[LispM-Hackers] BR-*-LIKELY

[James A. Crippen]

Paul Fuqua <pf@ti.com> writes:

> If we were doing this, we wouldn't want to fall out just because of a
> branch, so we had the LIKELY branches to allow the common case to stay
> in RISC mode.  However, we could only sustain it for about four
> macroinstructions anyway due to memory-bandwidth limitations, so LIKELY
> branches weren't a big benefit and I don't think the compiler ever
> generated them.

Interesting...  RISC mode indeed.  I'll look in the compiler to see if
it has LAP for LIKELY instructions.  If not then we can gleefully
ignore them.

There are a few other insns that I've found in the miscops and auxops
that are probably not implemented.  I've listed them in the
appropriate enums anyway, because we might want to actually implement
and use some of them someday.  In any case, they're there for
completeness.

'james

-- 
James A. Crippen <james@unlambda.com> ,-./-.  Anchorage, Alaska,
Lambda Unlimited: Recursion 'R' Us   |  |/  | USA, 61.20939N, -149.767W
Y = \f.(\x.f(xx)) (\x.f(xx))         |  |\  | Earth, Sol System,
Y(F) = F(Y(F))                        \_,-_/  Milky Way.