How to build pytorch using multicore cpu

Hi!

I want to reduce the build time of pytorch.
How can i do?
I want to know about build options such that make -j4

I think by default it’s using as many cores as it can when building.