google-deepmind / google-deepmind/alphafold3

Segmentation fault for RTX PRO6000

Open
#735 0 comments 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
8.6k
Forks
1.4k
PR merge metrics
No merged PRs in 30d

Description

Hi,
I try to configure alphafold3 in RTX PRO6000, but got errors:
Fatal Python error: Segmentation fault

Current thread 0x00007a8c45781740 (most recent call first):
File "/home/guodong/anaconda3/envs/alphafold3/lib/python3.12/site-packages/jax/_src/compiler.py", line 362 in backend_compile_and_load
File "/home/guodong/anaconda3/envs/alphafold3/lib/python3.12/site-packages/jax/_src/profiler.py", line 359 in wrapper
File "/home/guodong/anaconda3/envs/alphafold3/lib/python3.12/site-packages/jax/_src/compiler.py", line 746 in _compile_and_write_cache
File "/home/guodong/anaconda3/envs/alphafold3/lib/python3.12/site-packages/jax/_src/compiler.py", line 478 in compile_or_get_cached
File "/home/guodong/anaconda3/envs/alphafold3/lib/python3.12/site-packages/jax/_src/interpreters/pxla.py", line 2854 in _cached_compilation
File "/home/guodong/anaconda3/envs/alphafold3/lib/python3.12/site-packages/jax/_src/interpreters/pxla.py", line 3073 in from_hlo
File "/home/guodong/anaconda3/envs/alphafold3/lib/python3.12/site-packages/jax/_src/interpreters/pxla.py", line 2527 in compile
File "/home/guodong/anaconda3/envs/alphafold3/lib/python3.12/site-packages/jax/_src/pjit.py", line 1620 in _pjit_call_impl_python
File "/home/guodong/anaconda3/envs/alphafold3/lib/python3.12/site-packages/jax/_src/pjit.py", line 146 in _python_pjit_helper
File "/home/guodong/anaconda3/envs/alphafold3/lib/python3.12/site-packages/jax/_src/pjit.py", line 264 in cache_miss
File "/home/guodong/anaconda3/envs/alphafold3/lib/python3.12/site-packages/jax/_src/traceback_util.py", line 195 in reraise_with_filtered_traceback
File "/home/guodong/opt/alphafold3/run_alphafold.py", line 502 in run_inference
File "/home/guodong/opt/alphafold3/run_alphafold.py", line 610 in predict_structure
File "/home/guodong/opt/alphafold3/run_alphafold.py", line 889 in process_fold_input
File "/home/guodong/opt/alphafold3/run_alphafold.py", line 1071 in main
File "/home/guodong/anaconda3/envs/alphafold3/lib/python3.12/site-packages/absl/app.py", line 355 in _run_main
File "/home/guodong/anaconda3/envs/alphafold3/lib/python3.12/site-packages/absl/app.py", line 410 in run
File "/home/guodong/opt/alphafold3/run_alphafold.py", line 1092 in
Would you please help me to solve the problem? Thank you. Here I attached the installed packages and full log.txt.

[installed.txt](https://github.com/user-attachments/files/31831595/installed.txt)

[log.txt](https://github.com/user-attachments/files/31831596/log.txt)

Contributor guide

Open the contributing guide

Research direction

Start by reviewing the attached installed.txt and log.txt, then trace the inference entry points in run_alphafold.py at lines 502, 610, 889, and 1071. Compare the reported Python and package environment with the RTX PRO6000 execution path. Done means identifying a reproducible compatibility cause or documented configuration that allows inference to run without the segmentation fault.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
machine-learning
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Active
Clarity
Needs clarification
Newbie friendliness
38/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.