cpuNumaStarter does not reliably work on Hemera defq

ComputationalRadiationPhysics / picongpu

Performance-Portable Particle-in-Cell Simulations for the Exascale Era :sparkles:

https://picongpu.readthedocs.io

Other

694 stars 218 forks source link

cpuNumaStarter does not reliably work on Hemera defq #3331

Closed sbastrakov closed 1 year ago

sbastrakov commented 4 years ago

Not every time, but rather consistently running LWFA with 1.cfg on Hemera defq results in error

numa_sched_setaffinity_v2_int() failed; abort
: Invalid argument
sched_setaffinity: Invalid argument

I do not know (yet) whether it's our problem or something is broken on Hemera, and if this is specific for 1 process. Workaround: remove numactl call and replace this line with just $*

PrometheusPi commented 4 years ago

@sbastrakov great catch - could you already determine the cause of the error?

sbastrakov commented 4 years ago

No, as this issue stopped reproducing for me after that day (but it was for sure present more than one time before), and then I was busy with some other bugs.

sbastrakov commented 1 year ago

I am not sure what is the status of that, but I guess it is not as originally posted in this issue anyways. So i will close it.