[clang] [flang] [lld] [llvm] [mlir] [RFC] Use pre-compiled headers to speed up LLVM build (~1.5-2x) (PR #173868)

[Andrew Pinski via cfe-commits]

pinskia wrote:

Llvm CPU time: 5497.8 usr, 297.6 sys
Vs
Gcc 5232.6 usr, 559.8 sys

Hmm. So from the looks of it clang/llvm could be optimized more. Gcc is most likely spending more time in system as it writes out the .s file and then gas reads it back in and assembles it.
So gcc is spending less time in user space to do work, ~5% if my math is correct. But 2x as much time in the kernel. Which does corresponding to writing out a .s file.


https://github.com/llvm/llvm-project/pull/173868