Skip to content

Issues: NVIDIA/Fuser

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

166 Open 534 Closed
166 Open 534 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
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

Race reported between Write access and Read access in fusion using async copy bug Something isn't working
#3428 opened Nov 17, 2024 by liqiangxl
1
2
thread predicate is missing in async copy
#3427 opened Nov 17, 2024 by liqiangxl
Dynamic shape host latency is slow dynamic shapes Host Latency
#3419 opened Nov 15, 2024 by jacobhinkle
1
@jacobhinkle
2
Failed to create/execute the same FusionDefinition multiple times. Python API Issues related to the Python API
#3389 opened Nov 8, 2024 by wujingyue
@rdspring1
Tracks performance issues related to inner reduction scheduler perf Schedulers & Heuristics Issues related to Schedulers & Heuristics.
#3293 opened Oct 28, 2024 by liqiangxl
@liqiangxl
Memory inefficient order of execution with multiple parallel paths Thunder
#3290 opened Oct 28, 2024 by IvanYashchuk
@jacobhinkle
4
Bind sharded input/output tensors with DID-parallelized allocation domains. Multi-GPU
#3282 opened Oct 25, 2024 by wujingyue
2
@wujingyue
14
Tracking perf optimization of HopperMatmulTest.HSH_NT_128BSwizzle for problem size (M=2048, N=2048, K=8192), CTA tile size (128, 256) H100 Perf improve performance on H100 Matmuls
#3279 opened Oct 25, 2024 by zasdfgbnm
@zasdfgbnm
4
Tracks performance issues related to inner outer persistent scheduler perf Schedulers & Heuristics Issues related to Schedulers & Heuristics.
#3272 opened Oct 25, 2024 by liqiangxl
@liqiangxl
1
Add dtype parameter to Tensor state in python frontend. good first issue Good for newcomers Python API Issues related to the Python API
#3269 opened Oct 24, 2024 by rdspring1
@rdspring1
4
Predicating resize Indexing Ops
#3257 opened Oct 23, 2024 by naoyam
@naoyam
Examine the broadcast size for aliased input/output when identifying RW race Segmentation Issues related to nvFuser Segmentation Thunder
#3251 opened Oct 22, 2024 by Priya2698
@Priya2698
2
NVFuserTest and mpi_test aren't compatible. Multi-GPU Testing e.g. improving test infra and test coverage
#3214 opened Oct 18, 2024 by wujingyue
@wujingyue
4
std::get error from getBankConflictInfo Instrumentation
#3188 opened Oct 15, 2024 by liqiangxl
@protonu
2
Add back the UCC backend for Bcast_sharded/PipelineTestTwoStages tests. Multi-GPU
#3124 opened Oct 7, 2024 by wujingyue
@samnordmann
13
CommunicationTest.SendRecv/UCC hangs. Multi-GPU
#3120 opened Oct 7, 2024 by wujingyue
@samnordmann
5
Check Tensor.dims equals the rank of the corresponding TensorView. Python API Issues related to the Python API
#3117 opened Oct 5, 2024 by wujingyue
1
@rdspring1
2
Fused pointwise + layernorm kernel slower than unfused kernels perf Schedulers & Heuristics Issues related to Schedulers & Heuristics.
#3095 opened Oct 3, 2024 by cowanmeg
@liqiangxl
[RFC] Multi-Gpu Python Frontend API Multi-GPU Python API Issues related to the Python API
#3094 opened Oct 3, 2024 by rdspring1
@rdspring1
14
A Python distributed test don't exit properly when one rank fails. Multi-GPU Testing e.g. improving test infra and test coverage
#3092 opened Oct 3, 2024 by wujingyue
@wujingyue
1
Gdb a python distributed test. Multi-GPU
#3091 opened Oct 3, 2024 by wujingyue
@wujingyue
RoPE-like slice-concat patterns may result in cyclic exact graphs Indexing Ops rope
#3072 opened Oct 1, 2024 by naoyam
@naoyam
2
Unable to reshard non-contiguous inputs. Multi-GPU
#3071 opened Sep 30, 2024 by wujingyue
@cowanmeg
1
Inlining RoPE-like patterns require an analysis akin to ComputeAtLogicalDomainMap Indexing Ops rope
#3068 opened Sep 30, 2024 by naoyam
@naoyam
2
FusionExecutorCache not synchronizing with multiple devices Multi-GPU
#3042 opened Sep 27, 2024 by cowanmeg
@wujingyue
1
ProTip! Type g i on any issue or pull request to go back to the issue listing page.