Does newer version of Clara Parabricks cannt used in RTX3090?

Healthcare & Life Sciences Parabricks
1

The same param used in version 4.0.0 is succeeded，but in version 4.2.1 got a fail msg
“Illegal instruction (core dumped)” ,does newer version of Clara Parabricks not support RTX3090 ?

docker run
–gpus all
–rm
–volume $(pwd):/workdir
–volume $(pwd):/outputdir
nvcr.io/nvidia/clara/clara-parabricks:4.0.0-1
pbrun fq2bam
–ref /workdir/parabricks_sample/Ref/Homo_sapiens_assembly38.fasta
–in-fq /workdir/parabricks_sample/Data/sample_1.fq.gz /workdir/parabricks_sample/Data/sample_2.fq.gz
–out-bam /outputdir/fq2bam_output.bam

2

Hi,

Can you share the full log with us to take a look? Also Parabricks v4.3.0 has been released, it would be good to also try this latest version.

3

this is 4.0.0 version success
singularity exec --nv clara-parabricks_4.0.0/ pbrun fq2bam --ref ./parabricks_sample/Ref/Homo_sapiens_assembly38.fasta --in-fq ./parabricks_sample/D
ata/sample_1.fq.gz ./parabricks_sample/Data/sample_2.fq.gz --out-bam ./outputdir/fq2bam_output.bam
Please visit NVIDIA Clara - NVIDIA Docs for detailed documentation

[Parabricks Options Mesg]: Checking argument compatibility
[Parabricks Options Mesg]: Automatically generating ID prefix
[Parabricks Options Mesg]: Read group created for …

this is version 4.4.0 ,any greater than 4.0.0 all failed

image
image1709×723 78.1 KB

Driver Version: 550.120 CUDA Version: 12.4 OS Version ubantu 22.04

4

Can you post the specifications for your machine? It looks like the container started executing correctly but then was killed. It is possible that the run ran out of RAM and the OOM killed killed the process.