Assuming one fills the components section with a few different component analyses, is there a way to run the heavy gta.setup() part using multiple threads?
As far as I understood in a multi-component analysis one is basically creating a cube (and related output files) for each component and then when when calling anything else like fit or tsmap fermipy will use all of them and do a joint likelyhood - am I correct?
I was thinking to write a config file for each component (so basically copy-paste the "base" config file and overwrite the component section with the specific component I want) and then use multiprocessing to launch a gta.setup() run for each child config file.
Is this a correct thing to do? Are there better options, maybe even built-in?
Assuming one fills the
componentssection with a few different component analyses, is there a way to run the heavygta.setup()part using multiple threads?As far as I understood in a multi-component analysis one is basically creating a cube (and related output files) for each component and then when when calling anything else like
fitortsmapfermipy will use all of them and do a joint likelyhood - am I correct?I was thinking to write a config file for each component (so basically copy-paste the "base" config file and overwrite the
componentsection with the specific component I want) and then use multiprocessing to launch agta.setup()run for each child config file.Is this a correct thing to do? Are there better options, maybe even built-in?