We want to run the quicksurvey on 1000s of mocks for dark time. @forero mentioned that it was not possible to submit quicksurvey to compute node due to memory issue. Is their a workaround to run many instances of quicksurvey on NERSC to process 1000 mocks efficiently?
It will be great if someone can help us figure out how for example run following instance of quicksuvey on some compute node rather than head node?
Hi,
We want to run the quicksurvey on 1000s of mocks for dark time. @forero mentioned that it was not possible to submit quicksurvey to compute node due to memory issue. Is their a workaround to run many instances of quicksurvey on NERSC to process 1000 mocks efficiently?
It will be great if someone can help us figure out how for example run following instance of quicksuvey on some compute node rather than head node?
quicksurvey -T /global/project/projectdirs/desi/users/shadaba/Buzzard/FiberAssign/quicksurvey_example/targets/no_spectra/dark_data/ -E /global/project/projectdirs/desi/users/shadaba/Buzzard/FiberAssign/quicksurvey_example/survey/all_exposures_dark.fits --output_dir /global/project/projectdirs/desi/users/shadaba/Buzzard/FiberAssign/quicksurvey_example/zcat/dark_data/ -f $(which fiberassign) -t /global/project/projectdirs/desi/users/shadaba/Buzzard/FiberAssign/quicksurvey_example/fiberassign/template_fiberassign_dark.txt -D /global/project/projectdirs/desi/users/shadaba/Buzzard/FiberAssign/quicksurvey_example/fiberassign/buzzard_dark_fiberassign_dates.txt