mahmood naderan

Hi Werner, I have reinstalled one of the nodes and I want to know how can I use the slurm roll to install daemons on the new compute nodes. Is that possible? Or I have to reinstall on the frontend again?

Hi and thanks for the tool. Regarding the integer performance, I see this kernel: __global__ void CZCudaCalcKernelInteger32( void *buf /*!<[in] Data buffer. */ ) { int index = blockIdx.x * blockDim.x + threadIdx.x; int *arr = (int*)buf; int val1 = index; int val2 = arr[index]; int i; for(i = 0; i < CZ_CALC_BLOCK_LOOPS; i++) { CZ_CALC_IMAD32_256(val1, val2); CZ_CALC_IMAD32_256(val1, val2); CZ_CALC_IMAD32_256(val1, val2); CZ_CALC_IMAD32_256(val1, val2); CZ_CALC_IMAD32_256(val1, val2); CZ_CALC_IMAD32_256(val1,...

Hi Werner, I did the following steps: 1- eth1 with 10.1.1.1 on frontend. 2- installed slurm on frontend and nodes. 3- everything was fine. 4- unpluged eth1 and plugged the cable to eth2 with 10.1.1.1 and disabled eth1. Now, rocks sync slurm command shows some messages like this for each compute node. compute-0-1: File "/opt/rocks/bin/rocks", line 260, in <module> compute-0-1: command.runWrapper(name, args[i:]) compute-0-1: File "/opt/rocks/lib/python2.7/site-packages/rocks/commands/__init__.py",...

Can I downgrade 18 to 17? I think I have to use the last working one for myself.

OK I did that. But it doesn't work. Please note that I have connected to the Gnome through a vnc session [mahmood@rocks7 ~]$ echo $DISPLAY :1 [mahmood@rocks7 ~]$ srun --spankx11 --nodelist=rocks7 -n 1 -c 6 --mem=8G -A y8 -p RUBY xclock Error: Can't open display: 172.20.54.10:13.0 srun: error: rocks7: task 0: Exited with exit code 1 [mahmood@rocks7 ~]$ srun --spankx11 --nodelist=compute-0-3 -n 1 -c 6 --mem=8G -A y8 -p RUBY xclock Error: Can't open display: localhost:10.0 srun: error: compute-0-3:...

OK I did that. But it doesn't work. Please note that I have connected to the Gnome through a vnc session [mahmood@rocks7 ~]$ echo $DISPLAY :1 [mahmood@rocks7 ~]$ srun --spankx11 --nodelist=rocks7 -n 1 -c 6 --mem=8G -A y8 -p RUBY xclock Error: Can't open display: 172.20.54.10:13.0 srun: error: rocks7: task 0: Exited with exit code 1 [mahmood@rocks7 ~]$ srun --spankx11 --nodelist=compute-0-3 -n 1 -c 6 --mem=8G -A y8 -p RUBY xclock Error: Can't open display: localhost:10.0 srun: error: compute-0-3:...

OK. But there is no sign of spank [root@rocks7 ~]# rpm -q slurm slurm-18.08.3-1.el7.centos.x86_64 [root@rocks7 ~]# rpm -q slurm-spank-x11 package slurm-spank-x11 is not installed [root@rocks7 ~]# rpm -qa | grep slurm slurm-pam_slurm-18.08.3-1.el7.centos.x86_64 slurm-slurmd-18.08.3-1.el7.centos.x86_64 slurm-18.08.3-1.el7.centos.x86_64 slurm-devel-18.08.3-1.el7.centos.x86_64 slurm-torque-18.08.3-1.el7.centos.x86_64 slurm-rolldoc-7.0.0-17.11.5.x86_64 slurm-slurmdbd-18.08.3-1.el7.centos.x86_64 slurm-perlapi-18.08.3-1.el7.centos.x86_64...

Excuse me, but I get this [root@rocks7 ~]# rpm -q slurm* package slurm-7.0.0-17.11.05.00.01.x86_64.disk1.iso is not installed package slurm-7.0.0.185-18.08.03.01.00.x86_64.disk1.iso is not installed package slurm-7.0.0.186-18.08.03.01.01.x86_64.disk1.iso is not installed package slurm-spank-x11-0.2.5 is not installed package slurm-spank-x11-0.2.5.tar.gz is not installed [root@rocks7 ~]# rpm -qa | grep slurm* grep: slurm-spank-x11-0.2.5: Is a directory I wonder how srun works then! [root@rocks7 ~]#...