Skip to content

calcPi_intrinsic rework

Martin Kröning requested to merge mkroening/gi4:v6-intrinsics into master

This is a rework of calcPi_intrinsic in v6 of the Pi calculation examples.

This should now work in macOS, but I can't test that.

I replaced built-in functions with proper Intel intrinsics and reworked the __m128d variables.

On my machine calcPi_intrinsic is now faster than calcPi_SSE.

@stvogel, what do you think?

Merge request reports