Skip to content
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.

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

FusionPersistentSoftmaxLocalShared failure #163

Closed
naoyam opened this issue Apr 12, 2023 · 0 comments · Fixed by #185
Closed

FusionPersistentSoftmaxLocalShared failure #163

naoyam opened this issue Apr 12, 2023 · 0 comments · Fixed by #185
Assignees

Comments

@naoyam
Copy link
Collaborator

naoyam commented Apr 12, 2023

On my Titan RTX, the FusionPersistentSoftmaxLocalShared_CUDA test fails:

C++ exception with description "(dynamic_smem_size) < (available_dynamic_smem_without_reconfiguration + additional_dynamic_smem_available_through_reconfiguration) INTERNAL ASSERT FAILED at "/home/nmaruyama/pytorch/debug3/nvfuser/csrc/executor.cpp":910, please report a bug to PyTorch. The total shared memory allocation is larger than available memory. Dynamic size: 66048. Available size: 49136. Configured smem size: 49152. Device limit size: 65536

It seems this issue started to happen at PR #148. Unclear why the PR could affect shared memory usage.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging a pull request may close this issue.

2 participants