ATI 6.12.4 on ultraSparc runs slow

[Michel Dänzer]

On Mit, 2010-03-24 at 00:41 +0100, jean-francois simon wrote: 
> 2010/3/24 Michel Dänzer <michel at daenzer.net>
>         
>         
>         >
>         > so i am assuming that most of the work is done by some X
>         server
>         > "painting code" to the shadowfb in system ram, which is then
>         DMA'ed to
>         > the scan out buffer in the GPU VRAM.
>         
>         
>         It's only DMA'ed if you implemented that. ;) 
> 
> so it is PIO not DMA? just curious, where is the code that is doing
> that transfer?

shadowUpdatePacked() in xserver/miext/shadow/shpacked.c .


>         > since i see a big difference between solaris/sparc and
>         linux/x86-64,
>         > one idea that comes to mind is that there could some
>         overhead due to
>         > the difference in endianess between the solaris and the
>         linux box.
>         
>         
>         I don't think that matters here, we can use GPU facilities to
>         handle
>         endianness for direct CPU access to the framebuffer.
>         
> 
> sorry i dont understand what you mean. are you saying i should no
> longer use shadowfb?

No, just that the code copying from the shadow buffer to VRAM doesn't
need to do anything about endianness. 

-- 
Earthling Michel Dänzer           |                http://www.vmware.com
Libre software enthusiast         |          Debian, X and DRI developer