epi2me-labs / wf-artic

ARTIC SARS-CoV-2 workflow and reporting
https://labs.epi2me.io/
Other
49 stars 36 forks source link

UCX package #47

Closed jftoupet closed 2 years ago

jftoupet commented 2 years ago

We have this error, In addition, the UCX support is also built but disabled by default. To enable it, first install UCX (conda install -c conda-forge ucx). Then, set the environment variables OMPI_MCA_pml="ucx" OMPI_MCA_osc="ucx" before launching your MPI processes. Equivalently, you can set the MCA parameters in the command line: mpiexec --mca pml ucx --mca osc ucx ... Note that you might also need to set UCX_MEMTYPE_CACHE=n for CUDA awareness via UCX. Please consult UCX's documentation for detail.

however I notice that ucx is not installed in the 1.1.14 image. It is impossible to install it because the node.js 16 package is not present in the epi2melabs repo.

It's a shame not to be able to use cuda on a 50M$ gridion

Thank's Best regards

cjw85 commented 2 years ago

Hi @jftoupet we don't currently support using the NVIDIA CUDA runtime in our workflow containers. This is on the TODO list.

For wf-artic very little of the workflow is GPU capable so there is little to gain.

jftoupet commented 2 years ago

Ok why not, but why ha e a 16M$ GPU just to sequencing with minknow ?

Le mar. 19 avr. 2022, 18:39, cjw85 @.***> a écrit :

Hi @jftoupet https://github.com/jftoupet we don't currently support using the NVIDIA CUDA runtime in our workflow containers. This is on the TODO list.

For wf-artic very little of the workflow is GPU capable so there is little to gain.

— Reply to this email directly, view it on GitHub https://github.com/epi2me-labs/wf-artic/issues/47#issuecomment-1102862619, or unsubscribe https://github.com/notifications/unsubscribe-auth/AHEKY6JV4U5N3N7SOYMEJATVF3OUHANCNFSM5TY557HQ . You are receiving this because you were mentioned.Message ID: @.***>

mattdmem commented 2 years ago

Hello again @jftoupet - The primary use of the GPU on the GridION is basecalling.

For wf-artic we use existing community developed tools like FieldBioinformatics for analysis, which itself uses tools that are not GPU enabled.

I will close this issue now, please do raise new issues in the future if you have any problems with our workflows.

jftoupet commented 2 years ago

Hi Thank's Best regards Jf

Le mer. 20 avr. 2022 à 14:21, Matthew Parker @.***> a écrit :

Hello again @jftoupet https://github.com/jftoupet - The primary use of the GPU on the GridION is basecalling.

For wf-artic we use existing community developed tools like FieldBioinformatics for analysis, which itself uses tools that are not GPU enabled.

I will close this issue now, please do raise new issues in the future if you have any problems with our workflows.

— Reply to this email directly, view it on GitHub https://github.com/epi2me-labs/wf-artic/issues/47#issuecomment-1103868378, or unsubscribe https://github.com/notifications/unsubscribe-auth/AHEKY6OTHSCJDG7IJ7GSATTVF7ZFHANCNFSM5TY557HQ . You are receiving this because you were mentioned.Message ID: @.***>

-- Jean-François TOUPET