GridSearchCV with NMF not use all cores

[DiegoCorrea]

Hi Nicholas and supporters,

Can you help me? This question is related with #195, were the last guy show a image with the Grid search using just one core.

So, I am using the Surprise in a machine with 5,19 TB de ram and 240 cores for my master degree and I get the same problem. This happens when I use the Google Cloud Plataform too. In some time, like 12 hours maybe all core are used. But I don't now when this start to work "normal".

gs = GridSearchCV(instance, params, measures=measures, cv=folds, n_jobs=-2, joblib_verbose=100)

I want to use 239 cores at the same time. In the code snnippet above, I use the NMF.

OBS: in my PC with 4 cores, that problem with the same code not happens. I think that is a problem with more than 8 cores. Python 3.7 With anaconda3/2020.11

[callowaysutton]

I am also facing this issue, any update?