I wouldn’t automatically expect a linear decrease in compile time with growing core count. That would have to be tried.