Some more details on the GPU3 core, regarding OpenCL

[GlitchPC]

Here is an update and some more details on the third generation GPU core. As I stated before, this core is built on OpenMM, which brings the science further along and allows us to make easier updates. OpenMM supports OpenCL...

More... (http://folding.typepad.com/news/2010/01/some-more-details-on-the-gpu3-core-regarding-opencl.html)

[jjFarking]

.. This means that the core will only roll out for NVIDIA first. ATI has depreciated Brook, but does not have a fully-functioning OpenCL implementation, so we are stuck in between support on the ATI side

What the...?
It most certainly does and in preliminary tests is actually faster than nVidia's implementation.
When I still had the 9800GT in my system, I ran OpenCLBench on both, only to find that my 5850's score was almost triple that of the 9800GT..

That's just a crud statement..
The more I look around the forum there, the more I see fully-fledged support for anything Intel/nVidia related, but sweet fa for AMD/ATI hardware.
Are they sponsored by those camps or something?

/rant
:p

EDIT: OpenCL support has been advanced in the 9.12 HotFix & will be further expanded in the about-to-be-released v10(.1) Catalyst drivers, together with DirectCompute. F@H is in an ideal situation to make use of both of these implementations, in order to facilitate huge speed advantages. Also, only AMD can offer the advantage of double-precision calculations for projects that require extra-ultra-mega-careful precision for their respective outcomes.
A quick Google around the web, shows that there are many amateur coders out there, who can easily implement both OpenCL & DirectCompute for their projects (including one guy who got the attention from AMD about how efficient his coding was - more so than their own), so there's no excuse for the good people at Stanford to not come up with something very decent indeed.

My 2¢ worth
So there :p

[BC_Programming]

so there's no excuse for the good people at Stanford to not come up with something very decent indeed.


Actually, there is a very good reason. They are pompous douchebags who like to make handwavy generalizations from a position of deep ignorance (for example, say, saying that ATI doesn't have a full OpenCL implementation, when, as you've demonstrated, they do)

[jjFarking]

I just don't get it though.
Aren't we all involved in this, in order to make things better for everyone concerned?

Not that any of this will stop me folding, but it beggars disbelief in regards to actually utilising, to its fullest extent, ALL options available to Stanford in particular, and the general community at large, to have the most efficient computing power for the requirements at hand.
Just think of all the untold teraflops of underutilised (or completely wasted) computing power they can so easily tap in to..

/me is lost
:confused: