i965/fs: Don't CSE negated multiplies with saturation.

It's not correct to CSE these multiplies

   mul.sat dst1, -a, b
   mul.sat dst2,  a, b

by emitting a negated MOV from dst1 to dst2:

   mul.sat dst1, -a, b
   mov     dst2, -dst1

Take 2.0*2.0 for example. The first multiply would produce 0.0 and the
second would produce 1.0.

Fixes bad generated code in 18 to 22 shaders:

instructions in affected programs: 432 -> 464 (7.41%)
helped: 4
HURT: 18

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
This commit is contained in:
Matt Turner 2016-02-22 10:25:38 -08:00
parent 3da789f1e9
commit 1567da1e28

View file

@ -139,6 +139,8 @@ operands_match(const fs_inst *a, const fs_inst *b, bool *negate)
ys[1].f = ys1_imm;
*negate = (xs0_negate != xs1_negate) != (ys0_negate != ys1_negate);
if (*negate && (a->saturate || b->saturate))
return false;
return ret;
} else if (!a->is_commutative()) {
bool match = true;