Chaofan Lin committed
A
[CI] Use stable torch for the ROCm leg (#2910)
andyluo7 committed
M
[BugFix][ROCm] Lower vector Select conditions lane-wise (#2889)
morluto committed
A
[Testing] Gate the issue #2883 regression test on CUDA (#2908)
andyluo7 committed
August 6, 2026
C
[Docs] Define backend architecture and integration skill (#2900)
Chaofan Lin committed
P
[CPU] Support reduce ops on CPU (#2893)
penguin_wwy committed
C
[Test] Add regression test for issue #2883 (#2899)
Chaofan Lin committed
K
[BugFix] Deduplicate replicated reducer updates (#2881)
KellyFrog committed
A
[Doc] ROCm CI is no longer disabled (#2896)
andyluo7 committed
August 5, 2026
C
[Backend] Add unified backend resolution policy (#2318)
Chaofan Lin committed