Dear CUDA users, CDUA Toolkit was updated. -System: Login node and ACSG -Version: CDUA Toolkit 7.5 # Compaile and link Users can use CUDA in the same way as it was before this update. [username@greatwave ~]$ module load cuda [username@greatwave ~]$ nvcc test.cu Ex) Sample script for job execution [username@greatwave ~]$ vi acsg-gpu.sh #!/bin/sh #------ pjsub option --------# #PJM -L rscunit=gwacsg #PJM -L rscgrp=gpu #PJM -L vnode=1 #PJM -L vnode-core=1 #PJM -L elapse=10:00 #PJM -x gpu_per_vnode=2 #PJM -g Q99999 #PJM -j #------- Program execution -------# ./a.out [username@greatwave ~]$ pjsub acsg-gpu.sh [INFO] PJM 0000 pjsub Job 12345 submitted. # When users want to use older version please refer the following. [username@greatwave ~]$ module available cuda ----------- /gwfefs/opt/modulefiles/greatwave/compilers ------------ cuda/6.5 cuda/7.0 cuda/7.5(default) [username@greatwave ~]$ module load cuda/7.0 [username@greatwave ~]$ which nvcc /gwfefs/opt/x86_64/cuda/7.0/bin/nvcc Pleae contact hpc@riken.jp if you have questions.