• 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.

SMP + GPU very inconsistent

N

natermeister

Limp Gawd
Joined
Dec 26, 2005
Messages
463
I'm folding again on my Q6600 (at 3.4GHz because it's summer time) and my 8800GT. The 88GT is folding as you would expect, somewhere between 3.5k and 4k ppd depending on the WU. But the SMP client is all over the place. When it's working well, I get somewhere around 2.7k to 3k ppd from it. Right now it's bouncing around just under a thousand ppd.

I keep getting 'timered checkpoint triggered' on it as well. I don't know what this means. Could it be some problem relating to affinity?
 
That is what I'm getting out of my Q6600 at stock speeds running two SMP clients.
I would definately run something like WinAFC if you arent already.
 
How do you have the SMP client set up? Are you using one instance of the Windows SMP client or one or two Linux VMs? Also, please post which workunits you've been running, as well as the contents of your FAHlog.txt file.
 
natermeister said:
I keep getting 'timered checkpoint triggered' on it as well. I don't know what this means. Could it be some problem relating to affinity?
Click to expand...

This simply means that it's saving the data that's been processed up to that point, just in case of a crash or power loss. When you restart the client, it will pick back up at the last checkpoint triggered. ;)

Keep Folding!!!
 
I'm running a single WinSMP client. This is not a dedicated folding machine so I'd rather not mess with setting up virtual machines. The SMP WUs that I'm usually getting are 2653. I usually get 5780s on the GPU side of things.
 
Okay, 2.7k to 3k PPD sounds pretty good for the WinSMP client with 2653 units. Can you post the logfile output during a time period when you're seeing very low PPD? Please post an excerpt from the log while your CPU is performing normally as well.
 
I just trolled through my BIOS and saw that I'd only turned off the C1E halt state and forgot to nix EIST. My production seems to be back to normal now (~2500 on SMP). The strange thing is that Real Temp never showed my clock speeds fluctuating while I was folding.
 
That's odd, EIST shouldn't have any effect on folding performance. I have it enabled on two C2D-based systems and I haven't had any PPD drops. Good to know that you've got the issue dealt with though.
 
its possible that the cpu was down clocking just for a split second when it was doing the check point saving.. since the client does stop when it saves to disc.. but its only for a split second.. so its possible everytime it was saving the cpu went back to idle then sped back up causing the PPD to look like it was fluxuating but it really wasnt..
 
sirmonkey1985 said:
its possible that the cpu was down clocking just for a split second when it was doing the check point saving.. since the client does stop when it saves to disc.. but its only for a split second.. so its possible everytime it was saving the cpu went back to idle then sped back up causing the PPD to look like it was fluxuating but it really wasnt..
Click to expand...
PPD is calculated based on the time between frames, so the CPU clocking down for a moment during a checkpoint save wouldn't cause the PPD to appear low.
 
You must log in or register to reply here.
Back
Top