In some cases, gcc-11 does a good enough job translating normal looking
C expressions so use just those, but other times need to dig around for
an appropriate intrinsic.
Also, now need to disable psapi warnings when compiling for anything
less than avx.
It turns out gcc optimizes the obvious code nicely. It doesn't do so
well for cmul, but I decided to use obvious code anyway (the instruction
counts were the same, so maybe it doesn't get better for a single pair
of operands).
For int, long, float and double. I've been meaning to add them for a
while, and they're part of the new Ruamoko instructions set (which is
progressing nicely).
