[Mesa-dev] [PATCH 1/2 v3] i965: add XRGB to tiled_memcpy
courtney at lunarg.com
Tue Nov 12 09:04:56 PST 2013
Does Matt's explanation help?
I wanted to tell the performance story and show the progression of
performance enhancements with these two patches.
On Mon, Nov 11, 2013 at 3:01 PM, Matt Turner <mattst88 at gmail.com> wrote:
> On Mon, Nov 11, 2013 at 1:19 PM, Chad Versace
> <chad.versace at linux.intel.com> wrote:
> > On 11/07/2013 01:59 PM, Courtney Goeltzenleuchter wrote:
> >> MESA_FORMAT_XRGB8888 is equivalent to MESA_FORMAT_ARGB8888 in terms
> >> of storage on the device, so okay to use this optimized copy routine.
> >> This series builds on work from Frank Henigman to optimize the
> >> process of uploading a texture to the GPU. This series adds support for
> >> MESA_XRGB_8888 and full miptrees where were found to be common
> >> in the Smokin' Guns game. The issue was found while profiling the app
> >> but that part is not benchmarked. Smokin-Guns uses mipmap textures with
> >> an internal format of GL_RGB (MESA_XRGB_8888 in the driver).
> >> These changes need a performance tool to run against to show how they
> >> improve execution performance for specific texture formats. Using this
> >> benchmark I've measured the following improvement on my Ivybridge
> >> Intel(R) Xeon(R) CPU E3-1225 V2 @ 3.20GHz.
> >> Using 1024x1024, RGBA 8888 source, mipmap
> >> <<THIS PATCH>>
> > I don't understand. What do you mean by ``<<THIS PATCH>>``? That all
> > numbers were obtained with this patch? But that doesn't make sense,
> > these are before-and-after numbers. And it can't be just this patch,
> > these numbers are identical to the numbers quoted in patch 2.
> The first column is "before", the second is after patch 1, and the
> third is after patch 2. Instead of doing before->after patch 1 in
> patch 1's commit, and after patch 1->after patch 2 in patch 2's
> commit, he just pasted the same data in and added <<THIS PATCH>> above
> the column that corresponds to the patch.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the mesa-dev