Skip to content

Pull requests: NVIDIA/Fuser

New pull request New
Clear current search query, filters, and sorts
46 Open 337 Closed
46 Open 337 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

Introduce use_stmatrix parameter to MatmulParams
#4100 opened Mar 18, 2025 by jacobhinkle • Review required
1
4
Translate MmaOp patterns properly on Hopper
#4072 opened Mar 13, 2025 by jacobhinkle • Draft
12
[WIP] simple L2 model for setting grid swizzle and cta order
#4044 opened Mar 7, 2025 by jacobhinkle • Draft
1
Automatically save MatmulParams in extra_info in benchmarks
#4031 opened Mar 6, 2025 by jacobhinkle Loading…
9
Add nvfuser benchmark executor and unify test_matmul.py
#4021 opened Mar 6, 2025 by jacobhinkle • Draft
6
Remove MmaOp::AxisMapping
#3995 opened Feb 28, 2025 by jacobhinkle • Draft
3
Fix slice to empty not being converted to FullOp
#3882 opened Feb 13, 2025 by jacobhinkle Loading…
7
[WIP] Check that replaceOutput does not change alias behavior
#3839 opened Feb 6, 2025 by jacobhinkle • Draft
1
Set __launch_bounds__ in kernel whenever we are able
#3794 opened Jan 29, 2025 by jacobhinkle Loading…
20
[wgmma] Insert commit_group and wait_group after mma_async Matmuls
#3573 opened Dec 11, 2024 by jacobhinkle • Draft
1
9
Handle empty tensors during definition of cat
#3313 opened Oct 30, 2024 by jacobhinkle • Draft
9
Export DEBUG_SERDE=true in compare_codegen.sh
#3280 opened Oct 25, 2024 by jacobhinkle • Draft
1
[DO NOT MERGE] Testing matmul bench CI
#3163 opened Oct 11, 2024 by jacobhinkle • Draft
2
[WIP] Accept multi-matmul fusions for automatic scheduling
#3018 opened Sep 25, 2024 by jacobhinkle • Draft
1
Concretize resize
#2835 opened Aug 22, 2024 by jacobhinkle Loading…
1
26
Avoid Reduction/Broadcast domains in ValGraph scheduling
#2609 opened Jul 16, 2024 by jacobhinkle Loading…
4
Apply AbstractTensor schedule to TensorView
#2512 opened Jul 2, 2024 by jacobhinkle Loading…
101
[WIP DO NOT REVIEW] "Fold" operations to generalize Reductions
#2307 opened May 28, 2024 by jacobhinkle • Draft
[WIP] Check that matmul epilogue contains only pointwise ops Matmuls Segmentation Issues related to nvFuser Segmentation Top-Down Matmul Dev
#2225 opened May 9, 2024 by jacobhinkle • Draft
1
Don't reuse segmented kernel runtimes if we can avoid segmentation
#2163 opened Apr 30, 2024 by jacobhinkle • Draft
36
[matmul] Merge batch dims with BIDx tile dim Matmuls
#2140 opened Apr 24, 2024 by jacobhinkle • Draft
2
5
Reset semaphore in grid sync
#2081 opened Apr 16, 2024 by jacobhinkle • Draft
1
5
[DO NOT REVIEW] Introduce Features as a replacement for {Enable,Disable}Option
#2077 opened Apr 15, 2024 by jacobhinkle • Draft
Perform cancellation in SimplifyingIrBuilder::addExpr
#2020 opened Apr 2, 2024 by jacobhinkle Loading…
16
Propagate negative inequalities for 1 step
#1981 opened Mar 22, 2024 by jacobhinkle • Draft
ProTip! Exclude everything labeled bug with -label:bug.