I was struggling with inconsistent behavior when running the SEDM-KP pipeline: in short, I would get different values for SPREAD_MODEL inside my sextractor catalogs, depending on if I was running the pipeline on one image at a time or on multiple images. When re-running the pipeline several times with multiple images, those SPREAD_MODEL values would continue to vary.
It was a strange behavior, but the fix was to set max_n_cpu=1. I did this first globally, (export MAX_N_CPU...) then diagnosed that Sextractor is the processor to limit to avoid this weird behavior (not PSFex like I suspected in our call today, @robertdstein). Hopefully this is a small change for everyone!
I was struggling with inconsistent behavior when running the SEDM-KP pipeline: in short, I would get different values for
SPREAD_MODELinside my sextractor catalogs, depending on if I was running the pipeline on one image at a time or on multiple images. When re-running the pipeline several times with multiple images, thoseSPREAD_MODELvalues would continue to vary.It was a strange behavior, but the fix was to set max_n_cpu=1. I did this first globally, (export MAX_N_CPU...) then diagnosed that Sextractor is the processor to limit to avoid this weird behavior (not PSFex like I suspected in our call today, @robertdstein). Hopefully this is a small change for everyone!