On 02/18/2016 03:03 PM, Roland Scheidegger wrote:
Looks ok to me, albeit the sse2 code I posted a while ago for handling
swizzled stores probably would be faster still (though only with x86
sse2...). At least I think it would hit this (if not the next patch in
the series should...).
It would be
Looks ok to me, albeit the sse2 code I posted a while ago for handling
swizzled stores probably would be faster still (though only with x86
sse2...). At least I think it would hit this (if not the next patch in
the series should...).
Matt Turner reviewed the first 2 patches in the series (albeit I
On 18/02/16 15:52, Brian Paul wrote:
This reduces a glTexImage(GL_RGBA, GL_UNSIGNED_BYTE) hot spot in when
storing the texture as BGRA.
---
src/mesa/main/format_utils.c | 74 +---
1 file changed, 69 insertions(+), 5 deletions(-)
diff --git a/src/mesa/ma
This reduces a glTexImage(GL_RGBA, GL_UNSIGNED_BYTE) hot spot in when
storing the texture as BGRA.
---
src/mesa/main/format_utils.c | 74 +---
1 file changed, 69 insertions(+), 5 deletions(-)
diff --git a/src/mesa/main/format_utils.c b/src/mesa/main/format_
