nir/algebraic: Optimize some ifind_msb to ufind_msb

On Intel platforms, the uclz lowering if ufind_msb is either one
instruction better (Gfx7 and newer) or two instructions better (all
older platforms) than the ifind_msb implementations.

On platforms that use lower_find_msb_to_reverse, there should be no
difference.

All Haswell and newer Intel platforms had similar results. (Ice Lake shown)
total instructions in shared programs: 19938662 -> 19938634 (<.01%)
instructions in affected programs: 850 -> 822 (-3.29%)
helped: 2 / HURT: 0

total cycles in shared programs: 858467067 -> 858465538 (<.01%)
cycles in affected programs: 10080 -> 8551 (-15.17%)
helped: 2 / HURT: 0

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19042>

diff --git a/src/compiler/nir/nir_opt_algebraic.py b/src/compiler/nir/nir_opt_algebraic.py
index 2a58f91..0162c33 100644 (file)
--- a/src/compiler/nir/nir_opt_algebraic.py
+++ b/src/compiler/nir/nir_opt_algebraic.py
@@ -1814,6 +1814,8 @@ optimizations.extend([
    (('find_lsb', ('bitfield_reverse', a)), ('ufind_msb_rev', a), 'options->has_find_msb_rev'),
    (('ufind_msb_rev', ('bitfield_reverse', a)), ('find_lsb', a), '!options->lower_find_lsb'),
 
+   (('ifind_msb', ('f2i32(is_used_once)', a)), ('ufind_msb', ('f2i32', ('fabs', a)))),
+
    (('~fmul', ('bcsel(is_used_once)', c, -1.0, 1.0), b), ('bcsel', c, ('fneg', b), b)),
    (('~fmul', ('bcsel(is_used_once)', c, 1.0, -1.0), b), ('bcsel', c, b, ('fneg', b))),
    (('~fmulz', ('bcsel(is_used_once)', c, -1.0, 1.0), b), ('bcsel', c, ('fneg', b), b)),