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.

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

Jump to bottom

gcc should be listed in the README as required dependency for running GATK #6012

Open
TedBrookings opened this issue Jun 20, 2019 · 2 comments · May be fixed by #8196
Open

gcc should be listed in the README as required dependency for running GATK #6012

TedBrookings opened this issue Jun 20, 2019 · 2 comments · May be fixed by #8196

Comments

@TedBrookings
Copy link
Contributor

Documentation request

Description

I propose that installation of gcc be added to the instructions on the GATK Github README.md

If gcc is not installed, HaplotypeCaller complains that the AVX instruction set is not available, even when it is. It falls back to slower LOGLESS_CACHING PairHMM. The fault is missing libgomp1, which is a required dependency of gcc.

Since this documentation request is related to a "bug" that comes about from not installing necessary libraries, I'll include the bug report format below, in case someone else searches for solutions to this problem, as suggested by @lbergelson

Affected tool(s) or class(es)

HaplotypeCaller, or any other tool that uses PairHMM

Affected version(s)

-I think all as of 2019-06-20. I tested on release version 4.1.2.0

Steps to reproduce

Run HaplotypeCaller from a released jar on an Ubuntu VM that supports the AVX instruction set. Critically, do NOT install gcc on the VM. Installing gcc fixes this problem.

Expected behavior

If you install gcc, that results in the installation of libgomp1, which allows the Intel library to load and use AVX acceleration. You could probably install libgomp1 on its own, but I did not test that.

14:51:01.013 INFO NativeLibraryLoader - Loading libgkl_utils.so from jar:file:/home/ubuntu/gatk-4.1.2.0/gatk-package-4.1.2.0-local.jar!/com/intel/gkl/native/libgkl_utils.so
14:51:01.015 INFO NativeLibraryLoader - Loading libgkl_pairhmm_omp.so from jar:file:/home/ubuntu/gatk-4.1.2.0/gatk-package-4.1.2.0-local.jar!/com/intel/gkl/native/libgkl_pairhmm_omp.so
14:51:01.053 INFO IntelPairHmm - Using CPU-supported AVX-512 instructions
14:51:01.053 INFO IntelPairHmm - Flush-to-zero (FTZ) is enabled when running PairHMM
14:51:01.054 INFO IntelPairHmm - Available threads: 16
14:51:01.054 INFO IntelPairHmm - Requested threads: 8
14:51:01.054 INFO PairHMM - Using the OpenMP multi-threaded AVX-accelerated native PairHMM implementation

Actual behavior

Without libgomp1, AVX acceleration doesn't work:

19:43:36.387 INFO NativeLibraryLoader - Loading libgkl_utils.so from jar:file:/home/ubuntu/gatk-4.1.2.0/gatk-package-4.1.2.0-local.jar!/com/intel/gkl/native/libgkl_utils.so
19:43:36.389 WARN NativeLibraryLoader - Unable to load libgkl_utils.so from native/libgkl_utils.so (/tmp/libgkl_utils5391341743604217103.so: libgomp.so.1: cannot open shared object file: No such file or directory)
19:43:36.389 WARN IntelPairHmm - Intel GKL Utils not loaded
19:43:36.389 INFO PairHMM - OpenMP multi-threaded AVX-accelerated native PairHMM implementation is not supported
19:43:36.389 INFO NativeLibraryLoader - Loading libgkl_utils.so from jar:file:/home/ubuntu/gatk-4.1.2.0/gatk-package-4.1.2.0-local.jar!/com/intel/gkl/native/libgkl_utils.so
19:43:36.390 WARN NativeLibraryLoader - Unable to load libgkl_utils.so from native/libgkl_utils.so (/tmp/libgkl_utils3484179251394006588.so: libgomp.so.1: cannot open shared object file: No such file or directory)
19:43:36.390 WARN IntelPairHmm - Intel GKL Utils not loaded
19:43:36.390 WARN PairHMM - ***WARNING: Machine does not have the AVX instruction set support needed for the accelerated AVX PairHmm. Falling back to the MUCH slower LOGLESS_CACHING implementation!
@scalavision
Copy link

scalavision commented Oct 13, 2019 •
edited
Loading

If you wrap the gatk python script with something like:

#! /bin/bash -e
export LD_LIBRARY_PATH='<path to gcc-7.4.0-lib>/lib'${LD_LIBRARY_PATH:+':'}$LD_LIBRARY_PATH
exec "gatk" "$@"

It should work

@yonniejon
Copy link

yonniejon commented Dec 25, 2022 •
edited
Loading

I have gcc installed and have also tried the suggestion that @scalavision has suggested and still I get the warning "Machine does not have the AVX instruction set support needed for the accelerated AVX PairHmm. Falling back to the MUCH slower LOGLESS_CACHING implementation!".

I am not sure that I used the correct path. I found the path to gcc 7.4.0 path by running
gcc -print-prog-name=cc1 -v

@lbergelson lbergelson mentioned this issue Feb 9, 2023
Closed
droazen added a commit that referenced this issue Feb 10, 2023
@droazen
Add GCC OpenMP library to list of GATK requirements in the README
7d6c4b4 
The GCC OpenMP library, libgomp1, a required dependency of GCC, needs to be present in order to run the GKL accelerated PairHMM in tools like HaplotypeCaller. We now mention this requirement in the GATK README.

Resolves #6012
@droazen droazen linked a pull request Feb 10, 2023 that will close this issue
Open
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

Add GCC OpenMP library to list of GATK requirements in the README broadinstitute/gatk
3 participants
@yonniejon @scalavision @TedBrookings