Why have P10102 - 10106 such (relatively) low credits ?

Moderators: slegrand, Site Moderators, PandeGroup

Re: Why have P10102 - 10106 such (relatively) low credits ?

Postby Grandpa_01 » Tue Mar 09, 2010 4:10 pm

domboy wrote:
stevehat1 wrote:
COOLDUDEGAMER wrote:At least these WUs run quicker and cooler than the old 511-point series WUs on my GTS250.

I always thought these 1010x WU's were the re-incarnation of the 511's LOL


And I thought they were a re-incarnation of the 548 point WU I used to see on the ATI GPU client... :?:
Those were the best (highest PPD) I've seen on my 4850. I wish Folding@Home had assigned these new projects to the ATI folders, especially since a lot of nVidia folders don't like them anyway. Though my GT 220 doesn't complain when folding them...


Now isn't that the trouth, I would be willing to bet there are allot of ATI folders that would just love to have a GPU WU that caused our cards to fold at 70C+ since most of the WU's that ATI fold you have to turn the fan up to 75% to fold at 80C+ and get half the points Nvidia cards do. Suck it up and fold on. :mrgreen: :lol:
Image
2 - SM H8QGi-F AMD 6xxx=112 cores @ 3.2 & 3.9Ghz
5 - SM X9QRI-f+ Intel 4650 = 320 cores @ 3.15Ghz
2 - I7 980X 4.4Ghz 2-GTX680
1 - 2700k 4.4Ghz GTX680
Total = 464 cores folding
User avatar
Grandpa_01
 
Posts: 1757
Joined: Wed Mar 04, 2009 7:36 am

Re: Why have P10102 - 10106 such (relatively) low credits ?

Postby domboy » Wed Mar 10, 2010 2:01 pm

Grandpa_01 wrote:
Now isn't that the trouth, I would be willing to bet there are allot of ATI folders that would just love to have a GPU WU that caused our cards to fold at 70C+ since most of the WU's that ATI fold you have to turn the fan up to 75% to fold at 80C+ and get half the points Nvidia cards do. Suck it up and fold on. :mrgreen: :lol:


Right. I have to smile when reading threads like this one. My nVidia cards only get 1 Billion ppd with this new work unit instead of 1.1 Billion, and it runs hotter! :roll: If you want something to complain about try running an ATI GPU client. But even the ATI client is a heck of a lot faster than the Uniprocessor client, which is why I run it.
I've heard rumors that the GPU3 core may level the playing field a bit, i.e. ATI might gain a bit of speed and nVidia might loose a bit. I can only imagine the outcry should that be true.
domboy
 
Posts: 137
Joined: Thu Oct 02, 2008 1:42 pm
Location: Wilmington NC

Re: Why have P10102 - 10106 such (relatively) low credits ?

Postby Sidicas » Thu Mar 11, 2010 5:54 pm

I doubt the GPU3 core would level the playing field, I think that the nvidia cards are just designed in such a way that they're more efficient at getting the kind of work folding@home needs, done.. Of course, such designs have huge drawbacks as well, it's been what? 6 months? 8 months? and it will still be several more weeks before you can even buy an nvidia card that supports DX11.. Also, previous nvidia designs have proven to not be scalable. The GT200/GT200b chip designs simply couldn't, by design, accommodate the mainstream / low end market.. So nvidia just kept respinning their previous chips (g92, etc.)

According to Hardware canucks, the new nvidia Fermi chip has a VERY modular design similar to how AMD can disable a core and sell off a chip as a triple-core if there's a manuf. defect. This wlil allow Fermi to scale from top to bottom easily so they will have full market coverage on DX11 parts this year.. Thank god, huh? I doubt I would have been able to tolerate yet another respin of a chip two generations back.. Even better news, is that since Fermi is more modular, it will be priced competitively. Unlike the GT200/GT200b which, even with the smallest manuf. defect, ends up in the trash bin. This has kept nvidia's prices on their mainstream and high end cards quite high for quite some time now..

For gaming, I wouldn't even consider an nvidia chip right now.. But I have hope Fermi will save the year for nvidia..
Edit: Also, by my guesstimations, since the design of the individual shaders (now called "CUDA" cores) hasn't changed, based off of CUDA Core count and clocks the GTX 470 and GTX 480 should hit 10k PPD and 11k PPD at stock clocks.. What worries me is that a lot of these tiny atom count nvidia workunits simply aren't parallel enough for Gigathread and other Fermi architecture specific features to work.. In other words, I just can't see any way at all that the high end Fermi cards are going to come anywhere near beating the current generation of cards in PPD for your dollar with the current lineup of workunits as they are.. Which really makes me want to wait for the mid range Fermis when it comes to Folding..
Sidicas
 
Posts: 232
Joined: Sun Feb 17, 2008 4:46 pm

Previous

Return to NVIDIA specific issues

Who is online

Users browsing this forum: No registered users and 3 guests

cron