I am using Mac with Intel processor. I can't figure out why RTorch package only uses 1 core for tensor operation while PyTorch does all 4 cores available on my laptop. There is some known issue about OpenMP on Mac. I tried to use my fixed makevars file and option remotes::install_github("mlverse/torch", build = TRUE ) for installations but this does not help. More over simultaneously usage of data.table and torch in that case leads to R session crash.
I am using Mac with Intel processor. I can't figure out why RTorch package only uses 1 core for tensor operation while PyTorch does all 4 cores available on my laptop. There is some known issue about OpenMP on Mac. I tried to use my fixed
makevarsfile and optionremotes::install_github("mlverse/torch", build = TRUE )for installations but this does not help. More over simultaneously usage of data.table and torch in that case leads to R session crash.