By: jp (jipe4153.delete@this.gmail.com), July 27, 2012 9:47 am
Room: Moderated Discussions
aaron spink (aaronspink.delete@this.notearthlink.net) on July 27, 2012 9:36 am wrote:
> jp (jipe4153.delete@this.gmail.com) on July 27, 2012 9:18 am wrote:
> > About
> bandwidth, the GPUs already
> > have the fastest RAM out there ( over 250
> GB/s ) and they have no reason not to
> > continue this lead (read mentioned
> FLOPS/bandwidth ratio "issue")
> >
>
> GPUs have high local bandwidth but
> rather poor global bandwidth. Not to mention rather limited capacity.
>
Global bandwidth on GPUs is much higher than on any other CPU ( 5-6 times higher), you seem to have no backing for your claim.
> >
> Looking at the articles at hpcwire about
> > new clusters over the last 1.5
> years its obvious that almost everyone is buying
> > Nvidia Tesla cards. In
> fact ORNL:s newest cluster (fastest in the US) will be
> > based on the new
> Kepler cards (K20).
> >
> Very few are buying Tesla cards. And most of the
> data says they are not any better than CPUs with much more complex programming
> models.
Could you please refer me to "most of the data"? Again you have no backing for these claims and opinions of yours.
Here's a list of ~1300 applications that have had great benefits from GPU acceleration in a wide range of fields: http://www.nvidia.com/object/cuda-apps-flash-new.html#.
You see that above is called a referene.
> jp (jipe4153.delete@this.gmail.com) on July 27, 2012 9:18 am wrote:
> > About
> bandwidth, the GPUs already
> > have the fastest RAM out there ( over 250
> GB/s ) and they have no reason not to
> > continue this lead (read mentioned
> FLOPS/bandwidth ratio "issue")
> >
>
> GPUs have high local bandwidth but
> rather poor global bandwidth. Not to mention rather limited capacity.
>
Global bandwidth on GPUs is much higher than on any other CPU ( 5-6 times higher), you seem to have no backing for your claim.
> >
> Looking at the articles at hpcwire about
> > new clusters over the last 1.5
> years its obvious that almost everyone is buying
> > Nvidia Tesla cards. In
> fact ORNL:s newest cluster (fastest in the US) will be
> > based on the new
> Kepler cards (K20).
> >
> Very few are buying Tesla cards. And most of the
> data says they are not any better than CPUs with much more complex programming
> models.
Could you please refer me to "most of the data"? Again you have no backing for these claims and opinions of yours.
Here's a list of ~1300 applications that have had great benefits from GPU acceleration in a wide range of fields: http://www.nvidia.com/object/cuda-apps-flash-new.html#.
You see that above is called a referene.
