sorry if this question has already been asked. I simply couldn't figure out right vocabulary to search for it. I have multiple experiments that I need to run in a cluster with GPU. But for now I start one job and wait for it to finish before launching the other. What can I do to "give command" to run all the scripts sequentially? I have something like the following:
./experiment.sh configs/job1.sh --train <gpu-ID>
./experiment.sh configs/job2.sh --train <gpu-ID>
./experiment.sh configs/job2.sh --train <gpu-ID>
where for <gpu-ID> I put 0 or 1 for specific GPUs
Right now I start a job say:
./experiment.sh configs/job1.sh --train <gpu-ID>
And periodically check whether it finished so that I can start job2.
What's a more efficient way to automate this?
&&likecmd1 && cmd2, this means, cmd2 will run after cmd1 fineshed sucessfullyjob1.shfinishes,job2.shwill start.