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

[JayliN]

Just got the unstable system error when I tried to run 2x gpu clients. Looks like I need to go to radio shack to get some resistors.

With that said, running 2x ati cards will probably kill my cpu's ppd. I'm willing to test and see what happens though.

 

[BUCK NASTY]

A question for the pros:

I'm running 2xvmware notfred images right now which utilize all four cores to 100%. ~ 4200ppd

I also have 2 4670s in this system that are capable of folding at ~1300ppd each. I want to fold with them but also game when I want to. How difficult is it to set up 2 gpu clients in addition to the 2 smp clients I'm running now? What would I have to do to pause the gpu clients when I want to game? I guess I can forget about crossfire too right?

Just wondering if its worth it to fold with my 4670s. any input would help.

It's not hard to set up separate instances for the GPU's. You must create separate folders for F@H GPU2(gpu-0, gpu-1) that assign the client to a chosen GPU. Follow the multi-gpu directions in the sticky section.

 

[thebluebumblebee]

A question for the pros:

I'm running 2xvmware notfred images right now which utilize all four cores to 100%. ~ 4200ppd

I also have 2 4670s in this system that are capable of folding at ~1300ppd each. I want to fold with them but also game when I want to. How difficult is it to set up 2 gpu clients in addition to the 2 smp clients I'm running now? What would I have to do to pause the gpu clients when I want to game? I guess I can forget about crossfire too right?

Just wondering if its worth it to fold with my 4670s. any input would help.

Read the sticky for setting up two Nvidia clients. because it's the same. You just need 2 separate GPU client folders and add -gpu 0 for the first card and -gpu 1 for the second card. picture Whether it's worth it is up to you and I think you have to disable crossfire in the software first (you don't have to remove the bridge). Yes, you should stop the gpu clients before gaming (especially the no crossfire thing). BTW, it is my humble opinion, that if you are going to be popping in and out of Folding, that using the tray client is better as long as you don't use the viewer. The viewer drastically lowers PPD and tends to crash the client, especially for ATI.

 

[JayliN]

Read the sticky for setting up two Nvidia clients. because it's the same. You just need 2 separate GPU client folders and add -gpu 0 for the first card and -gpu 1 for the second card. picture Whether it's worth it is up to you and I think you have to disable crossfire in the software first (you don't have to remove the bridge). Yes, you should stop the gpu clients before gaming (especially the no crossfire thing). BTW, it is my humble opinion, that if you are going to be popping in and out of Folding, that using the tray client is better as long as you don't use the viewer. The viewer drastically lowers PPD and tends to crash the client, especially for ATI.

Thanks for your help. I went to the one radio shack that didn't sell resistors in hopes of making a dummy plug to fold but it looks like what I was missing is the -gpu # parameters. I'll give that a try and as long as the gpus don't kill my cpu ppd, look for me to put out ~7k ppd total in the near future

 

[BUCK NASTY]

Hy there.
I actually had a Problem yesterday. I´m using a gpu client for folding and yesterday it just peaked my cpu(not gpu) usage to 100% a just froze. Before that i already folded two jobs without problems. Could that be a bug? I am using Win 7 x64.

Thanks.

Help him out guy's. He has a ATI4850 folding GPU2. Does anyone know the flag commands(flush interval?) for lower cpu usage with ATI and will it help?

 

[stanhemi]

Brook_Yield= 2
Cal_No_Flush= 1
Flush_Interval= 64 or 128 or 256

try this flag: -local -advmethods -forceasm -verbosity 9 -gpu 0

 

[PaulieG]

Does it take a few days for PPD's to stabilize when I start folding after some time off? I ask because my PPD today seem very low for 3 gtx260's. Also, when overclocking in Rivatuner, should I be overclocking just performance 3D?

 

[BUCK NASTY]

Does it take a few days for PPD's to stabilize when I start folding after some time off? I ask because my PPD today seem very low for 3 gtx260's. Also, when overclocking in Rivatuner, should I be overclocking just performance 3D?

Forget Rivatuner and go with Precision. I have switched over all my rigs and I am very satisfied with the stability and control I have with Precision.

Do you have your priorities set to normal(low) for the GPU's(I know your crunching every CPU)?

 

[PaulieG]

Forget Rivatuner and go with Precision. I have switched over all my rigs and I am very satisfied with the stability and control I have with Precision.

Do you have your priorities set to normal(low) for the GPU's(I know your crunching every CPU)?

I have everything set to default in the console. I never really got too familiar with folding, so I have no idea how to tweak things for the best performance.

 

[JayliN]

Q9550 solo = 4300ppd
Q9550 + 1 HD4670 = 3900

I knew that ati cpu usage would ruin everything

 

[thebluebumblebee]

Does it take a few days for PPD's to stabilize when I start folding after some time off? I ask because my PPD today seem very low for 3 gtx260's.

