nir/algebraic: Don't distrubte absolute-value into dot-products

Dot product is multiplication followed by addition, and absolute value
does not distribute into addition.

Only vec4 platforms are affected by this change as scalar-only platforms
never have any of the fdot_replicated instructions.  In the shader-db
results, below, shaders in MANY different applications are affected.
Trine, Doom3, Enemy Territory: Quake Wars, Counter Strike: Global
Offensive, Mad Max, Metro Last Light, and on and on...  I'm really
shocked that there were no test regressions!

All Haswell and earlier platforms had similar results. (Haswell shown)
total instructions in shared programs: 16219743 -> 16219820 (<.01%)
instructions in affected programs: 12171 -> 12248 (0.63%)
helped: 1
HURT: 78
helped stats (abs) min: 1 max: 1 x̄: 1.00 x̃: 1
helped stats (rel) min: 0.78% max: 0.78% x̄: 0.78% x̃: 0.78%
HURT stats (abs)   min: 1 max: 1 x̄: 1.00 x̃: 1
HURT stats (rel)   min: 0.35% max: 2.38% x̄: 0.91% x̃: 1.06%
95% mean confidence interval for instructions value: 0.92 1.03
95% mean confidence interval for instructions %-change: 0.78% 1.00%
Instructions are HURT.

total cycles in shared programs: 538481383 -> 538491045 (<.01%)
cycles in affected programs: 470796 -> 480458 (2.05%)
helped: 149
HURT: 142
helped stats (abs) min: 1 max: 1338 x̄: 71.13 x̃: 4
helped stats (rel) min: 0.06% max: 40.99% x̄: 2.76% x̃: 0.67%
HURT stats (abs)   min: 1 max: 2092 x̄: 142.68 x̃: 12
HURT stats (rel)   min: 0.07% max: 55.38% x̄: 5.07% x̃: 1.07%
95% mean confidence interval for cycles value: -5.28 71.69
95% mean confidence interval for cycles %-change: -0.07% 2.19%
Inconclusive result (value mean confidence interval includes 0).

Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Fixes: 62795475e8f ("nir/algebraic: Distribute source modifiers into instructions")
Closes: #3129
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5581>

diff --git a/src/compiler/nir/nir_opt_algebraic.py b/src/compiler/nir/nir_opt_algebraic.py
index 575024c04bf45a19f2795f533d5492bdb70b69d6..4b5fc5ecc9abd8c8668ed29e7f5c4e29e99ed0f3 100644 (file)
--- a/src/compiler/nir/nir_opt_algebraic.py
+++ b/src/compiler/nir/nir_opt_algebraic.py
@@ -2041,6 +2041,9 @@ distribute_src_mods = [
    (('fdot_replicated4', ('fneg', a), ('fneg', b)), ('fdot_replicated4', a, b)),
    (('fneg', ('fneg', a)), a),
 
+   (('fneg', ('fmul(is_used_once)', a, b)), ('fmul', ('fneg', a), b)),
+   (('fabs', ('fmul(is_used_once)', a, b)), ('fmul', ('fabs', a), ('fabs', b))),
+
    (('fneg', ('ffma(is_used_once)', a, b, c)), ('ffma', ('fneg', a), b, ('fneg', c))),
    (('fneg', ('flrp(is_used_once)', a, b, c)), ('flrp', ('fneg', a), ('fneg', b), c)),
    (('fneg', ('fadd(is_used_once)', a, b)), ('fadd', ('fneg', a), ('fneg', b))),
@@ -2050,21 +2053,18 @@ distribute_src_mods = [
    (('fneg', ('fmin(is_used_once)', a, b)), ('fmax', ('fneg', a), ('fneg', b))),
    (('fneg', ('fmax(is_used_once)', a, b)), ('fmin', ('fneg', a), ('fneg', b))),
 
+   (('fneg', ('fdot_replicated2(is_used_once)', a, b)), ('fdot_replicated2', ('fneg', a), b)),
+   (('fneg', ('fdot_replicated3(is_used_once)', a, b)), ('fdot_replicated3', ('fneg', a), b)),
+   (('fneg', ('fdot_replicated4(is_used_once)', a, b)), ('fdot_replicated4', ('fneg', a), b)),
+
    # fdph works mostly like fdot, but to get the correct result, the negation
    # must be applied to the second source.
    (('fneg', ('fdph_replicated(is_used_once)', a, b)), ('fdph_replicated', a, ('fneg', b))),
-   (('fabs', ('fdph_replicated(is_used_once)', a, b)), ('fdph_replicated', ('fabs', a), ('fabs', b))),
 
    (('fneg', ('fsign(is_used_once)', a)), ('fsign', ('fneg', a))),
    (('fabs', ('fsign(is_used_once)', a)), ('fsign', ('fabs', a))),
 ]
 
-for op in ['fmul', 'fdot_replicated2', 'fdot_replicated3', 'fdot_replicated4']:
-   distribute_src_mods.extend([
-       (('fneg', (op + '(is_used_once)', a, b)), (op, ('fneg', a), b)),
-       (('fabs', (op + '(is_used_once)', a, b)), (op, ('fabs', a), ('fabs', b))),
-   ])
-
 print(nir_algebraic.AlgebraicPass("nir_opt_algebraic", optimizations).render())
 print(nir_algebraic.AlgebraicPass("nir_opt_algebraic_before_ffma",
                                   before_ffma_optimizations).render())