• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Preliminary numbers for Nvidia GPU client

^^^Good post. But the variances are frustrating. I have a 512MB GTS, overclocked to around 800/1850/1000, and I'm only seeing 3000ppd max, even though I've seen others reporting 5500 ppd with similar configurations (q6600 @ 3ghz feeding it, with no other CPU clients running). This is with the official cuda drivers (174.55 i think).
I gave the leaked 177 drivers a whirl, and tried the modded inf file, but that just screwed up my boxen. I think a re-install of Winders might be my next move.
 
jebo_4jc said:
^^^Good post. But the variances are frustrating. I have a 512MB GTS, overclocked to around 800/1850/1000, and I'm only seeing 3000ppd max, even though I've seen others reporting 5500 ppd with similar configurations (q6600 @ 3ghz feeding it, with no other CPU clients running). This is with the official cuda drivers (174.55 i think).
I gave the leaked 177 drivers a whirl, and tried the modded inf file, but that just screwed up my boxen. I think a re-install of Winders might be my next move.
Click to expand...

drivers seem to be the key... I noticed on F@H forum that another guy with a GTX was getting EXACTLY my points on his, we used the same driver and had about the same system so it is consistent. It just varies how clean the install is. How much you tax your cpu core that your GPU is running on, etc.



 
Please keep in mind that the WU's we are doing are tiny... they are not using the GPU to its fullest... I expect to see this increase in the future.

Side note I need to OC my GTX again (lost it with the win xp 32bit install) once I do I might break 5k.. I had a pretty solid OC and since it is water cooled... :)



 
kirk121 said:
I just wanted to note that everyone that is talking about performance, these are beta drivers and a beta client for nvidia folding.

Everything takes time to get working the right way, we just need to give the folding guys some time to get the client up to par. After all it is the very first release for us to test out for them and to find and report any problems that we find so that they can get everything working. Then we can compare stats and drivers and PPD's, but for now we should just enjoy that we can finally fold with our speedy nvidia cards. :D

And I can under stand wanting to make sure that you have the best possible hardware to run this on, I cant wait for native 64bit windows support as well as sli support. Then we can all say that today is a good day to fold.:D
Click to expand...


It is always a good day to fold! Don't let a lack of official 64 bit drivers stop you. I got my 64 bit Vista rig folding with the new drivers.

GPU: BFG Geforce 8800GTX Stock
CPU: Q6600 @ 3.2 (also running 2 SMP instances)
CPU load while folding: 100%
Board Connection (PCI-E, AGP, etc.): PCI-E
Memory used by the client: 20MB (FahCore _11)
Temps of the GPU Diode: 75 C GPU 71C
GPU Clocks: core 576/mem 900/shader 1350
PPD (as stated by FAHMON or other program): 4200-4500PPD

GPU: EVGA 8800GT Superclocked Edition
CPU: Q6600 @ 3.2 (also running 2 SMP instances)
CPU load while folding: 100%
Board Connection (PCI-E, AGP, etc.): PCI-E
Memory used by the client:
Temps of the GPU Diode: 71-74 C
GPU Clocks:
PPD (as stated by FAHMON or other program): 4200-4500PPD -- same as a 8800GTX??
This client is on a wireless connection and always seems to be listed as hung in FAHMon (the other clients don't seem to be listed as hung nearly as much)

Seems odd that the 8800GT is the same as an 8800GTX

Fahmon is saying my total production is going from 9-10k ppd to 16-19k ppd with the existing hardware. WOW! :eek:

0.png
 
It's so tempting right now to rack a few GPU cards and double-triple our production :p

/raise fist

DAMN NIVIDIA

/lower fist


 
Is anyone getting the high numbers 4-5kppd on a 8800 class machine using windows vista and the 174 drivers? It seems to me the biggest variance is that. I'm sure hoping thats the issue, I'm one of the 2k ppd people, I have a gtx and a original low memory gts they are in the same range of each other both 2kish.
I hope they release official 177's for 8800's soon.
 
The G92 is getting bigger numbers vs the G80 but the GTX/Ultra should be doing over 4000 ppd.

