[PATCH xserver 0/3] glamor: Accelerate XY format images

[Adam Jackson]

On Fri, 2016-09-30 at 22:43 -0700, Keith Packard wrote:

>        1                 2                 Operation
> ------------   -------------------------   -------------------------
>      10900.0        99900.0 (     9.165)   PutImage XY 10x10 square 
>       1740.0         2160.0 (     1.241)   PutImage XY 100x100 square 
>         83.2           90.4 (     1.087)   PutImage XY 500x500 square 

90 ops per second is still shameful. No blit should take 11ms.

>      11800.0       351000.0 (    29.746)   PutImage XYBitmap 10x10 square 
>       6280.0        81600.0 (    12.994)   PutImage XYBitmap 100x100 square 
>        752.0        14300.0 (    19.016)   PutImage XYBitmap 500x500 square 

Where does one get a copy of x11perf that has this test?

Also I assume this is measuring before the whole series vs after. Would
be nice to see the impact between 2/3 and 3/3 too.

> I'd like to know why the GPU expansion version has so much GL
> overhead; it "should" be faster for everything as it uploads a lot
> less data to the GPU.

On i965, at least, scissor updates are an appreciable amount of gl
driver time, so small ops get punished.

> I'd also like to try writing an XYPixmap that
> did all of the plane merging in the fragment shader. That "should" be
> faster for large images than merging on the CPU.

With a trivial planemask, sure. With an interesting planemask you
either need a bitwise write mask to the output fragment (and I'm not
sure glsl 1.30 gives you that) or you need one of the fb fetch
extensions.

- ajax