nir: intel/compiler: Move ufind_msb lowering to NIR

Fossil-db results:

All Intel platforms had similar results. (Ice Lake shown)
Cycles in all programs: 9098346105 -> 9098333765 (-0.0%)
Cycles helped: 6

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

src/compiler/nir/nir.h

@@ -3413,6 +3413,8 @@ typedef struct nir_shader_compiler_options {
    bool lower_find_msb_to_reverse;
    /** Lowers ifind_msb to uclz and logic ops*/
    bool lower_ifind_msb_to_uclz;
+   /** Lowers ufind_msb to 31-uclz */
+   bool lower_ufind_msb_to_uclz;
    /** Lowers find_lsb to ufind_msb and logic ops */
    bool lower_find_lsb;
    bool lower_uadd_carry;

src/compiler/nir/nir_opt_algebraic.py

@@ -2041,6 +2041,10 @@ optimizations.extend([
      ('ufind_msb_rev', 'value')),
     'options->lower_find_msb_to_reverse'),
 
+   (('ufind_msb', 'value@32'),
+    ('isub', 31, ('uclz', 'value')),
+    'options->lower_ufind_msb_to_uclz'),
+
    (('uclz', a), ('umin', 32, ('ufind_msb_rev', a)), 'options->lower_uclz'),
 
    (('find_lsb', 'value'),