If you're talking about one of the stats sites, the PPD numbers are based on the average of your work over the last 7 days. Look at your points for the last 24 hours.

 

[BUCK NASTY]

Q9550 solo = 4300ppd
Q9550 + 1 HD4670 = 3900

I knew that ati cpu usage would ruin everything

you running VMware?

 

[JayliN]

you running VMware?

yes. latest notfred vmware images x 2.

 

[hat]

When it says IP address to bind core to for viewer, this means fahmon right? Which IP would I use, my internet address or that pc's network address?

 

[BUCK NASTY]

When it says IP address to bind core to for viewer, this means fahmon right? Which IP would I use, my internet address or that pc's network address?

I don't know anyone that has used that setting. We always leave it @ default.

 

[El Fiendo]

yes. latest notfred vmware images x 2.

Are you using a priority manager? Prifinity, though it has its problems in Win7 x64, is currently the program I use. VM Ware will choke GPU production no matter what, simply because VM Ware defaults to a higher priority than the GPUs. Using Prifinity, set the priority of the GPU consoles to higher than the VM Ware consoles, and set the program to remember this setting. Whenever the program is running, it will automatically keep the process at a higher priority. This is handy as you'd have to reset it every time a work unit completes if you didn't use a priority manager.


Make sure you have FahCore 11 and FahCore 14 both added to Prifinity. 1 for each card. For instance, on my i7, I'll have 3 FahCore11s and 3 FahCore14s in the list, 1 for each card it's setting priority for.

 

[JayliN]

Are you using a priority manager? Prifinity, though it has its problems in Win7 x64, is currently the program I use. VM Ware will choke GPU production no matter what, simply because VM Ware defaults to a higher priority than the GPUs. Using Prifinity, set the priority of the GPU consoles to higher than the VM Ware consoles, and set the program to remember this setting. Whenever the program is running, it will automatically keep the process at a higher priority. This is handy as you'd have to reset it every time a work unit completes if you didn't use a priority manager.


Make sure you have FahCore 11 and FahCore 14 both added to Prifinity. 1 for each card. For instance, on my i7, I'll have 3 FahCore11s and 3 FahCore14s in the list, 1 for each card it's setting priority for.

Thanks for the tips, I'll play around with it more to see how I can maximize my ppd/watt.

I was running both my cpu and gpu clients under low priority as set in task manager. cpu client uses "idle" in advanced client config and gpu client uses "low" in advanced client config. But the hit on my cpu production was massive. Before the ppd breakdown was

2150+2150 = 4300ppd

With the 4670, it was = 1450+1450+1000(gpu) = 3900

With the cooler weather, I just overclocked my q9550 to 3.6g and now I'm pushing 5000ppd (no gpu client).

 

[BUCK NASTY]

Anyone having problems with the 787 pt WU's today? I've got 2 rigs crashing in the same manner and there folding 787's. Kinda weird...

 

[hertz9753]

Anyone having problems with the 787 pt WU's today? I've got 2 rigs crashing in the same manner and there folding 787's. Kinda weird...

I haven't got a single 787 today. Just one 353 after the other.

 

[thebluebumblebee]

my 8800GTS 512 is 75% through one without any problems.

 

[Crazybc]

you asking about the GPUv2 gromacs then I,ve been getting lots of them today and they have all completed Just got another one..

 

[msgclb]

Anyone having problems with the 787 pt WU's today? I've got 2 rigs crashing in the same manner and there folding 787's. Kinda weird...

I know earlier this week I saw 787 pt WUs but I can't remember which rig(s) they were on. Also earlier in the week I had a X58, w7 with a pair of GTX 260 cards repeatedly crash after running for some 6 hours but I wasn't paying any attention to WUs.

Now that same rig has been running more than 120 hours with one of those GTX 260 cards and all I've seen is 353 pt WUs.

I just decided to install the other GTX 260 card in another rig to see what happens.

 

[JayliN]

fixed my problems. Just had to add the environmental variables and make a dummy plug. Now happily folding away at ~7000ppd.

I wish there was a way to fold while crossfired. Turning it on and off to game gets tedious.

 

[sneekypeet]

Buck...my 280 is failing.....

I need some one on one time to see if we can sort this out.

new drivers, new install of the F@H client. Ran all night and faild this morning. really odd.

the usual unstable machine error leading to EUE. Project: 5767 (Run 1, Clone 190, Gen 969) is what failed.

 

[bogmali]

Buck...my 280 is failing.....

I need some one on one time to see if we can sort this out.

new drivers, new install of the F@H client. Ran all night and faild this morning. really odd.

the usual unstable machine error leading to EUE. Project: 5767 (Run 1, Clone 190, Gen 969) is what failed.

I have one these once in a while, do one of the following:

1. Reconfig and change the WU size (you choices are small, normal, big) from what it is currently.
2. If you're OCed, drop the clocks down.

If it persists, post some screenies bro.