-
Notifications
You must be signed in to change notification settings - Fork 16
/
Copy pathsbatch.sh
40 lines (29 loc) · 875 Bytes
/
sbatch.sh
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
#!/bin/sh
#SBATCH -J icn # Job name
#SBATCH -o sbatch_log/pytorch-1gpu.%j.out # Name of stdout output file (%j expands to %jobId)
#SBATCH -p A100 # queue name or partiton name titanxp/titanrtx/2080ti
#SBATCH -t 3-00:00:00 # Run time (hh:mm:ss) - 1.5 hours
#SBATCH --gres=gpu:1 # number of gpus you want to use
#SBATCH --nodes=1
##SBATCH --exclude=n13
##SBTACH --nodelist=n12
##SBTACH --ntasks=1
##SBATCH --tasks-per-node=1
##SBATCH --cpus-per-task=1
cd $SLURM_SUBMIT_DIR
echo "SLURM_SUBMIT_DIR=$SLURM_SUBMIT_DIR"
echo "CUDA_HOME=$CUDA_HOME"
echo "CUDA_VISIBLE_DEVICES=$CUDA_VISIBLE_DEVICES"
echo "CUDA_VERSION=$CUDA_VERSION"
srun -l /bin/hostname
srun -l /bin/pwd
srun -l /bin/date
module purge
echo "Start"
export NCCL_NSOCKS_PERTHREAD=4
export NCCL_SOCKET_NTHREADS=2
export WANDB_SPAWN_METHOD=fork
nvidia-smi
date
squeue --job $SLURM_JOBID
echo "##### END #####"