On my terrible laptop which only has a small amount of memory available before culling the tests, we noticed that this test (at least) balloons up to 3+GB before being culled by the out-of-memory killer.
This was brought up to at least be aware of due to the architecture of perlmutter having 64 or 128 cores and 512GB ram, therefore if the memory usage goes much higher than this we may be limited on the number of instances we can run at once.
tests that I have had it fail on (among others):
pipeline/test_pipeline.py::test_data_flow
improc/test_alignment.py::test_warp_decam
On my terrible laptop which only has a small amount of memory available before culling the tests, we noticed that this test (at least) balloons up to 3+GB before being culled by the out-of-memory killer.
This was brought up to at least be aware of due to the architecture of perlmutter having 64 or 128 cores and 512GB ram, therefore if the memory usage goes much higher than this we may be limited on the number of instances we can run at once.
tests that I have had it fail on (among others): pipeline/test_pipeline.py::test_data_flow improc/test_alignment.py::test_warp_decam