Skip to content

Pull requests: NVIDIA/Fuser

New pull request New
123 Open 2,772 Closed
123 Open 2,772 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
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Resize indexing misc bug fixes
#3632 opened Dec 22, 2024 by naoyam • Draft
1
[WIP] C++ layer to implement Triu
#3631 opened Dec 21, 2024 by protonu • Draft
Limit unrolling of all circular buffered loops to depth equal to prefetch
#3627 opened Dec 20, 2024 by jacobhinkle • Draft
5
Ring Allgather + GEMM Overlap HostIR Implementation Multi-GPU
#3626 opened Dec 20, 2024 by nsarka Loading…
@nsarka
2
Resolve conflicts by recomputation rope
#3625 opened Dec 20, 2024 by naoyam Loading…
5
Remove unnecessary NVF_API from scheduler/utils.h
#3623 opened Dec 19, 2024 by wujingyue Loading…
2
cacheInputs propagates allocation only for matmul schedulers.
#3621 opened Dec 19, 2024 by wujingyue Loading…
1
3
Add support for smem_epilogue when mma output is not cast to half
#3620 opened Dec 19, 2024 by protonu Loading…
23
Support outer reduction scheduler with SOL autotuning Autotune Generate heuristics through machine learning models.
#3618 opened Dec 19, 2024 by rdspring1 • Draft
1
Fix matmul incorrect results when k dim for CTA tile is a multiple of 16 Matmuls
#3616 opened Dec 19, 2024 by rdspring1 • Draft
5
Use tensor_options which specifies the right GPU.
#3615 opened Dec 18, 2024 by wujingyue Loading…
1
Host IR: make stream synchronization non blocking
#3608 opened Dec 18, 2024 by samnordmann Loading…
6
Lower distributed matmul to pipelined algorithm for fine-grained overlap
#3606 opened Dec 18, 2024 by samnordmann Loading…
2 tasks
4
Host IR: add GetCurrentStream
#3605 opened Dec 18, 2024 by samnordmann Loading…
4
Support TMA with 64-bit indexing
#3599 opened Dec 16, 2024 by jacobhinkle Loading…
1
2
Accept Hopper matmuls and update default heuristic
#3579 opened Dec 12, 2024 by jacobhinkle • Draft
3
[wgmma] Insert commit_group and wait_group after mma_async Matmuls
#3573 opened Dec 11, 2024 by jacobhinkle • Draft
1
1
[do not merge] nsight compute for #3563 Matmuls
#3568 opened Dec 11, 2024 by zasdfgbnm • Draft
[do not merge] Register stealing Matmuls
#3566 opened Dec 11, 2024 by zasdfgbnm • Draft
1
[Do not merge] Pull elect-sync predicate out Matmuls
#3564 opened Dec 11, 2024 by zasdfgbnm • Draft
[do not merge] Enable smem-epilogue and manual fix sync Matmuls
#3563 opened Dec 11, 2024 by zasdfgbnm • Draft
1
[Do not merge] fix-sync Matmuls
#3560 opened Dec 10, 2024 by zasdfgbnm • Draft
3
[Do not merge] Dump a kernel and run profiling on current main
#3559 opened Dec 10, 2024 by zasdfgbnm • Draft
toy rope benchmark
#3551 opened Dec 10, 2024 by jjsjann123 Loading…
rope_benchmark
#3550 opened Dec 10, 2024 by jjsjann123 Loading…
1 task done
23
ProTip! no:milestone will show everything without a milestone.