Mesa (master): tgsi: Mark the tgsi_exec_channel and tgsi_double_channel ALIGN16.

[GitLab Mirror]

Module: Mesa
Branch: master
Commit: a2efa2e833ded46da03666253de0d0e9f77beaf2
URL:    http://cgit.freedesktop.org/mesa/mesa/commit/?id=a2efa2e833ded46da03666253de0d0e9f77beaf2

Author: Eric Anholt <eric at anholt.net>
Date:   Mon May  3 15:19:46 2021 -0700

tgsi: Mark the tgsi_exec_channel and tgsi_double_channel ALIGN16.

We allocate them all align16, so mark the unions (and their container
structs) that way so the compiler can do aligned SSE load/stores.

glmark2 -b loop FPS +0.197265% +/- 0.117633% (n=1906)

Reviewed-by: Adam Jackson <ajax at redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10604>

---

 src/gallium/auxiliary/tgsi/tgsi_exec.c | 4 ++--
 src/gallium/auxiliary/tgsi/tgsi_exec.h | 6 +++---
 2 files changed, 5 insertions(+), 5 deletions(-)

diff --git a/src/gallium/auxiliary/tgsi/tgsi_exec.c b/src/gallium/auxiliary/tgsi/tgsi_exec.c
index fd0471fd2f5..cf950e89fb3 100644
--- a/src/gallium/auxiliary/tgsi/tgsi_exec.c
+++ b/src/gallium/auxiliary/tgsi/tgsi_exec.c
@@ -80,9 +80,9 @@ union tgsi_double_channel {
    unsigned u[TGSI_QUAD_SIZE][2];
    uint64_t u64[TGSI_QUAD_SIZE];
    int64_t i64[TGSI_QUAD_SIZE];
-};
+} ALIGN16;
 
-struct tgsi_double_vector {
+struct ALIGN16 tgsi_double_vector {
    union tgsi_double_channel xy;
    union tgsi_double_channel zw;
 };
diff --git a/src/gallium/auxiliary/tgsi/tgsi_exec.h b/src/gallium/auxiliary/tgsi/tgsi_exec.h
index 982cad592e0..257769819eb 100644
--- a/src/gallium/auxiliary/tgsi/tgsi_exec.h
+++ b/src/gallium/auxiliary/tgsi/tgsi_exec.h
@@ -76,12 +76,12 @@ union tgsi_exec_channel
    float    f[TGSI_QUAD_SIZE];
    int      i[TGSI_QUAD_SIZE];
    unsigned u[TGSI_QUAD_SIZE];
-};
+} ALIGN16;
 
 /**
   * A vector[RGBA] of channels[4 pixels]
   */
-struct tgsi_exec_vector
+struct ALIGN16 tgsi_exec_vector
 {
    union tgsi_exec_channel xyzw[TGSI_NUM_CHANNELS];
 };
@@ -286,7 +286,7 @@ typedef void (* apply_sample_offset_func)(
 /**
  * Run-time virtual machine state for executing TGSI shader.
  */
-struct tgsi_exec_machine
+struct ALIGN16 tgsi_exec_machine
 {
    /* Total = program temporaries + internal temporaries
     */