-
Notifications
You must be signed in to change notification settings - Fork 644
Pull requests: tile-ai/tilelang
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
[Transform] Replace CPU fallback thread placeholder with a constant-zero logical thread index
#2718
opened Jul 20, 2026 by
penguin-wwy
Contributor
Loading…
[BugFix] Pack 32-lane 8-bit CUDA vectors correctly
#2701
opened Jul 19, 2026 by
xiaoyouPREG
Collaborator
Loading…
[Bugfix] Emit st.bulk destination as a shared write to fix missing barrier and potential compilation-introduced races
#2700
opened Jul 19, 2026 by
arxgy
Loading…
[BugFix] Align multi-version shared buffer strides for TMA
#2694
opened Jul 18, 2026 by
UnsettingGalaxy
Collaborator
Loading…
[BugFix] Skip warp specialization for pipeline loops inside while loops
#2692
opened Jul 17, 2026 by
zyy3077
Collaborator
Loading…
[BugFix] Gate stochastic FP4/FP8 casts on sm_100a
#2691
opened Jul 17, 2026 by
xiaoyouPREG
Collaborator
Loading…
[BugFix] Skip TMA promotion for copies whose base pointer is bound inside the kernel body
#2687
opened Jul 17, 2026 by
Lyscoria
Collaborator
Loading…
[BugFix] Fallback non-16B cluster bulk copies
#2683
opened Jul 16, 2026 by
UnsettingGalaxy
Collaborator
Loading…
[Bugfix] Guard TMA copies against unaligned box offsets
#2681
opened Jul 15, 2026 by
UnsettingGalaxy
Collaborator
Loading…
[BugFix] Fix wrong offset when scanning a non-zero-offset buffer sub-region
#2680
opened Jul 15, 2026 by
li-ruinan
Collaborator
Loading…
[BugFix] Fix T.pow/T.power for constant integer exponent y <= 0
#2677
opened Jul 15, 2026 by
Hughshine
Contributor
Loading…
[Example][Opt] deepseek_v32 topk_selector kernel memory access optimization (~1.9× faster)
#2659
opened Jul 14, 2026 by
mengmeexix
Contributor
Loading…
[BugFix][Transform] Never classify side-effecting binds as replayable (atomics re-executed at every use site since v0.1.11)
#2651
opened Jul 13, 2026 by
zkyue
Loading…
[BugFix] Use blockDim as workspace stride in batch AllReduce
#2621
opened Jul 11, 2026 by
ColmaLiu
Collaborator
Loading…
[BugFix] Pass buffer row stride to 2D scan kernel to fix silent miscomputation
#2620
opened Jul 11, 2026 by
ColmaLiu
Collaborator
Loading…
[BugFix] Correct IEEE math intrinsic names for fp64/fp16/bf16
#2619
opened Jul 11, 2026 by
ColmaLiu
Collaborator
Loading…
[BugFix] Mark bare data vars passed to call_extern as written
#2618
opened Jul 11, 2026 by
Chennesxu
Contributor
Loading…
[BugFix] Reject non-power-of-two AllReduce widths
#2611
opened Jul 11, 2026 by
zyy3077
Collaborator
Loading…
[Example][DeepSeek-V3.2] Adaptive threads for sparse MLA backward
#2592
opened Jul 10, 2026 by
Butterfingrz
Loading…
[BugFix] Guard T.atomic_addx4 return type for sliced destinations
#2590
opened Jul 10, 2026 by
Lyscoria
Collaborator
Loading…
[BugFix] Avoid unsafe MergeIfStmt across condition writes
#2557
opened Jul 8, 2026 by
Chennesxu
Contributor
Loading…
[Feature] Enhance blockscaled gemm support for sm100 and examples
#2546
opened Jul 7, 2026 by
Rachmanino
Collaborator
•
Draft
1 of 4 tasks
[FEATURE] Add block-causal attention for dLLM example
#2499
opened Jun 30, 2026 by
perkyfever
Loading…
Previous Next
ProTip!
no:milestone will show everything without a milestone.