 
Gray_area said:
Is anyone getting the high numbers 4-5kppd on a 8800 class machine using windows vista and the 174 drivers? It seems to me the biggest variance is that. I'm sure hoping thats the issue, I'm one of the 2k ppd people, I have a gtx and a original low memory gts they are in the same range of each other both 2kish.
I hope they release official 177's for 8800's soon.
Click to expand...

my 8800GTS is pulling ~2k ppd

 
Buahahahahaha, that's crazy !!!!!!

I restored my Vista x64 that I backed on the Windows Home Server, installed the 177.35 driver then fired the GPU2 client. I'm now doing 4980 ppd with a 8800GTS 640MB overclocked to 700/2000/1750 (gpu/mem/shader). That's a record for a old G80 card !!!

/me kiss the A3 core which is the same as a 8800 Ultra :D



 
Xilikon said:
The G92 is getting bigger numbers vs the G80 but the GTX/Ultra should be doing over 4000 ppd.
Click to expand...

There definitely seems to be an advantage to the g92s. It's also got to have something to do with the number of stream processors because the 9600s aren't doing all that great compared to the 8800GT and g92 8800GTS.

I think driver might have quite a bit to do with performance, though. I think the 177 series is giving a pretty hefty boost.

Everyone who has posted results already, please go back to your post and update it with what driver version you're running and on what OS.
 
Holy crap...12 minutes since I downloaded it and its at 21% on wimpy 8800GS......:eek::eek:
 
Logan321 said:
Now I'm curious. FahSpy says I'm running 5001 (5-135-1) work unit, 34secs per %, credit listed as 98, PPD 2490, on a dedicated core.

This is with a Q6600 @ 3.6GHz and a 8800GT OC @ 625/900MHz
Click to expand...

The above was with the 75.xx drivers, upgraded to the 77.35 hacked drivers (what a pain in the a** that was >.< ) and now I'm getting 18secs per %, PPD 4540.

And I'm running a SMP on Vista32 for another 1800PPD, totaling 6200... bit of a step up from the 2x SMP I was running before, only getting 1600PPD per smp.
 
Logan321 said:
The above was with the 75.xx drivers, upgraded to the 77.35 hacked drivers (what a pain in the a** that was >.< ) and now I'm getting 18secs per %, PPD 4540.

And I'm running a SMP on Vista32 for another 1800PPD, totaling 6200... bit of a step up from the 2x SMP I was running before, only getting 1600PPD per smp.
Click to expand...

hacked drivers?!?! :confused:

 
schenksmill said:
the 177.35 drivers are for the new cards, but with an .inf hack they work just fine on the 8 and 9 series, with significant improvements to the CUDA driver in them

Click to expand...

is there a how to? I'm all about messing with stuff...... just need to know what to mess with :p

 
Updated number with 177 drivers

GPU: 8800 GTS 512MB G92
CPU: Q6600 @ 3.4
2GB RAM @866Mhz
CPU load while folding: ~20%
Board Connection: PCI-E
Memory used by the client: 15 megs for the core
WinXP32
Temps of the GPU (NVMonitor): 54C
GPU Clocks: Core/Mem 740/1075
PPD FahSpy 5010 ppd w/175 drivers and 5645ppd w/177 drivers
 
Ok, so now i'm starting to get upset. without anything else running i'm still only getting about 2560 in total ppd out of the 8800GT.

Running vista 32bit and the 174.55 driver, anyone else with this type of config not getting 4k+ in fahmon?
 
GPU: Geforce 8800 Ultra
CPU: C2D E6750 @ Stock 2.66
4G Ram
CPU Load: 61%
Board Connection: PCI-E
Memory Used: Roughly 21megs
Vista x64
PPD 5216
 
Imitation said:
Ok, so now i'm starting to get upset. without anything else running i'm still only getting about 2560 in total ppd out of the 8800GT.

Running vista 32bit and the 174.55 driver, anyone else with this type of config not getting 4k+ in fahmon?
Click to expand...

