On Sat, Aug 15, 2020, Camila Clemente wrote:
>
>I'm trying to run pmemd.cuda on a cluster that it owns:
>
>### gpu (Tesla K20X): 1-2
>### gpu (Tesla M2090): 1-3
>
>By running this:
>
>$AMBERHOME/bin/pmemd.cuda -O -i eq.mdin -p r_500.wet.complex.prmtop -c
>r_500_term.rst7 -r r_500_eq.rst7 -o r_500_eq.mdout -x r_500_eq.nc
>
>And the error I get is this:
>Error searching for compatible GPU.
Please report (a) what version of Amber you are using; (b) what the result of
typing "nvidia-smi" is; (3) post the entire mdout file, so that we can
see the entire error message and its context.
You could certainly try "export CUDA_VISIBLE_DEVICES=0" so see if that makes
a difference. You could also try running the test cases, but I'm guessing
that will give you the same error repeatedly.
....dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sun Aug 16 2020 - 05:30:02 PDT
