The trace files are stored in a directory that is given by the arguments provided to the application. If these arguments are long this can result in simulation jobs failing without any given reason. We should limit the length of the argument string so we don't end up with excessively long directory names.
An example of a job that fails due to a long argument string is below. It can be fixed by shortening the directory name of the trace files and removing the arguments from the define-all-apps.yml, but ideally we wouldn't need this workaround.
Using logfiles ['/scratch/tgrogers-disk01/a/barnes88/private-accel-sim/util/job_launching/../job_launching/logfiles/sim_log.1ggnn.23.06.14-Wednesday.txt']
squeue.id Node App AppArgs Version Config RunningTime Mem JobStatus Basic GPGPU-Sim Stats
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
146046 UNKNOWN sift10k_multi __base_filename_data sift10k_multi.accels QV100-SASS UNKNOWN UNKNOWN NOT_RUNNING_NO_OUTPUT
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Passed:0/1, No error:0/1, Failed/Error:1/1, Running:0/1, Waiting:0/1
Contents :
All 1 Tests Done.
Something did not pass.
The trace files are stored in a directory that is given by the arguments provided to the application. If these arguments are long this can result in simulation jobs failing without any given reason. We should limit the length of the argument string so we don't end up with excessively long directory names.
An example of a job that fails due to a long argument string is below. It can be fixed by shortening the directory name of the trace files and removing the arguments from the define-all-apps.yml, but ideally we wouldn't need this workaround.