Mesa (staging/21.2): Revert "nir/algebraic: Convert some f2u to f2i"

[GitLab Mirror]

Module: Mesa
Branch: staging/21.2
Commit: 5fb386c48dcfb6ca1db0f1d1987625bd07ba2787
URL:    http://cgit.freedesktop.org/mesa/mesa/commit/?id=5fb386c48dcfb6ca1db0f1d1987625bd07ba2787

Author: Ian Romanick <ian.d.romanick at intel.com>
Date:   Fri Aug  6 14:20:25 2021 -0700

Revert "nir/algebraic: Convert some f2u to f2i"

Per https://gitlab.freedesktop.org/mesa/mesa/-/issues/5178#note_1019666,
the assumption fundamental to this optimization is false.  Section
2.4.1 (Float to Integer) of Ivy Bridge PRMs describes the situation.
The wording of the section is somewhat confusing (because it doesn't
clearly delineate between signed and unsigned integers), but the last
two rows of the table make it clear that F->UD conversion clamps
negative float values to 0.

All other hardware mentioned in that thread seems to behave the same
way.

The real problem is that, with hardware that behaves in this ways,
converting f2u(2147483648.0) to f2i(2147483648.0) changes the bit pattern
that would be produced from 0x80000000 to 0x7fffffff.

This reverts commit ad059202583e8c86bbccf0d65c5ce35bc4ab20f1.

Reviewed-by: Jason Ekstrand <jason at jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12297>
(cherry picked from commit 84d2e537890acea01da2795e5dd3bac1319635f5)

---

 .pick_status.json                     |  2 +-
 src/compiler/nir/nir_opt_algebraic.py | 13 -------------
 2 files changed, 1 insertion(+), 14 deletions(-)

diff --git a/.pick_status.json b/.pick_status.json
index 51ab287a53a..be26ffa833f 100644
--- a/.pick_status.json
+++ b/.pick_status.json
@@ -202,7 +202,7 @@
         "description": "Revert \"nir/algebraic: Convert some f2u to f2i\"",
         "nominated": true,
         "nomination_type": 2,
-        "resolution": 0,
+        "resolution": 1,
         "main_sha": null,
         "because_sha": "ad059202583e8c86bbccf0d65c5ce35bc4ab20f1"
     },
diff --git a/src/compiler/nir/nir_opt_algebraic.py b/src/compiler/nir/nir_opt_algebraic.py
index 31e9d83e454..99078c66696 100644
--- a/src/compiler/nir/nir_opt_algebraic.py
+++ b/src/compiler/nir/nir_opt_algebraic.py
@@ -1155,19 +1155,6 @@ optimizations.extend([
    (('iabs', 'a(is_not_negative)'), a),
    (('fsat', 'a(is_not_positive)'), 0.0),
 
-   # Section 5.4.1 (Conversion and Scalar Constructors) of the GLSL 4.60 spec
-   # says:
-   #
-   #    It is undefined to convert a negative floating-point value to an
-   #    uint.
-   #
-   # Assuming that (uint)some_float behaves like (uint)(int)some_float allows
-   # some optimizations in the i965 backend to proceed.
-   (('ige', ('f2u', a), b), ('ige', ('f2i', a), b)),
-   (('ige', b, ('f2u', a)), ('ige', b, ('f2i', a))),
-   (('ilt', ('f2u', a), b), ('ilt', ('f2i', a), b)),
-   (('ilt', b, ('f2u', a)), ('ilt', b, ('f2i', a))),
-
    (('~fmin', 'a(is_not_negative)', 1.0), ('fsat', a), '!options->lower_fsat'),
 
    # The result of the multiply must be in [-1, 0], so the result of the ffma