Copy link
@guyharris

guyharris Sep 15, 2017

Author Member

whether MSVC uses FPU for float calculations?

As in "uses the FPU or uses SSE" - the FPU instructions do 80-bit floating-point internally, with 80-bit registers, but the SSE instructions do purely 64-bit floating-point (as do, I think, floating-point instructions on most other processors).