Skip to content

Pull requests: NVIDIA/Fuser

New pull request New
154 Open 2,985 Closed
154 Open 2,985 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

Create a string from the char* argument in TraceScope constructor
#3925 opened Feb 20, 2025 by rdspring1 Loading…
2
Add benchmarks for cross entropy loss
#3924 opened Feb 19, 2025 by protonu • Draft
1
Expose MultiDeviceExecutor and overlapped AG+matmul to python API
#3923 opened Feb 19, 2025 by samnordmann Loading…
17
Tensor memory allocation analysis
#3922 opened Feb 19, 2025 by zasdfgbnm Loading…
2
Add a script to run a single matmul configuration with custom MatmulParams Matmuls
#3918 opened Feb 18, 2025 by rdspring1 Loading…
12
avoid ublk tma out bound access
#3917 opened Feb 18, 2025 by liqiangxl • Draft
4
Convert KernelExecutor to always use KernelArgumentHolder.
#3916 opened Feb 18, 2025 by csarofeen Loading…
2
Build some infrastructure to locally run all nvFuser tests and parallalize across available devices.
#3915 opened Feb 17, 2025 by csarofeen Loading…
24
[Multidevice] Add backend type to Communication and expose to API
#3914 opened Feb 17, 2025 by samnordmann Loading…
9
[CudaIpc Tuto] Minimal snippet example
#3912 opened Feb 17, 2025 by samnordmann Loading…
1
[CudaIpc 3/3]: p2p get-Zcopy
#3911 opened Feb 17, 2025 by samnordmann Loading…
1
[CudaIpc 2/3]: Ipc handle exchange
#3910 opened Feb 17, 2025 by samnordmann Loading…
6
[CudaIpc 1/3]P2PCommunication: add backend type
#3909 opened Feb 17, 2025 by samnordmann Loading…
6
[WIP] Traverse all paths (including cycles)
#3906 opened Feb 16, 2025 by naoyam • Draft
2
should not remove predicate for 1d TMA
#3899 opened Feb 14, 2025 by liqiangxl • Draft
3
Do some more simplifications specific to extents
#3891 opened Feb 14, 2025 by naoyam Loading…
11
Working towards returning KernelArgumentHolder/PolymorphicVal's from fusion execution.
#3890 opened Feb 14, 2025 by csarofeen Loading…
43
Move host benchmarks to a subdirectory for easier distinction
#3887 opened Feb 13, 2025 by Priya2698 Loading…
8
Create Communication scheduler, move meta-ops to ExprEval scheduler
#3885 opened Feb 13, 2025 by nsarka Loading…
@nsarka
9
P2p comms through Cuda Ipc
#3883 opened Feb 13, 2025 by samnordmann • Draft
Fix slice to empty not being converted to FullOp
#3882 opened Feb 13, 2025 by jacobhinkle Loading…
7
[WIP] Set allocation as loop in fusion segmentor
#3880 opened Feb 13, 2025 by Priya2698 • Draft
2
IdModel: Fix invalid promotion selection idmodel
#3877 opened Feb 12, 2025 by naoyam Loading…
1
10
DID loop split on reshaped IDs
#3875 opened Feb 11, 2025 by Priya2698 Loading…
14
Use nvtx3 to avoid deprecation warning
#3872 opened Feb 11, 2025 by Priya2698 Loading…
8
ProTip! Type g i on any issue or pull request to go back to the issue listing page.