Skip to content

Pull requests: NVIDIA/Fuser

New pull request New
174 Open 4,024 Closed
174 Open 4,024 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

Relaxing the exactness requirement of the greedy scheduler greedy scheduler
#5176 opened Sep 17, 2025 by naoyam Loading…
1
Allow ScatterOp with multiple dimensions as long as extents are the same greedy scheduler
#5175 opened Sep 17, 2025 by naoyam Loading…
4
Support layout op in scheduler
#5174 opened Sep 17, 2025 by jjsjann123 • Draft
3
Allow broadcast and squeeze in the greedy scheduler greedy scheduler
#5171 opened Sep 17, 2025 by naoyam Loading…
1
4
Patch allocation logic to produce outputs with correct logical size
#5170 opened Sep 17, 2025 by jjsjann123 Loading…
9
Migrate Tutorial.Memcpy to direct bindings Direct Bindings Python extension with direct mapping to NvFuser CPP objects. Python API Issues related to the Python API
#5169 opened Sep 17, 2025 by rdspring1 Loading…
6
Add support for KernelExecutor to direct bindings Direct Bindings Python extension with direct mapping to NvFuser CPP objects. Python API Issues related to the Python API
#5168 opened Sep 17, 2025 by rdspring1 Loading…
4
Replay loop and allocation independently in segmentor
#5167 opened Sep 16, 2025 by Priya2698 • Draft
3
[SMOKE TEST] Reproduce failure in jit_python_tests_20_B100_TNVF
#5164 opened Sep 16, 2025 by rdspring1 Loading…
6
Modernize nvFuser to use C++20 std::ranges more widely (Fixes #4571)
#5161 opened Sep 13, 2025 by varunjain3 Loading…
1
1
[EXPERIMENT] A runtime function to quantize to NVFP4
#5148 opened Sep 10, 2025 by protonu • Draft
2
[SmokeTest][DoNotReview]
#5143 opened Sep 9, 2025 by jjsjann123 • Draft
2
Introduce Cutlass scheduler and CutlassExecutor
#5142 opened Sep 9, 2025 by jacobhinkle Loading…
10
Convert python benchmarks to direct bindings except benchmark_overlap.py Direct Bindings Python extension with direct mapping to NvFuser CPP objects. Python Benchmarks
#5139 opened Sep 9, 2025 by rdspring1 Loading…
3
Fix TensorDomain::setContiguity
#5125 opened Sep 5, 2025 by zasdfgbnm Loading…
4
Add a repro
#5124 opened Sep 5, 2025 by wujingyue • Draft
@wujingyue
2
[DoNotReview] Use initial loop domain in domain_map
#5122 opened Sep 5, 2025 by jjsjann123 • Draft
4
Add unsafe_reduce presegmentation pass
#5121 opened Sep 5, 2025 by tbqh • Draft
34
Creating a CuTe TV Layout in NvFuser Cutlass Matmuls
#5117 opened Sep 4, 2025 by rdspring1 Loading…
3 tasks
4
Add runAndValidate() to wrap scheduleAndRun() and testValidate() pattern
#5113 opened Sep 4, 2025 by tbqh Loading…
3
max.fp16
#5111 opened Sep 4, 2025 by liqiangxl • Draft
1
P2PCommunication Benchmark over Cuda IPC
#5102 opened Sep 2, 2025 by nsarka Loading…
17
[WIP] Dump option to print CuTE layouts for scheduled Fusion
#5085 opened Aug 28, 2025 by jacobhinkle • Draft
1
Use ATen ops on meta tensor to compute output shapes and strides for ExprEval segments
#5082 opened Aug 27, 2025 by zasdfgbnm • Draft
4
Accept some patterns with Cutlass scheduler if option enabled
#5077 opened Aug 26, 2025 by jacobhinkle • Draft
ProTip! Follow long discussions with comments:>50.