new grapics card just came

[dz_jad]

got it off the classies, $250. X1900XT. Gonna reboot now and set 'er up. :D
I wasnt expecting to get it until next week...so I'll have some folding results hopefully in the next couple days :D

[dz_jad]

what's the points value on these suckers? This is what I've got going on so far...
[23:38:27] Project: 2734 (Run 1, Clone 343, Gen 0)
[23:38:27]
[23:38:27] Assembly optimizations on if available.
[23:38:27] Entering M.D.
[23:38:33] Will resume from checkpoint file
[23:38:42] Working on 1258 GPUlambda
[23:38:42] Starting GUI Server
[23:38:42] Resuming from checkpoint
[23:38:42] Verified work/wudata_01.log
[23:38:42] Verified work/wudata_01.edr
[23:38:42] Verified work/wudata_01.trr
[23:38:42] Verified work/wudata_01.xtc
[23:48:00] Completed 1
[23:55:31] Completed 2
[00:02:47] Completed 3
[00:09:43] Completed 4
[00:16:47] Completed 5
[00:23:42] Completed 6

how does she fair? I cant seem to find a good thread on the GPU folding :(



ok, just answered my own question. found the project on the summary page, 330 points, and its hanging about 688 PPD if it keeps with the just under 7 min per frame...nice.
**untouched clocks 621/720
**temps hanging around 97 C....not too bad, ehh?

[muddocktor]

Damn, that is some high temps! :eek:

[LandShark]

you sure want to crank up the fan!! I feel comfortable at high 70s to low 80s even tho a few people say that's hot already!

[dz_jad]

I cant seem to find the option :shrug:

[LandShark]

do download ATI Tray Tool (my personal preference) or ATITool (most other use), you can adjust the stock fan speed there according by temp plus you can use it to keep it at 3D clock and o/c. your card should have no problem to do 650+core thus equal to mid-high 700s ppd. :cool:

also, from my experience, F@H seems like core speed over memory speed!

[dz_jad]

thanks. I just did some poking in ATITool again and found the fan speed thingy...so down to 84C with fan at 100%.....that still seems high. Not sure why its so hot, I'm in my basement even :shrug:
hmm...just dropped to 76 for a split second, now hovering around 80...

[pscout]

thanks. I just did some poking in ATITool again and found the fan speed thingy...so down to 84C with fan at 100%.....that still seems high. Not sure why its so hot, I'm in my basement even :shrug:
hmm...just dropped to 76 for a split second, now hovering around 80...

Sounds like you might need some more cooling assuming you are running in a case?

If you can keep the temps down, you should be able to raise the volts on the core (i don't play with memory V on any of my flavours of x1900's).
The max i run any of my cores at is 1.375 V.

From what i have read, your xt is likely to oc as well as the 1900xtx's do.

Not sure if the unadjustable memory latencies are the some tho. My sapphire xt 256 lite will only run 700/700 and is only slightly faster than my aiw's at 675/612 :shrug:

my xtx runs at 700/800 and has a tpf just around 5:45 so around 800 ppd.

my aiw 1900's run at 675/612 and get a tpf of 8:10-8:20 for just under 600 ppd.

For case cooling you could remove a side cover from the case or add a fan blowing into it. The stock gpu fans are pretty noisy at 100%.

Both sides of the card get quite toasty so you need good air to the cpu area for chipset/mofset cooling to keep the cpu stable at OC.

[muddocktor]

It also probably wouldn't hurt to remove the heatsink off the vid card and clean that factory tim off and apply some AS5 or Ceramique too. I did that with my old X800XT-PE card and lowered it's temps 5-10 C after I got it.

[WarriorII]

Wow, that's some nice ppd there.

you're gonna be giving Pete a run for his $$ now.

:attn:

[Sleepy_Steve]

How many pixel pipelines does that have?? -- And when GPU folding, whats more important, memory or pipelines?

Crazzy temps though... Im glad i'll be able to cool mine with water.

[dz_jad]

thanks for the info guys
the side of my case IS off :P the fan blows out back anyway, I'll just have to get creative, and like muddocktor was saying, I've thought about ripping it off and getting some good goop on there.
doing good, got the core bumped up to 688, and mem still @ 720, and I'm running at 780 PPD, thats just a 'throw 'er up there' overclock.
anyway, thanks guys!

[PaciFIST69]

Make sure you test your OC by using "3d view" in atitool. You can turn your screensaver off and let it run for like an hour or so, then you know you're good. You can probably bump your memory too...

[pscout]

3dmark2005 (or 2006 which takes longer to run) is also good ... in fact i use it as my final stability test before letting them fold.

The benchmark scores aren't useful to predict folding perf tho since it weights the score with CPU perf etc ... and gpu perf doesn't seem to depend very much at all on cpu perf.

[WarriorII]

NM.... reread 1st post.

[dz_jad]

NM.... reread 1st post.
lol.


yeah, I've just got this thing set up, I've by no means found the limit of its clock, I just wanted to get a few more PPD right now, so I threw up the core frequency, and got another 100 ppd :)

[Lancelot]

Would such a thing ever be written for the nVidia GPU's? My C2D Merom laptop is my folder, and the 20-pipe 256MB 7900GS inthere would be great for folding since it comes with the (heavier) dual heatpipe cooling needed for the 7900GTX...

[dz_jad]

Would such a thing ever be written for the nVidia GPU's? My C2D Merom laptop is my folder, and the 20-pipe 256MB 7900GS inthere would be great for folding since it comes with the (heavier) dual heatpipe cooling needed for the 7900GTX...
stanford is working on expanding to other cards...since the architectures are different of each card, they have to re-port all of the code and everything, evidently it takes a lot of time and effort. You just might see it someday :shrug:

[PaciFIST69]

stanford is working on expanding to other cards...since the architectures are different of each card, they have to re-port all of the code and everything, evidently it takes a lot of time and effort. You just might see it someday :shrug:

It's probably not going to happen for his card.

The R580 (in the X1900XT, etc.) performs particularly well for molecular dynamics, due to its 48 pixel shaders. Currently, other cards (such as those from nVidia and other ATI cards) do not perform well enough for our calculations as they have fewer pixel shaders. Also, nVidia cards in general have some technical limitations beyond the number of pixel shaders which makes them perform poorly in our calculations.

Try SMP though, that will get you even greater ppd than gpu.

[LandShark]

lol.


yeah, I've just got this thing set up, I've by no means found the limit of its clock, I just wanted to get a few more PPD right now, so I threw up the core frequency, and got another 100 ppd :)
glad it's working out great!! you know, for a sec, I was afraid it might be a sublime card when recommended it for you.... http://landshark-racing.com/icon/thumbsup.gif