Mesa (main): freedreno/ir3: Call nir_opt_find_array_copies().

[GitLab Mirror]

Module: Mesa
Branch: main
Commit: 1bcd848816f3c9abf40362fd3ca3ca175774094b
URL:    http://cgit.freedesktop.org/mesa/mesa/commit/?id=1bcd848816f3c9abf40362fd3ca3ca175774094b

Author: Emma Anholt <emma at anholt.net>
Date:   Fri Apr 15 12:08:07 2022 -0700

freedreno/ir3: Call nir_opt_find_array_copies().

gfxbench vk-5-normal has a shader that sampels into a texels[] array at
the top, then in a loop calls a GLSL function passing texels[] in by
value.  This resulted in a copy to a temp inside the loop, which got
lowered to scratch stores since it was pretty big.

By doing find_array_copies(), we notice that it's equivalent to
copy_deref, then get to copy-propagate from the array at the top.  Then we
only have to set up the scratch array outside of the loop and load_scratch
from it in the called function inside the loop.  This also causes there to
be less spilling, stps 1144 -> 354 and ldps 826->36.

However, it doesn't seem to change performance on the test.  So, while
this seems to be an improvement for the shader, and we could maybe even do
better by rematerializing the txl samples inside the loop instead of
storing the texture fetches to scratch in the first place, it doesn't
currently seem worth pursuing more optimization of this shader.

No change on freedreno shader-db.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15982>

---

 src/freedreno/ir3/ir3_nir.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/src/freedreno/ir3/ir3_nir.c b/src/freedreno/ir3/ir3_nir.c
index 96529f35727..c08377e8ab9 100644
--- a/src/freedreno/ir3/ir3_nir.c
+++ b/src/freedreno/ir3/ir3_nir.c
@@ -81,8 +81,6 @@ ir3_optimize_loop(struct ir3_compiler *compiler, nir_shader *s)
       progress = false;
 
       OPT_V(s, nir_lower_vars_to_ssa);
-      progress |= OPT(s, nir_opt_copy_prop_vars);
-      progress |= OPT(s, nir_opt_dead_write_vars);
       progress |= OPT(s, nir_lower_alu_to_scalar, NULL, NULL);
       progress |= OPT(s, nir_lower_phis_to_scalar, false);
 
@@ -90,6 +88,11 @@ ir3_optimize_loop(struct ir3_compiler *compiler, nir_shader *s)
       progress |= OPT(s, nir_opt_deref);
       progress |= OPT(s, nir_opt_dce);
       progress |= OPT(s, nir_opt_cse);
+
+      progress |= OPT(s, nir_opt_find_array_copies);
+      progress |= OPT(s, nir_opt_copy_prop_vars);
+      progress |= OPT(s, nir_opt_dead_write_vars);
+
       static int gcm = -1;
       if (gcm == -1)
          gcm = env_var_as_unsigned("GCM", 0);
@@ -165,6 +168,8 @@ ir3_optimize_loop(struct ir3_compiler *compiler, nir_shader *s)
       progress |= OPT(s, nir_opt_remove_phis);
       progress |= OPT(s, nir_opt_undef);
    } while (progress);
+
+   OPT(s, nir_lower_var_copies);
 }
 
 static bool