By: blue (blue.delete@this.blue.com), August 1, 2016 8:58 am
Room: Moderated Discussions
Andrei Frumusanu (andrei.delete@this.anandtech.com) on August 1, 2016 6:21 am wrote:
> John H (john.heritage.delete@this.gmail.com) on August 1, 2016 6:13 am wrote:
> > David Kanter (dkanter.delete@this.realworldtech.com) on August 1, 2016 12:01 am wrote:
> > > This is my first new article in a while, but its a treat. It's the first video I've done for the site.
> > >
> > > Starting with the Maxwell and Pascal architectures, Nvidia high-performance GPUs use tile-based
> > > immediate-mode rasterizers, instead of conventional full-screen immediate-mode rasterizers. Using
> > > simple DirectX shaders, we demonstrate the tile-based rasterization in Nvidia's Maxwell and Pascal
> > > GPUs and contrast this behavior to the immediate-mode rasterizer used by AMD.
> > >
> > > http://www.realworldtech.com/tile-based-rasterization-nvidia-gpus/
> > >
> > > I look forward to the discussion.
> > >
> > > David
> >
> > David - excellent video and article. You deserve much respect as usual..
> >
> > In your estimation - if there were an 'incompatible situation' for this type of
> > rendering/rasterization, could the drivers switch Maxwell back to the 'traditional
> > PC mode' for output? If yes, any idea of what kind of performance hit?.
> >
> > Could this explain why Polaris hasn't caught up with Pascal on the perf/watt efficiency curve,
> > despite the "in theory" advantage of Samsung/GloFo's 14nm process over TSMC's 16nm? (as evidenced
> > by the Apple A9 that is slightly smaller and lower power on the Samsung 14nm process).
> >
> > Thanks!
> > John H
>
> 14LPP has a density advantage but it has a power disadvantage (~10%)
> over 16FF+. TSMC A9's were lower power than Samsung ones.
>
Samsung A9 was LPE, and would be equal to TSMC 16FF. TSMC ended up canning it's 16FF and did 16FF+ as the initial process.
To LPP should close most or all of the gap.
> John H (john.heritage.delete@this.gmail.com) on August 1, 2016 6:13 am wrote:
> > David Kanter (dkanter.delete@this.realworldtech.com) on August 1, 2016 12:01 am wrote:
> > > This is my first new article in a while, but its a treat. It's the first video I've done for the site.
> > >
> > > Starting with the Maxwell and Pascal architectures, Nvidia high-performance GPUs use tile-based
> > > immediate-mode rasterizers, instead of conventional full-screen immediate-mode rasterizers. Using
> > > simple DirectX shaders, we demonstrate the tile-based rasterization in Nvidia's Maxwell and Pascal
> > > GPUs and contrast this behavior to the immediate-mode rasterizer used by AMD.
> > >
> > > http://www.realworldtech.com/tile-based-rasterization-nvidia-gpus/
> > >
> > > I look forward to the discussion.
> > >
> > > David
> >
> > David - excellent video and article. You deserve much respect as usual..
> >
> > In your estimation - if there were an 'incompatible situation' for this type of
> > rendering/rasterization, could the drivers switch Maxwell back to the 'traditional
> > PC mode' for output? If yes, any idea of what kind of performance hit?.
> >
> > Could this explain why Polaris hasn't caught up with Pascal on the perf/watt efficiency curve,
> > despite the "in theory" advantage of Samsung/GloFo's 14nm process over TSMC's 16nm? (as evidenced
> > by the Apple A9 that is slightly smaller and lower power on the Samsung 14nm process).
> >
> > Thanks!
> > John H
>
> 14LPP has a density advantage but it has a power disadvantage (~10%)
> over 16FF+. TSMC A9's were lower power than Samsung ones.
>
Samsung A9 was LPE, and would be equal to TSMC 16FF. TSMC ended up canning it's 16FF and did 16FF+ as the initial process.
To LPP should close most or all of the gap.
