Skip to content

Pull requests: NVIDIA/Fuser

New pull request New
112 Open 1,927 Closed
112 Open 1,927 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

Rename rfactor -> producer projection or logical
#2612 opened Jul 16, 2024 by zasdfgbnm • Draft
1
Make setShardedAllocationDomain a preseg pass.
#2611 opened Jul 16, 2024 by wujingyue Loading…
1
Make insertShardedAxisReordering a preseg pass.
#2610 opened Jul 16, 2024 by wujingyue Loading…
1
Avoid Reduction/Broadcast domains in ValGraph scheduling
#2609 opened Jul 16, 2024 by jacobhinkle Loading…
4
Allow alias analysis mark candidate on segmented fusion inputs
#2608 opened Jul 16, 2024 by jjsjann123 • Draft
5
Make InsertReshardings a preseg pass.
#2606 opened Jul 16, 2024 by wujingyue Loading…
1
Unit test to try schedule the store of the Mma output from shared to global memory using TMA.
#2605 opened Jul 16, 2024 by protonu Loading…
7
Root rename
#2602 opened Jul 16, 2024 by zasdfgbnm • Draft
fix replay pos in TransformReplay::replayCasP
#2595 opened Jul 15, 2024 by liqiangxl Loading…
1
3
[Do not merge] Sharded transformer forward pass
#2590 opened Jul 13, 2024 by cowanmeg • Draft
[WIP] SDPA Bwd Mapping
#2589 opened Jul 12, 2024 by Priya2698 • Draft
2
set vectorization factor for each IO tensor
#2586 opened Jul 12, 2024 by liqiangxl • Draft
[Tutorial Multidevice]: communicator, DeviceMesh, Simple pipelining
#2585 opened Jul 12, 2024 by samnordmann Loading…
18
Host irs: Reduce-Scatter based pipelining
#2584 opened Jul 12, 2024 by samnordmann • Draft
skip further checks in reductionInterferingView if non of the view ops has ID merges
#2583 opened Jul 12, 2024 by liqiangxl • Draft
1
Print torch/nvfuser versions in python error repro
#2578 opened Jul 11, 2024 by jacobhinkle Loading…
3
FusionExecutorCache runs a communication-only fusion.
#2575 opened Jul 11, 2024 by wujingyue Loading…
15
Use force_iter_type in broadcast cases
#2568 opened Jul 10, 2024 by Priya2698 • Draft
3
Use _reduce_scatter_base instead of reduce_scatter when possible
#2562 opened Jul 10, 2024 by samnordmann Loading…
3
Host irs: add support for MatmulOp
#2546 opened Jul 8, 2024 by samnordmann Loading…
7
Check ATen-evaluated tensor sizes
#2534 opened Jul 5, 2024 by jacobhinkle Loading…
8
Apply AbstractTensor schedule to TensorView
#2512 opened Jul 2, 2024 by jacobhinkle Loading…
99
[Hopper matmul] Refactor the code which schedules TMA loads for the input Operands
#2508 opened Jul 1, 2024 by protonu Loading…
24
preseg optimization pass - merge neighboring PadOp
#2490 opened Jun 28, 2024 by jjsjann123 Loading…
4
[DO NOT MERGE] Python API for SdpaFwdOp
#2483 opened Jun 27, 2024 by Priya2698 • Draft
ProTip! What’s not been updated in a month: updated:<2024-06-16.