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

[headshot119]

Still having stats problems with my 2 X 5770s, they both complete work units at the same time and both upload the units within minutes of each other, but only 1 unit ever shows up on the stats pages.

Any ideas?

 

[xvi]

Still having stats problems with my 2 X 5770s, they both complete work units at the same time and both upload the units within minutes of each other, but only 1 unit ever shows up on the stats pages.

Any ideas?

Could you post your two client.cfg and FAHlog.txt files? Are you getting near your expected PPD?

 

[thebluebumblebee]

Still having stats problems with my 2 X 5770s, they both complete work units at the same time and both upload the units within minutes of each other, but only 1 unit ever shows up on the stats pages.

Any ideas?

Look at the log.

1.)Are the numbers the same? Such as: Project: 5798 (Run 17, Clone 292, Gen 9)
2.)If you use HFM.net, it will tell you if there are conflicts between the clients.
3.)Check your setup for the clients and make sure that they have different machine ID's. Normally, this would mean that GPU0 would be #2 and GPU1 would be #3.
4.) Do you have 2 shortcuts with 1 being for GPU0 and the second for GPU1?
5,)Since you have 5xxx series ATI GPU's, switch to Client V7. Link

We could also do a Teamviewer session if you want further help.

 

[headshot119]

Ok So here is the shortcut for card 1

C:\Users\Karl\Desktop\Downloads\Folding@home-Win32-GPU_Vista-623\Folding@home-Win32-GPU.exe -gpu 0 -forcegpu ati_r700

and card 2

"C:\Users\Karl\Desktop\Downloads\Folding@home-Win32-GPU_Vista-623 - Copy\Folding@home-Win32-GPU.exe" -gpu 1 -forcegpu ati_r700

Card 1 config file

[settings]
username=headshot119
team=50711
passkey=
asknet=no
machineid=3
local=4

[http]
active=no
host=localhost
port=8080
usereg=no


and card 2

[settings]
username=headshot119
team=50711
passkey=
asknet=no
machineid=2
local=4

[http]
active=no
host=localhost
port=8080
usereg=no

 

[thebluebumblebee]

HS119, No guarantees with this, but try adding -local to each of those shortcuts.

 

[headshot119]

ok quick update.

One thing I forgot to mention was that the work units where always failing 4 or 5 times to upload before they finally went through. After adding the -local flag they have gone through first time. I've also staggered one card about 6 minutes behind the other to see if that helps. We'll see when the stats next update.

 

[headshot119]

Sorted well at least I think it is. Score has jumped from 7k to 8k which fits with me completing 2 work units. thanks to both of you.

 

[Athlonite]

hmmm that's a rather old version of FAH why are you not using 6.40 you'll not need the -forcegpu ati_r700 with the newer version and why is the machine ID 3 and 4 instead of 2 and 3

 

[thebluebumblebee]

hmmm that's a rather old version of FAH why are you not using 6.40 you'll not need the -forcegpu ati_r700 with the newer version and why is the machine ID 3 and 4 instead of 2 and 3

He got it working, don't fix what isn't broken.

Machine ID doesn't matter, as long as each number is only used once per machine, 1-16.

HS119, did you input the Environmental Variables. (I don't recall saying anything to you....) Until Fahcore_16 is released for the ATI cards, there is no reason to change anything. I saw where someone with an ATI card installed client V7 and they found out it was still running GPU2.

 

[bogmali]

hmmm that's a rather old version of FAH why are you not using 6.40 you'll not need the -forcegpu ati_r700 with the newer version and why is the machine ID 3 and 4 instead of 2 and 3

Very true......v. 6.40 eliminates the need for the r700 flag. He's OK with the machine ID as long as they're different.

One question though, when you check your display configuration within windows can you see that your 2nd GPU is extended to the 1st one? Are you using a dummy plug?

 

[headshot119]

Environmental variables? What are those?

In regards to using an old version. It's the version I've always used it's always worked fine apart from now. I don't fancy changing what isn't broken.

Both cards are running in xfire, no dummy plug used and it seems to work.

Thanks Karl

 

[thebluebumblebee]

Environmental variables? What are those?

In regards to using an old version. It's the version I've always used it's always worked fine apart from now. I don't fancy changing what isn't broken.

Both cards are running in xfire, no dummy plug used and it seems to work.

Thanks Karl

You have to disable Xfire.

EV"S: Step 5 in the first post and read my post at #24 for more info at:http://www.techpowerup.com/forums/showthread.php?t=110212
The EV's will increase the utilization of the GPU's (more PPD) and reduce the load on the CPU.

