Page 1 of 1

Nvidia Maxwell

Posted: Thu Feb 20, 2014 6:53 am
by iceman1992
The new 750 Ti seems to have an edge in Explicit, Single Precision computation, even compared to the faster 760. Would this mean better PPD? http://www.anandtech.com/bench/product/1130?vs=1038

Re: Nvidia Maxwell

Posted: Thu Feb 20, 2014 9:35 am
by P5-133XL
Anandtech runs FAHBench as one of their benchmarks. You should be able to cross-compare folding potential from that.

Re: Nvidia Maxwell

Posted: Thu Feb 20, 2014 12:06 pm
by OzPapaSmurf
I would think it will be a better platform as it has 640 cuda cores on the Ti as opposed to 512 on the GTX 750 and 384 on the Kepler.
It will take a few months for us here in Australia to see any of these and they will be MUCH dearer than for buyers in America
and other overseas countries.

At $149.00 for the 2gb model, it will sure be a boost to people building FAH farms. :)

Re: Nvidia Maxwell

Posted: Thu Feb 20, 2014 2:49 pm
by VijayPande
I have high hopes for Maxwell. We'll be optimizing OpenMM for it.

Re: Nvidia Maxwell

Posted: Thu Feb 20, 2014 3:31 pm
by mdk777
I thought Folding attempted to remain hardware neutral.
I don't recall any similar effort to optimize for GCN.

Is Folding going in a new direction with this initiative?

Re: Nvidia Maxwell

Posted: Thu Feb 20, 2014 5:14 pm
by 7im
mdk777 wrote:I thought Folding attempted to remain hardware neutral.
I don't recall any similar effort to optimize for GCN.

Is Folding going in a new direction with this initiative?
You thought correctly.

However, every comment about a new advancement does NOT have to mention both Intel and AMD, or Windows and OSX, or NV and ATI to be considered a fair and balanced comment. There is no conspiracy or change in policy when a comment only mentions one side of the equation. And a lack of comment on the other side is just that. Nothing. No comment. No implied message. No favoritism. Just a comment.

If you would like to know more about the OpenMM initiatives, please feel free to read about them on their web site and look at their code optimizations on GitHub. It's all Open Sourced and very transparent, if you would just look instead of seeing a problem where there is none.

And considering the latest AMD beta driver (14.1) gave the AMD cards a big boost in folding performance, the software seems to be well optimized for AMD cards already, even though the GNC SDK was only released a few months ago. ;)

Re: Nvidia Maxwell

Posted: Sun Feb 23, 2014 1:56 am
by decali
7im wrote:
mdk777 wrote:I thought Folding attempted to remain hardware neutral.
I don't recall any similar effort to optimize for GCN.

Is Folding going in a new direction with this initiative?
You thought correctly.

However, every comment about a new advancement does NOT have to mention both Intel and AMD, or Windows and OSX, or NV and ATI to be considered a fair and balanced comment. There is no conspiracy or change in policy when a comment only mentions one side of the equation. And a lack of comment on the other side is just that. Nothing. No comment. No implied message. No favoritism. Just a comment.

If you would like to know more about the OpenMM initiatives, please feel free to read about them on their web site and look at their code optimizations on GitHub. It's all Open Sourced and very transparent, if you would just look instead of seeing a problem where there is none.

And considering the latest AMD beta driver (14.1) gave the AMD cards a big boost in folding performance, the software seems to be well optimized for AMD cards already, even though the GNC SDK was only released a few months ago. ;)
Thanks 7im, that's well put.

Advancing technology, whether it's better hardware or better software, is good for FAH.

Re: Nvidia Maxwell

Posted: Tue Feb 25, 2014 6:32 am
by OzPapaSmurf
decali wrote: Thanks 7im, that's well put.
Advancing technology, whether it's better hardware or better software, is good for FAH.
I also agree. FAH must use the latest and most favourable technologies available.