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

ttnn.linear fails with OOM issue #16268

Closed
keerthana-r-mcw opened this issue Dec 23, 2024 · 2 comments
Closed

ttnn.linear fails with OOM issue #16268

keerthana-r-mcw opened this issue Dec 23, 2024 · 2 comments

Comments

@keerthana-r-mcw
Copy link
Contributor

Describe the bug
ttnn.linear is failing with Out of Memory in the unit test of SDXL_Turbo model for the resolution 1024x1024 when memory config is set to ttnn.L1_MEMORY_CONFIG
To Reproduce
Steps to reproduce the behavior:

  1. Checkout keerthanar/sdxl_turbo_unit_test
  2. Run the following command,
    pytest tests/ttnn/unit_tests/operations/test_linear.py::test_sdxl_turbo_linear_1024x1024_l1_failing_case
  3. The test will fail with the following error will occur,
    RuntimeError: TT_THROW @ ../tt_metal/impl/allocator/allocator.cpp:145: tt::exception E info: E Out of Memory: Not enough space to allocate 41943040 B L1 buffer across 56 banks, where each bank needs to store 749568 B

Please complete the following environment information:
WH (n150)

@keerthana-r-mcw
Copy link
Contributor Author

Passes when dtype =. bfloat8

@keerthana-r-mcw
Copy link
Contributor Author

Closing this issue since this passes when dtype = bfloat8

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

No branches or pull requests

1 participant