COMMITS
August 5, 2026
L
[Compiler][Z3] Isolate analyzer contexts per kernel compilation (#2890)
Lei Wang committed
G
[Cleanup] Replace root reproducers with CPU regression coverage (#2878)
Gengyuan Bai committed
A
[CI] Re-enable ROCm CI on a gfx942 runner (#2874)
andyluo7 committed
J
[BugFix] Handle vectorized SelectNode in codegen_cuda (#2843)
Jinpeng Wang committed
A
[BugFix][ROCm] Emit a compiler barrier for tl.sync_warp on HIP (#2872)
andyluo7 committed
C
[TIR][Python] Split backend-specific op proxies (#2870)
Chaofan Lin committed
August 4, 2026
A
A
[Examples] Gate CUDA-only and flash_attn-dependent example tests (#2864)
andyluo7 committed
B
[Doc] Fix wrong loop bound in FlashAttention README example (#2868)
bing zhou committed
L
[Debug][TIR] Drive Pass Visualizer with PassInstrument (#2866)
Lei Wang committed
J
[BugFix] Use masked warpsync in in-warp allreduce (#2865)
jpwang committed
Y
[Refactor] Recycle `T.unroll(explicit=True)` for early explicit unrolling (#2859)
Yongqi Zhuo committed
M
[Typo] Correct source spelling errors (#2858)
morluto committed
Y
[BugFix][Metal] Resolve auto backend to torch and skip disk cache for torch (#2856)
Yichen Yan committed
L
[Docs][LSP] Announce the open-source TileLang LSP (#2862)
Lei Wang committed
P
[CI] [pre-commit.ci] autoupdate (#2861)
pre-commit-ci[bot] committed
Y
[CUDA] Adopt multi-staged buffers in examples (#2836)
Yongqi Zhuo committed
August 3, 2026
L
[CUDA][ROCm][Metal] Split backend-specific builtin ops (#2855)
Lei Wang committed
C
[TIR][Python] Trim redundant op proxy wrappers (#2850)
Chaofan Lin committed
L
[CUDA][TMA] Separate atomic-add dtype support from layout encoding (#2846)
Lei Wang committed
J
[BugFix] Fix absmax and abssum for uint dtypes (#2845)
jpwang committed
K
[Enhancement] Disable data race check by default, opt-in via env var (#2851)
KellyFrog committed
L
[Docs] Refresh README news and onboarding (#2849)
Lei Wang committed
T
L
[Release] Bump versin into 0.1.13 (#2826)
Lei Wang committed
August 1, 2026
Y
[BugFix] Correctly preserve loop step when unrolling loops (#2835)
Yongqi Zhuo committed
Y
Revert "[BugFix] Preserve loop step when unrolling loops" (#2834)
Yongqi Zhuo committed
M
[BugFix] Reject unsupported TMA atomic add dtypes (#2830)
morluto committed
July 31, 2026
P
[Enhancement] Aggregate VerifyParallelLoop race diagnostics with span (#2806)
penguin_wwy committed