BRIGHT SIDE OF NEWS About | Advertise | Contact BSN USER Login
| Register 		SUBSCRIBE Newsletter | RSS Feeds
Thursday, May 23, 2013
Email this to a friend.
Your friend's e-mail:
Your Name:
Your e-mail:
Message subject:
Comments
by: Anonymous on 10/14/2009
Just think of this...

All the time these "programmers" are messing around taking money from nVidia for using CUDA, they are killing cancer suffers because the software is not performing at full speed on ATi hardware, which is a massive percentage of users out there.

Think of how many work units have not been finished while these people take back-handers from nVidia!!!

Get it fixed, your supposed to be genius mathematicians and programmers!!!
Brook+ applications can also scale by: Gipsel on 10/8/2009
I have no idea what Folding is doing in its code. But you may want to look at two other distributed computing projects using ATI cards: Milkyway@home and Collatz Conjecture.

Both scale virtually perfectly with the higher shader count. That a HD5870 achieves this with the Collatz Conjecture project is also showing that ATI has done a tremendeous job with the memory controller efficiency as it is quite bandwidth hungry.

And I guess I don't have to tell you how pathetic a GTX285 looks compared to a HD5870 at the Milkyway project which uses double precision calculations (it barely matches the performance of an old and rusty HD3850). Nvidia badly needs a factor 8 improvement here to even approach a Cypress, let alone to pass it. And by the way, an nvidia engineer helped the project with its CUDA application, I guess one can assume it's near the maximum performance the hardware is capable of.
I have high hopes for Fermi, but frankly I still doubt a bit it will pass Cypress on pure number crunching.

Oh, by the way, both projects use the Brook+ layer as a base for the ATI applications. So that has nothing to do with the possible scaling.
by: Anonymous on 10/7/2009
Gee Theo, this is quite a news story. Some really fine detective work on an issue that we're all extremely concerned about.
Wow by: Anonymous on 10/7/2009
You used -forcegpu ati_r700 and was expecting it to be faster? A new client needs to be used as well as larger work units for it to become faster you twit.
by: Anonymous on 10/7/2009
"-forcegpu ati_r700"

Probably also makes the application assume 800 shaders are available at most. Just wait for the updated version. But yes, it's time for F@H to be ported to OpenCL.
How does it handle 3D and Photoshop CS4 processing by: Michael A. McKenney on 10/7/2009
Can you do any 3D CAD rendering and Photoshop CS4 processing benchmarks?
I think you must wait for next generation driver(s). by: Anonymous on 10/7/2009
ATI Radeon HD5870 will work properly when Windows 7 launched officially and by then the driver will mature with OpenCL optimization that could be implemented for folding@home.
Can't win 'em all by: Anonymous on 10/7/2009
Yessiree, ATI does suck in these computational tasks... But hey, they do all right in games.

I suppose their future does not depend on compute tasks because down the track they will be integrated in Fusion - AMD can afford to be a bit lax here. Still, its the wrong attitude to have
You do realize... by: Anshel Sag on 10/7/2009
That Folding at home was designed to make gaming cards into computing devices? At no point was F@H designed for workstation cards in mind... it was specifically designed to target the masses to spread the computing. Hence, the term Distributed Computing.
by: Anonymous on 10/6/2009
yeah always remember blame the GAMING video card for not being able to do WORK STATION jobs correctly because the application technology is not supported by your hardware.
Retarded article is like trying to bake a cake in the shower.
Nothing special : just working as R700 ( 800 shaders, R600 shader design ) by: Anonymous on 10/6/2009
Current Folding core is designed from the old HD 2000/HD 3000 generation ( R600 ), and when R700 came, the developer just added the support for the new hardware with the increased shaders ( 800 shaders vs. 320 in R600 )... but didn't add any R700 specific optimizations, so from arch. POV nothing changed to benefit from arch. improvement in R700 over R600.. this is why in HD4000 you will only gain performance from increased number of shaders and clock speed, nothing more...

the trick you made doesn't change anything at all, the core will treat R800 as R700, so only 800 shaders will be used ( from 1600 ) and coz the core it self doesn't have any R700 optimization, so the core will treat R800 same as R700 ( or just 800 shaders ), or just R600 with increased shaders and clocks !!

if you do the -forcegpu trick you will be able to fold on R800 normally, but with half of it's power as the core it self isn't able to see the whole 1600 shaders.. not to mention that it won't benefit from and shader performance improvement as the core still treat all shaders as R600 shaders...

the developer says that it mean nothing to develope the core to support any arch. change, but it may support the increased number of shaders... but about the optimization for the new shaders in both R700 and R800, he will do nothing...

the reason is simple, the current core is outdated and based on old stream platform ( Brook+, CAL, etc.. ) which has a lot of limitations that causes the current bad performance of all Radeon card in folding. he sees that consetrating on the next OpenCL based core is much more important... but the bad news is he just started the work in less than two weeks...

the OpenCL won't see the light till 2010, but no time frame to talk about... it's just too early to guess or estimate

Al-Khalaf, arabhardware.net
© 2009 - 2011 Bright Side Of News*, All rights reserved.
© 2009 - 2013 Bright Side Of News*, All rights reserved.