Got about 2900 ppd with Vista 32 bits but over 4900 ppd with Vista x64... It's weird and I can't explain why :(

 
My GPU seems to be pulling 4980.71 PPD by itself.. Just to give some insight. (177.35) Vista x64
 
Xilikon said:
Got about 2900 ppd with Vista 32 bits but over 4900 ppd with Vista x64... It's weird and I can't explain why :(

Click to expand...

So what your saying is I need to get vista 64 if we want to get the points we deserve ?

 
Same driver yes...

About the OS, I dunno since others have different results while some have the same result as mine. Do the experiment yourself and see if it works.

 
I went from 2400 to 4600 ppd just from a driver update... 75.xx to 77.35 That's with Vista 32 on a 8800GT OC
 
That's what i'm going to try tonight because right now something isn't right at all. My SMP client that's running along side the gpu client is taking an hour a frame instead of about 15mins. :(
 
Imitation said:
That's what i'm going to try tonight because right now something isn't right at all. My SMP client that's running along side the gpu client is taking an hour a frame instead of about 15mins. :(
Click to expand...

That was happening to me too..... Ok maybe a wont format reinstall..... cuz that would take days to get my system back to where it is :p

 
Yeah, my smp client took a hard nose dive too. Kept reading about how gpu wasn't affecting smp, and started worryin somethin was broke.
I'm really really hoping that its the stupid drivers, and they release the 177's for everyone soone.

they obviously work as is, are they just holding them back so only the new cards get good performance til all the reviewers are done comparilating them with the 8800's?
 
Hmm, my SMP remained nearly the same but I'm running 177.35's. It may have went up one minute per frame, but nothing drastic. I don't mess with VMs or multiple clients with affinity changer on this PC either, just one GPU and one Windows SMP.

Pre GPU:

Code: 
[13:31:37] Working on Unit 00 [June 10 13:31:37]
[13:31:37] + Working ...
[13:31:37] 
[13:31:37] *------------------------------*
[13:31:37] Folding@Home Gromacs SMP Core
[13:31:37] Version 1.74 (March 10, 2007)
[13:31:37] 
[13:31:37] Preparing to commence simulation
[13:31:37] - Looking at optimizations...
[13:31:37] - Created dyn
[13:31:37] - Files status OK
[13:31:48] - Expanded 4763909 -> 24426905 (decompressed 512.7 percent)
[13:31:48] - Starting from initial work packet
[13:31:48] 
[13:31:48] Project: 2665 (Run 3, Clone 397, Gen 9)
[13:31:48] 
[13:31:49] Assembly optimizations on if available.
[13:31:49] Entering M.D.
[13:31:49] + Results successfully sent
[13:31:49] Thank you for your contribution to Folding@Home.
[13:31:49] + Number of Units Completed: 109

[13:31:55] Rejecting checkpoint
[13:31:57] Protein: HGG in water
[13:31:57] Writing local files
[13:32:02] Extra SSE boost OK.
[13:32:03] Writing local files
[13:32:03] Completed 0 out of 250000 steps  (0 percent)
[13:45:45] Writing local files
[13:45:45] Completed 2500 out of 250000 steps  (1 percent)
[13:59:29] Writing local files
[13:59:29] Completed 5000 out of 250000 steps  (2 percent)
[14:13:08] Writing local files
[14:13:08] Completed 7500 out of 250000 steps  (3 percent)
[14:26:37] Writing local files
[14:26:37] Completed 10000 out of 250000 steps  (4 percent)
[14:40:05] Writing local files
[14:40:05] Completed 12500 out of 250000 steps  (5 percent)
[14:53:32] Writing local files
[14:53:32] Completed 15000 out of 250000 steps  (6 percent)
[15:07:02] Writing local files
[15:07:02] Completed 17500 out of 250000 steps  (7 percent)
[15:20:31] Writing local files
[15:20:31] Completed 20000 out of 250000 steps  (8 percent)
[15:34:02] Writing local files
[15:34:02] Completed 22500 out of 250000 steps  (9 percent)
[15:47:32] Writing local files
[15:47:32] Completed 25000 out of 250000 steps  (10 percent)

Post GPU:

