*** F@H Tech Assistance Thread ***

[theonedub]

Is it possible on the 191.xx drivers to have SLI enabled and still fold on both GPUs? I'd like to be able to go from F@H to a game without having to go to the Control Panel and adjust all the settings. Would like to use a SLI bridge too.

Im going to try it out later on after the SuperBowl, but if I fail (and I may) I would like some tips!

Thanks!

 

[hertz9753]

Is it possible on the 191.xx drivers to have SLI enabled and still fold on both GPUs? I'd like to be able to go from F@H to a game without having to go to the Control Panel and adjust all the settings. Would like to use a SLI bridge too.

Im going to try it out later on after the SuperBowl, but if I fail (and I may) I would like some tips!

Thanks!

I can only fold on one with SLI enabled.

 

[theonedub]

I guess I will see how Dirt2 runs with only one 275 at 2048x1152

 

[NastyHabits]

Is it possible on the 191.xx drivers to have SLI enabled and still fold on both GPUs? I'd like to be able to go from F@H to a game without having to go to the Control Panel and adjust all the settings. Would like to use a SLI bridge too.

Im going to try it out later on after the SuperBowl, but if I fail (and I may) I would like some tips!

Thanks!

I'm using the 195.62 drivers, have SLI enabled, fold on both cards (2 X GTS 250). I was also able to do the same on the previous driver version, which was 191.07.

 

[theonedub]

Well I can get both cards to Fold in SLI but only at 50% PPD. I got my fill of Dirt2 in (very fun) and turned SLI off. I learned that 2 GTX 275 are quite the capable gaming setup, not going to stop Folding though In the rare occasion when I play on the PC I will use one GPU and Fold on the other.

Looks like I was misinterpreting the information, they both Fold at 100% in SLI 191 drivers. My mistake

 

[Zenith]

What PPD can I expect from 260GTX?

 

[mstenholm]

8000 to 9300 depending on WU and OC.

 

[kyle2020]

my 260 returns anywhere up to 9k ppd.

 

[theonedub]

Anyone having issues with their GPU clients sending completed WUs? Mine is unable to contact host and my PPD is down even more so than I was expecting with my GSO gone.

 

[thebluebumblebee]

Anyone having issues with their GPU clients sending completed WUs? Mine is unable to contact host and my PPD is down even more so than I was expecting with my GSO gone.

This seems to be going around like a cold. I checked my clients and found that one had a problem earlier today but it was only about a 5 minute delay.

 

[theonedub]

they both just went though, I guess they are having random issues.

 

[theonedub]

My GPU clients are hanging either at 'sending results' or loaded 'queue successfully' deleted the work units, restarted the computer, and its still doing it. If I restart the client a couple times its fine. Getting to be a hassle since I am not at this computer often and I liked the fact it was a set and forget machine.

 

[msgclb]

I've had a devil of a time getting my 2nd -bigadv rig running. I settled on 3.0GHz as my max OC because the system (vr) gets to hot if I go any higher. I next had this rig crash twice but the second time I realized the culprit was the video card as the monitor turned a nasty pink. I swapped out the video card and have successfully completed a LinuxFAH v1.1 A3 core and it's running the 2nd one now. One of my problems is I haven't been able to use HFM to monitor the client on the DFI rig. Here are two images I'll discuss below.

You'll notice that the tooltip for the P6T VM SMP client says the 'same user has the same machine ID'. Its ID is 1. Notice that the DFI VM SMP client's color is the same as P6T and the tooltip for this client is also 'same user has the same machine ID'. Its ID is 2. The ID originally was 1 but after noticing this problem I changed it. Both of these clients are running in their VM clients successfully. Do I have to restart the VM or Linux to fix this?

The other problem is my network only show the FAH for the P6T VM client. I tried renaming the folder for the DFI VM Linux to FAH1 but that didn't work. How do I get a new network name for my DFI rig?

I hope the above is understandable!

 

[mstenholm]

My GPU clients are hanging either at 'sending results' or loaded 'queue successfully' deleted the work units, restarted the computer, and its still doing it. If I restart the client a couple times its fine. Getting to be a hassle since I am not at this computer often and I liked the fact it was a set and forget machine.

That does not work for me. The ATI server still work. It seems like the ATI people for once is out-folding us.

