Use multiple cores for model training

[fabianegli]

Dear MSqRob devs,

When running some initial tests with our data in MSqRob I realized that the tool only uses one CPU core and trains one model after the other. Couldn't there be some parallelization in training the different models?

I have to admit, I didn't do the reading to exclude the possibility that they need to be run sequentially. Do they?

Kind regards,
Fabian

PS: I am writing this issue because I am waiting for 619 models to be trained. Fitting models... now at Fitting model 108 of 619. after ~20' runtime.

[fabianegli]

In case this is important: I am working on a MacOS 10.14.6 with R4.0.2 and MSqRob 0.7.7 installed today with devtools from GitHub and while installing updating all dependencies as well.

[ludgergoeminne]

Dear fabianegli

It is certainly possible to use multithreading with MSqRob.
In fact, we already did this in our hurdle paper, see e.g. for an example.
However, parallelization has not yet been implemented in the GUI.

Best regards

Ludger