Mesa (master): nir/load_store_vectorize: don't ignore subgroup memory barriers

GitLab Mirror gitlab-mirror at kemper.freedesktop.org
Thu Jan 7 16:52:31 UTC 2021


Module: Mesa
Branch: master
Commit: f4eb833a12523142d908bf67e08904f4e8a866e9
URL:    http://cgit.freedesktop.org/mesa/mesa/commit/?id=f4eb833a12523142d908bf67e08904f4e8a866e9

Author: Rhys Perry <pendingchaos02 at gmail.com>
Date:   Fri Mar 13 16:07:26 2020 +0000

nir/load_store_vectorize: don't ignore subgroup memory barriers

Not sure why I thought this was correct, but we should consider them for
optimization purposes.

Fixes: ce9205c03bd ('nir: add a load/store vectorization pass')
Signed-off-by: Rhys Perry <pendingchaos02 at gmail.com>
Reviewed-by: Jason Ekstrand <jason at jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4202>

---

 src/compiler/nir/nir_opt_load_store_vectorize.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/src/compiler/nir/nir_opt_load_store_vectorize.c b/src/compiler/nir/nir_opt_load_store_vectorize.c
index e52006e40ef..41c968bfb25 100644
--- a/src/compiler/nir/nir_opt_load_store_vectorize.c
+++ b/src/compiler/nir/nir_opt_load_store_vectorize.c
@@ -1195,7 +1195,6 @@ handle_barrier(struct vectorize_ctx *ctx, bool *progress, nir_function_impl *imp
          release = nir_intrinsic_memory_semantics(intrin) & NIR_MEMORY_RELEASE;
          switch (nir_intrinsic_memory_scope(intrin)) {
          case NIR_SCOPE_INVOCATION:
-         case NIR_SCOPE_SUBGROUP:
             /* a barier should never be required for correctness with these scopes */
             modes = 0;
             break;



More information about the mesa-commit mailing list