 

[bogmali]

I think the machine ID's only apply in the same rig, i.e. you can't have a duplicate ID for a gpu and a cpu folder in the same machine. If they're on different rigs it would not matter. As far as SMP -bigadv clients, doesn't matter either because not all your vmware clients will have the same IP addresses.

To change your PC name which will be reflected on your network, right-click on My Computer, click Properties, under computer name, domain and workgroup settings click change settings which will give you this option:

 

[msgclb]

As far as SMP -bigadv clients, doesn't matter either because not all your vmware clients will have the same IP addresses.

http://img.techpowerup.org/100215/Untitled.jpg

The cure was to use the ip address, i.e.: \\ip address\fah

I'm running my DFI X58 rig one more time using: ./fah6 -smp 8 before I add -bigadv in about 6 hours.

Doesn't this look pathetic. At least the SMP clients are still running.

Does your User Name in HFM.NET have the same error message: 'Client's User Name does not match the configured User Name'? I know I'm getting credit for my WUs so this probably has to do with the way he combined the User Name and Team Number.

 

[mstenholm]

The cure was to use the ip address, i.e.: \\ip address\fah

I'm running my DFI X58 rig one more time using: ./fah6 -smp 8 before I add -bigadv in about 6 hours.

Doesn't this look pathetic. At least the SMP clients are still running.
http://img.techpowerup.org/100215/FH-2010-02-15-HFM-ErrorGone-3.jpg

Does your User Name in HFM.NET have the same error message: 'Client's User Name does not match the configured User Name'? I know I'm getting credit for my WUs so this probably has to do with the way he combined the User Name and Team Number.

It seems like you haven't changed the original settings:

Edit, Preferences, Web settings.

 

[BUCK NASTY]

Vijay Pande posted that the problem should be solved now. Can someone swing by my house and crank up 17-GPU2 clients for me?

Re: GPU server status 171.67.108.21, 171.64.65.71,171.67.108.26

Postby VijayPande » Mon Feb 15, 2010 3:27 pm
Thanks for the posts. It's early AM in Califorina (that's why this went unfixed for several hours), but I think we've got everything going again. I've contacted Joe regarding this issue: there was a WS bug.

I've also balanced the weights so the other NV WS's can get into the mix better and improve the redundancy.
Prof. Vijay Pande, PhD
Departments of Chemistry, Structural Biology, and Computer Science
Chair, Biophysics
Director, Folding@home Distributed Computing Project
Stanford University

User avatar
VijayPande
Pande Group Member

Posts: 1721
Joined: Fri Nov 30, 2007 6:25 am
Location: Stanford

 

[NastyHabits]

That does not work for me. The ATI server still work. It seems like the ATI people for once is out-folding us.

Over the last two days I've only had my ATI client register results. None of the Nvidia clients have been able to send results.

 

[theonedub]

Can anyone send me a dummy plug? I know they are not absolutely needed, but I think they are more convenient.

 

[bogmali]

Can anyone send me a dummy plug?

Can't you just make one? Or you don't have the DVI-VGA adapter?

 

[theonedub]

I guess I could go get the resistors and attempt it, but if someone already had one they would be saving me the cash and a trip to The Shack.

On another note, I just got a 8800GTS installed into my WCG0 rig in specs. I followed the same steps I did to get my other 275 folding together, but the card refuses to fold. Comes back with unstable machine seconds after displaying "entering MD" The GPU flag is correct and I have it hooked to my second input on my monitor. I tried the force G92 and g80 flags with no change. Any ideas?

EDIT: If I enable SLI I can get all three cards working, but eventually it will hang for no reason. I think that I will keep the last slot in the i7 rig for another 275 if I stumble on another good deal. I will put the 8800GTS in my Phenom setup.

 

[Fitseries3]

i seem to have forgot how to get SMP running on 8threads.

can anyone help?

 

[bogmali]

i seem to have forgot how to get SMP running on 8threads.

can anyone help?

Guide that I follow:

http://www.overclock.net/overclock-net-folding-home-team/607465-guide-vmware-3-0-8-core.html

Which is based off of this:

http://www.evga.com/forums/tm.aspx?&m=4464&mpage=1

 

[Fitseries3]

thanks but i have no room for VM on this drive.

can i get it to work without vmware?

guess i'll switch cpu to WCG