Memory per job

[DrSpringy]

Why does the "Memory per Job" on the "Devices" tab of deconvolution page, vary each time we use it?

[DrSpringy]

need to confirm max memory that runs on a node, 380GB RAM jobs fail, 360GB RAM jobs run fine

[DrSpringy]

need to test this against FIJI/microvolution plugin to ensure it's working correctly

[DrSpringy]

is there a faster way to run this calculation, ie on a dedicated hardware node?