On Fri, 4 May 2012 23:00:32 +0200 Daniel Vetter <daniel at ffwll.ch> wrote: > On Fri, May 04, 2012 at 01:40:27PM -0700, Jesse Barnes wrote: > > On Fri, 4 May 2012 18:58:59 -0700 > > Ben Widawsky <ben at bwidawsk.net> wrote: > > > > > I'm seeing about a 5% FPS improvement across various benchmarks on my > > > IVB i3. Rumor has it that the higher end parts show even more benefit. > > > > > > This derives from a patch originally given to me by Bernard. The docs > > > are confusing about the definition names (ie. medium really seems > > > like max), but it would seem it gives more cache to the GT at the > > > expense of uncore. This configuration makes the split most in favor > > > of the GT. I've not tried the other IDICOS values. > > > > Bringing the IRC discussion here. I think this is really at the > > expense (though not really) of the CPU, since it allows the GPU to use > > more of the shared CPU/GPU cache in the uncore. > > > > You can fix up the #define names if you want; the MEDIUM etc may be > > from an old, incorrect doc. They were added awhile back. > > Another perk from me: Can you do some benchmarking of this for snb, too? > -Daniel I didn't see such gains on my SNB. I ran half as many benchmarks, one had almost 2% improvement, the other had a slight performance regression. Honestly, I'm a bit surprised by this, I may spend some more time to see what happens. -- Ben Widawsky, Intel Open Source Technology Center
