[Intel-gfx] [PATCH] drm/i915: Support to create uncached user mapping for a Gem object

[Chris Wilson]

On Tue, Dec 09, 2014 at 08:48:38PM -0800, Chad Versace wrote:
> 
> 
> On 10/23/2014 06:23 AM, Chris Wilson wrote:
> > On Thu, Oct 23, 2014 at 12:56:46PM +0100, Chris Wilson wrote:
> >> On Thu, Oct 23, 2014 at 04:03:56PM +0530, akash.goel at intel.com wrote:
> >>> From: Akash Goel <akash.goel at intel.com>
> >>>
> >>
> >>> This is for improving the CPU write operation performance, as with such
> >>> mapping, writes are almost 50% faster than with mmap_gtt. Also it avoids the
> 
> 
> > Oops, bugs beware. I wasn't calling the ioctl correctly.
> > Having re-run the test, I now get:
> > 
> >    gtt       wc    Operation
> > --------  ------   ---------
> > 424000.0    1.30   ShmPutImage 10x10 square 
> >  29500.0    1.44   ShmPutImage 100x100 square 
> >   1510.0    1.38   ShmPutImage 500x500 square 
> > 
> > Supporting your claims. Very nice.
> 
> What are the units for these numbers?

ops/s, relative factor on a baby byt.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre