[Mesa-dev] [PATCH v3 13/14] squash i965/nir: Don't lower invocation index from local ID

Jordan Justen jordan.l.justen at intel.com
Sun May 29 22:38:49 UTC 2016


With our newer code, it is now more efficient to generate the ID from
the local invocation index. Therefore, we want to prevent nir core
code from deriving local invocation index from the local invocation
ID.

Signed-off-by: Jordan Justen <jordan.l.justen at intel.com>
---
 src/mesa/drivers/dri/i965/brw_compiler.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/src/mesa/drivers/dri/i965/brw_compiler.c b/src/mesa/drivers/dri/i965/brw_compiler.c
index bb06733..a4855a0 100644
--- a/src/mesa/drivers/dri/i965/brw_compiler.c
+++ b/src/mesa/drivers/dri/i965/brw_compiler.c
@@ -40,8 +40,7 @@
    .lower_fdiv = true,                                                        \
    .lower_flrp64 = true,                                                      \
    .native_integers = true,                                                   \
-   .vertex_id_zero_based = true,                                              \
-   .lower_cs_local_index_from_id = true
+   .vertex_id_zero_based = true
 
 static const struct nir_shader_compiler_options scalar_nir_options = {
    COMMON_OPTIONS,
-- 
2.8.1



More information about the mesa-dev mailing list