Code: 
[09:08:18] Working on Unit 09 [June 19 09:08:18]
[09:08:18] + Working ...
[09:08:18] 
[09:08:18] *------------------------------*
[09:08:18] Folding@Home Gromacs SMP Core
[09:08:18] Version 1.74 (March 10, 2007)
[09:08:18] 
[09:08:18] Preparing to commence simulation
[09:08:18] - Looking at optimizations...
[09:08:18] .
[09:08:22] - Starting from initial work packet
[09:08:22] 
[09:08:22] Project: 2665 (Run 3, Clone 121, Gen 12)
[09:08:22] 
[09:08:23] Assembly optimizations on if available.
[09:08:23] Entering M.D.
[09:08:44] al work packet
[09:08:44] 
[09:08:44] Project: 2665 (Run 3, Clone 121, Gen 12)
[09:08:44] 
[09:08:44] Entering M.D.
[09:08:50] Rejecting checkpoint
[09:08:52] Protein: HGG in water
[09:08:52] Writing local files
[09:08:58] Extra SSE boost OK.
[09:08:59] Writing local files
[09:08:59] Completed 0 out of 250000 steps  (0 percent)
[09:22:31] Writing local files
[09:22:31] Completed 2500 out of 250000 steps  (1 percent)
[09:36:05] Writing local files
[09:36:05] Completed 5000 out of 250000 steps  (2 percent)
[09:49:37] Writing local files
[09:49:37] Completed 7500 out of 250000 steps  (3 percent)
[10:03:10] Writing local files
[10:03:11] Completed 10000 out of 250000 steps  (4 percent)
[10:16:43] Writing local files
[10:16:43] Completed 12500 out of 250000 steps  (5 percent)
[10:30:16] Writing local files
[10:30:16] Completed 15000 out of 250000 steps  (6 percent)
[10:43:42] Writing local files
[10:43:43] Completed 17500 out of 250000 steps  (7 percent)
[10:57:00] Writing local files
[10:57:00] Completed 20000 out of 250000 steps  (8 percent)
[11:10:16] Writing local files
[11:10:16] Completed 22500 out of 250000 steps  (9 percent)
[11:23:33] Writing local files
[11:23:33] Completed 25000 out of 250000 steps  (10 percent)

 
Ok everyone. I just got done reading a bunch of the entries in the PPD GPU2 thread on the Folding forum. It's pretty obvious that you want to be running the 177.35 drivers. PPD takes a pretty big hit if you're running anything less.

I just copied the nv4_disp.inf from the 175.16 drivers over the one for the 177.35 drivers and it installed just fine (XP 32-bit).

If that doesn't work, give these a shot.

http://forums.nvidia.com/index.php?showtopic=70056
 
Hows this.... .does the performance number look good for an 8800GTX?

GPU-2.jpg
 
Close the viewer, it kills performance and use fahmon to check the ppd estimate.

 
It took some tinkering but I was able to disable SLI and move my monitor to my 2nd 8800GT and have my 1st one continuously folding without sacrificing general video performance (playing videos on VLC player). I can leave the client running all the time and have almost no noticeable difference while running windows. :D

8800GT @ 640 core / 950 mem (174.55 drivers ATM)
C2D E6550 @ 2.45 GHz
4,700 PPD / 50% CPU usage / 30 mins per WU
 
GPU: EVGA GeForce 9600GT
CPU: E8400 @ stock
CPU load while folding: 50%
Board Connection: PCI-E
Memory used by the client: 15MB
Temps of the GPU: 47C @ 75% fan...temps seem low?
GPU Clocks: 675 core / 1875 shader / 900mem
PPD (FAHMON): w/ 175.16 = 3200ppd
w/ 177.35 = 4100ppd


 
ah this is torture... i want to sell my 8800GTX's cause im broke... but i could pop it in two machines at work and let them fold all day for massive ppd... what to do lol :(
 
FLECOM said:
ah this is torture... i want to sell my 8800GTX's cause im broke... but i could pop it in two machines at work and let them fold all day for massive ppd... what to do lol
Click to expand...
How much would you sell the cards for, hypothetically speaking?

 
You must log in or register to reply here.
Back
Top