Mesa (main): mesa: Follow ASTC decode mode extension for RGBA8 output

[GitLab Mirror]

Module: Mesa
Branch: main
Commit: a07919edebfa95b990b386a3a74076db9311130d
URL:    http://cgit.freedesktop.org/mesa/mesa/commit/?id=a07919edebfa95b990b386a3a74076db9311130d

Author: Nanley Chery <nanley.g.chery at intel.com>
Date:   Tue Apr 12 15:00:23 2022 -0700

mesa: Follow ASTC decode mode extension for RGBA8 output

The GL extension, EXT_texture_compression_astc_decode_mode, enables
applications to specify the desired decoding precision when decoding
non-sRGB ASTC textures. The options for the channels are FP16 (the
default), UNORM8, and RGB9_E5.

The ASTC LDR decoder outputs to UNORM8 by doing the following
conversions: UNORM16 -> FP16 -> UNORM8. This doesn't seem to be defined
by any specification and is costly according to perf profiles. To
conform to the decode mode spec (and for better performance), we convert
UNORM16 to UNORM8 by simply storing/keeping the top 8 bits.

In a texture upload microbenchmark, this decreases the upload time for
textures in the linear color space by about 34%.

Reviewed-by: Kenneth Graunke <kenneth at whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17195>

---

 src/mesa/main/texcompress_astc.cpp | 36 ++++++++----------------------------
 1 file changed, 8 insertions(+), 28 deletions(-)

diff --git a/src/mesa/main/texcompress_astc.cpp b/src/mesa/main/texcompress_astc.cpp
index 660c53ad57a..e628c4c81bf 100644
--- a/src/mesa/main/texcompress_astc.cpp
+++ b/src/mesa/main/texcompress_astc.cpp
@@ -42,12 +42,6 @@
 static bool VERBOSE_DECODE = false;
 static bool VERBOSE_WRITE = false;
 
-static inline uint8_t
-uint16_div_64k_to_half_to_unorm8(uint16_t v)
-{
-   return _mesa_half_to_unorm8(_mesa_uint16_div_64k_to_half(v));
-}
-
 class decode_error
 {
 public:
@@ -1614,17 +1608,10 @@ void Block::write_decoded(const Decoder &decoder, uint16_t *output)
    if (is_void_extent) {
       for (int idx = 0; idx < decoder.block_w*decoder.block_h*decoder.block_d; ++idx) {
          if (decoder.output_unorm8) {
-            if (decoder.srgb) {
-               output[idx*4+0] = void_extent_colour_r >> 8;
-               output[idx*4+1] = void_extent_colour_g >> 8;
-               output[idx*4+2] = void_extent_colour_b >> 8;
-               output[idx*4+3] = void_extent_colour_a >> 8;
-            } else {
-               output[idx*4+0] = uint16_div_64k_to_half_to_unorm8(void_extent_colour_r);
-               output[idx*4+1] = uint16_div_64k_to_half_to_unorm8(void_extent_colour_g);
-               output[idx*4+2] = uint16_div_64k_to_half_to_unorm8(void_extent_colour_b);
-               output[idx*4+3] = uint16_div_64k_to_half_to_unorm8(void_extent_colour_a);
-            }
+            output[idx*4+0] = void_extent_colour_r >> 8;
+            output[idx*4+1] = void_extent_colour_g >> 8;
+            output[idx*4+2] = void_extent_colour_b >> 8;
+            output[idx*4+3] = void_extent_colour_a >> 8;
          } else {
             /* Store the color as FP16. */
             output[idx*4+0] = _mesa_uint16_div_64k_to_half(void_extent_colour_r);
@@ -1700,17 +1687,10 @@ void Block::write_decoded(const Decoder &decoder, uint16_t *output)
             };
 
             if (decoder.output_unorm8) {
-               if (decoder.srgb) {
-                  output[idx*4+0] = c[0] >> 8;
-                  output[idx*4+1] = c[1] >> 8;
-                  output[idx*4+2] = c[2] >> 8;
-                  output[idx*4+3] = c[3] >> 8;
-               } else {
-                  output[idx*4+0] = c[0] == 65535 ? 0xff : uint16_div_64k_to_half_to_unorm8(c[0]);
-                  output[idx*4+1] = c[1] == 65535 ? 0xff : uint16_div_64k_to_half_to_unorm8(c[1]);
-                  output[idx*4+2] = c[2] == 65535 ? 0xff : uint16_div_64k_to_half_to_unorm8(c[2]);
-                  output[idx*4+3] = c[3] == 65535 ? 0xff : uint16_div_64k_to_half_to_unorm8(c[3]);
-               }
+               output[idx*4+0] = c[0] >> 8;
+               output[idx*4+1] = c[1] >> 8;
+               output[idx*4+2] = c[2] >> 8;
+               output[idx*4+3] = c[3] >> 8;
             } else {
                /* Store the color as FP16. */
                output[idx*4+0] = c[0] == 65535 ? FP16_ONE : _mesa_uint16_div_64k_to_half(c[0]);