Skip to content

Pull requests: tile-ai/tilelang

New pull request New
25 Open 1,753 Closed
25 Open 1,753 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

[CUDA] Align swizzled TMA shared buffers
#2391 opened Jun 12, 2026 by LeiWang1999 Member Loading…
6
Add assert that block_K is a multiple of micro_size_k in CUDA MMA GEMM backends to prevent silent miscompilation.
#2390 opened Jun 12, 2026 by Federicorao Loading…
1
2
[WIP] [DO NOT REVIEW] [Backend] Support TMA lowering for arbitrary (swizzled) SMEM layout
#2380 opened Jun 11, 2026 by Yongqi-Zhuo Collaborator • Draft
6
[feat] add pass diff show for debugging
#2375 opened Jun 11, 2026 by erhsh Contributor Loading…
15
[Autotune] Stabilize autotune benchmarking across devices and fix the unconsistency in the autotuning process.
#2370 opened Jun 11, 2026 by Wazrrr Contributor Loading…
4
[BugFix][CuTeDSL] Fix TileKernels scan, optional-shape, and e5m6 paths
#2369 opened Jun 10, 2026 by JayceSu98 Contributor Loading…
8
[BugFix][Transform] Demote illegal-width residual cp.async to a synchronous copy
#2366 opened Jun 10, 2026 by Hughshine Contributor Loading…
1
7
[CUDA] Add SM120 NVF4 block-scale MMA support
#2364 opened Jun 9, 2026 by qqq-tao Contributor Loading…
7
[Feature] Flash Attention 1SM/2SM kernels with features supported
#2360 opened Jun 9, 2026 by chengyupku Contributor Loading…
3 tasks done
17
[Fix][Pipeline] Prevent double expansion of shared buffers across sibling pipelines
#2342 opened Jun 5, 2026 by harelhuang Contributor Loading…
3 tasks done
1
11
[TileOP] Add SM70 GEMM FMA fallback
#2339 opened Jun 5, 2026 by cklxx Contributor Loading…
2
[AMD][ROCm] Fix CI failures on gfx950, gfx1100, gfx1151, and gfx1201
#2326 opened Jun 3, 2026 by zhangnju Collaborator Loading…
11
[Stacked][Feature] Support NVFP4 Gemm on Blackwell arch (SM100,110,120)
#2324 opened Jun 3, 2026 by Hale423 Contributor Loading…
24
[Draft]: add tile level cost model for tilelang
#2320 opened Jun 2, 2026 by yyttt6 Collaborator • Draft
2
[Backend] Add unified backend resolution policy
#2318 opened Jun 2, 2026 by SiriusNEO Collaborator Loading…
12
[Enhancement] Add T.tma_copy barrier_rank argument and fold use_2cta into LowerTileOp
#2316 opened Jun 1, 2026 by Rachmanino Collaborator Loading…
10
[Feature] Support unaligned barrier sync
#2295 opened May 28, 2026 by Rachmanino Collaborator Loading…
8
Add SM120 dense block-scaled MMA and essential support for SageAttention3
#2253 opened May 23, 2026 by sepcnt Contributor Loading…
16
[Metal] M5 Cooperative Tensor T.gemm
#2252 opened May 23, 2026 by oraluben Collaborator • Draft
2
[ROCM] Fix buffer_load_lds support for gfx950
#2248 opened May 22, 2026 by benenzhu Contributor • Draft
9
[Feature] Support Blackwell FP4(float4_e2m1fn) GEMM for SM100 & SM120
#2182 opened May 11, 2026 by Hale423 Contributor Loading…
21
[Feature][Blackwell] Add SM120 T.float4_e2m1fn FP4 GEMM support.
#2171 opened May 8, 2026 by TerminusAkivili Contributor Loading…
16
[feat] Add Hexagon HMX backend support
#2155 opened May 6, 2026 by Calaweh Loading…
1
18
[Feature] Support mutable TMA descriptor and canonicalize usage in examples
#2113 opened Apr 28, 2026 by Rachmanino Collaborator • Draft
2
improve build skill description for AI agent discoverability
#2101 opened Apr 25, 2026 by oraluben Collaborator • Draft
3
ProTip! Filter pull requests by the default branch with base:main.