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

ptxas returned an error during compilation of ptx to sass #122

Closed
2subzero opened this issue Aug 12, 2021 · 9 comments
Closed

ptxas returned an error during compilation of ptx to sass #122

2subzero opened this issue Aug 12, 2021 · 9 comments
Labels
error report Something isn't working

Comments

@2subzero
Copy link

When running alphafold on Centos7, encountering ptxas error

I0811 21:45:02.256136 47363873189568 templates.py:270] Found an exact template match 3fwt_A. I0811 21:45:04.627611 47363873189568 run_alphafold.py:142] Running model model_1 I0811 21:46:29.244786 47363873189568 model.py:131] Running predict with shape(feat) = {'aatype': (4, 560), 'residue_index': (4, 560), 'seq_length': (4,), 'template_aatype': (4, 4, 560), 'template_all_atom_masks': (4, 4, 560, 37), 'template_all_atom_positions': (4, 4, 560, 37, 3), 'template_sum_probs': (4, 4, 1), 'is_distillation': (4,), 'seq_mask': (4, 560), 'msa_mask': (4, 508, 560), 'msa_row_mask': (4, 508), 'random_crop_to_size_seed': (4, 2), 'template_mask': (4, 4), 'template_pseudo_beta': (4, 4, 560, 3), 'template_pseudo_beta_mask': (4, 4, 560), 'atom14_atom_exists': (4, 560, 14), 'residx_atom14_to_atom37': (4, 560, 14), 'residx_atom37_to_atom14': (4, 560, 37), 'atom37_atom_exists': (4, 560, 37), 'extra_msa': (4, 5120, 560), 'extra_msa_mask': (4, 5120, 560), 'extra_msa_row_mask': (4, 5120), 'bert_mask': (4, 508, 560), 'true_msa': (4, 508, 560), 'extra_has_deletion': (4, 5120, 560), 'extra_deletion_value': (4, 5120, 560), 'msa_feat': (4, 508, 560, 49), 'target_feat': (4, 560, 22)} 2021-08-11 21:46:29.891255: W external/org_tensorflow/tensorflow/stream_executor/gpu/asm_compiler.cc:81] Couldn't get ptxas version string: Internal: Running ptxas --version returned 32512 2021-08-11 21:46:29.922824: F external/org_tensorflow/tensorflow/compiler/xla/service/gpu/nvptx_compiler.cc:479] ptxas returned an error during compilation of ptx to sass: 'Internal: ptxas exited with non-zero error code 32512, output: ' If the error message indicates that a file could not be written, please verify that sufficient filesystem space is provided. Fatal Python error: Aborted

@Augustin-Zidek
Copy link
Collaborator

Have you checked whether you have enough disk space as the error mentions?

If the error message indicates that a file could not be written, please verify that sufficient filesystem space is provided.

@Ftamsir
Copy link

Ftamsir commented Aug 15, 2021

Finally find same problem with mine. I'm running on Ubuntu2004 with 3080Ti cuda 11.4 cudnn8.2.1 and i see the ptxas error too. I tried other cuda and cudnn version like 11.1 + 8.1.1 but it didn't work. Otherwise cudnn samples and tensorflow.test shows my gpu is working. I guess it's the jax or jaxlib problem. Hope you make progress

@2subzero
Copy link
Author

Have you checked whether you have enough disk space as the error mentions?

If the error message indicates that a file could not be written, please verify that sufficient filesystem space is provided.

I checked, and there is enough space.

@benfulton
Copy link

I got this error when ptxas wasn't in my path. I added it to the path and the error went away.

@Ftamsir
Copy link

Ftamsir commented Sep 1, 2021

I got this error when ptxas wasn't in my path. I added it to the path and the error went away.

i used run_docker.py and i set
PATH=/usr/local/cuda-11.4/bin:$PATH
but it still got
I0901 12:14:38.792916 140640899000128 run_docker.py:193] 2021-09-01 04:14:38.791988: W external/org_tensorflow/tensorflow/stream_executor/gpu/asm_compiler.cc:81] Couldn't get ptxas version string: Internal: Couldn't invoke ptxas --version I0901 12:14:38.794684 140640899000128 run_docker.py:193] 2021-09-01 04:14:38.793811: F external/org_tensorflow/tensorflow/compiler/xla/service/gpu/nvptx_compiler.cc:479] ptxas returned an error during compilation of ptx to sass: 'Internal: Failed to launch ptxas' If the error message indicates that a file could not be written, please verify that sufficient filesystem space is provided. I0901 12:14:38.795017 140640899000128 run_docker.py:193] Fatal Python error: Aborted
can you please show me how you added it to the path? thank you anyway

@abridgland abridgland added the error report Something isn't working label Sep 2, 2021
@Lsz-20
Copy link

Lsz-20 commented Dec 8, 2021

the same question,have you solved this ?@2subzero
waiting for your help! @Ftamsir @benfulton

@Augustin-Zidek
Copy link
Collaborator

This should be fixed by bumping the CUDA version in AlphaFold v2.1.2. Feel free to reopen if that is not the case.

@hn42
Copy link

hn42 commented Jun 25, 2023

thenks @benfulton

add ptxas filepath
or
add ptxas file in project (I solved like this)

@asmit203
Copy link

asmit203 commented Mar 8, 2024

thenks @benfulton

add ptxas filepath or add ptxas file in project (I solved like this)

Can you please tell me how did you add then ptxas to the project.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
error report Something isn't working
Projects
None yet
Development

No branches or pull requests

8 participants