> $ qconf -sc | grep gpu > gpu gpu INT <= YES YES 0 0 > > Should be: > > # qconf -sc | grep gpu > gpu gpu INT <= YES JOB 0 0 > > I will poke at this on Monday and hopefully not mess things up too badly... > :-) > > Rajil, does this help you as well? I think this answers the same > question(s) that I had. > > -Dj
Yes, this seems to be working. I also set up a load_sensor script using the script mentioned here http://serverfault.com/questions/322073/howto-set-up-sge-for-cuda-devices something like this #qconf -sconf compute-4-0 compute-4-0.local: mailer /bin/mail xterm /usr/bin/X11/xterm execd_spool_dir /opt/gridengine/default/spool load_sensor /share/apps/load_sensor.sh The third submitted job is now in 'qw' as i wanted. #qstat 425 0.50500 j1 rajil r 12/05/2015 12:36:32 gpu.q@compute-4-0.local 8 426 0.50500 j2 rajil r 12/05/2015 12:37:17 gpu.q@compute-4-1.local 8 427 0.50500 j3 rajil qw 12/05/2015 12:39:18 8 Also i can see the gpu values to be 0 when they are being used #qhost -F gpu compute-4-0 linux-x64 32 1.73 126.1G 26.2G 1000.0M 0.0 Host Resource(s): hc:gpu=0.000000 compute-4-1 linux-x64 32 1.81 126.1G 24.4G 1000.0M 0.0 Host Resource(s): hc:gpu=0.000000 On the compute nodes i had to specify the gpu value #qconf -se compute-4-0 hostname compute-4-0.local load_scaling NONE complex_values gpu=1 But leave it the complex type empty in the queue #qconf -sq gpu.q complex_values NONE The queue submission script is like #!/bin/csh #$ -V #$ -S /bin/csh #$ -N j1 #$ -q gpu.q #$ -l gpu=1 #$ -m beas #$ -j y -o /home/rajil/tmp/tst/j1.qlog #$ -pe mpi 8 abaqus python /share/apps/abaqus/6.14-2/../abaJobHandler.py j1 /home/rajil/tmp/tst j1.fs.133848 0 j1.com model.inp -Rajil _______________________________________________ users mailing list users@gridengine.org https://gridengine.org/mailman/listinfo/users