 

[headshot119]

Done all the environmental variables. Looking at DC stats I've definitely uploaded two work units today, so both cards are now submitting.

Turned Xfire off and only 1 card would fold even after a reboot. So now back in xfire.

 

[thebluebumblebee]

Done all the environmental variables. Looking at DC stats I've definitely uploaded two work units today, so both cards are now submitting.

Turned Xfire off and only 1 card would fold even after a reboot. So now back in xfire.

It will never cease to amaze me how different the experiences are of different Folders. Sorry for the bad advice.

 

[headshot119]

In fairness I always thought you couldn't fold with Xfire on, and was most surprised when it actually worked.

 

[thebluebumblebee]

...And it's weird how and when things change. I don't Fold over the summer. When I turned my dual 9600GSO box off last spring, it was Folding in SLI. Turned it back on in the fall and it would not Fold with SLI and I hadn't changed anything! It's frustrating trying to help people when things change without warning.

 

[Athlonite]

Very true......v. 6.40 eliminates the need for the r700 flag. He's OK with the machine ID as long as they're different.

One question though, when you check your display configuration within windows can you see that your 2nd GPU is extended to the 1st one? Are you using a dummy plug?

with V6.40 there's no need of a dummy plug they fixed that particular little bug

 

[Athlonite]

any news on V7 yet ??? should be released real soon if it hasn't already they were say something about this quarter which is nearly up

 

[thebluebumblebee]

any news on V7 yet ??? should be released real soon if it hasn't already they were say something about this quarter which is nearly up

http://folding.typepad.com/news/2011/03/client-version-7-now-in-open-beta.html
http://folding.typepad.com/news/201...e-on-nvidia-gpu-support-for-older-boards.html

 

[Athlonite]

I hadn't been over there since I stopped foldin a while back because of the Power bill gettin to high so BIG thanks finally maybe I can start up again and do some real computing with my 2x HD5770

well D/L'd it installed it and WTF I'm still only doin core_11 WU's errrrr why did I even bother maybe core_16 WU's are in short supply

in other news V7 was the easiest setup of F@H to date so that's a good thing finally SMP + multi GPU used to be a biatch to get running but this version is a piece of cake to get it up and running

 

[headshot119]

Is it just because of server loads that my WUs never seem to want to send after 23:00 GMT+1? Then tend to keep retrying till 01:00?

 

[thebluebumblebee]

Is it just because of server loads that my WUs never seem to want to send after 23:00 GMT+1? Then tend to keep retrying till 01:00?

So you have a 2 hour period where you can not return results? Would you please post the contents of the fahlog for that?

 

[Athlonite]

even longer here when trying to send smp results

23:47:20:Sending unit results: id:00 state:SEND project:6053 run:1 clone:67 gen:169 core:0xa3 unit:0x2c27173b4d96e01000a90043000117a5
23:47:20:Unit 00: Uploading 22.73KiB
23:47:20:Connecting to 171.64.65.54:8080
23:47:21:WARNING: Exception: Failed to send results to work server: Upload failed
23:47:21:Trying to send results to collection server
23:47:21:Unit 00: Uploading 22.73KiB
23:47:21:Connecting to 171.67.108.25:8080
23:47:23:WARNING: WorkServer connection failed on port 8080 trying 80
23:47:23:Connecting to 171.67.108.25:80
23:47:24:ERROR: Exception: Failed to connect to 171.67.108.25:80: No connection could be made because the target machine actively refused it.

that msg is what I get everytime the new client tries to send work back so far it's upto attempt #22

 

[thebluebumblebee]

Is it just because of server loads that my WUs never seem to want to send after 23:00 GMT+1? Then tend to keep retrying till 01:00?

even longer here when trying to send smp results

23:47:20:Sending unit results: id:00 state:SEND project:6053 run:1 clone:67 gen:169 core:0xa3 unit:0x2c27173b4d96e01000a90043000117a5
23:47:20:Unit 00: Uploading 22.73KiB
23:47:20:Connecting to 171.64.65.54:8080
23:47:21:WARNING: Exception: Failed to send results to work server: Upload failed
23:47:21:Trying to send results to collection server
23:47:21:Unit 00: Uploading 22.73KiB
23:47:21:Connecting to 171.67.108.25:8080
23:47:23:WARNING: WorkServer connection failed on port 8080 trying 80
23:47:23:Connecting to 171.67.108.25:80
23:47:24:ERROR: Exception: Failed to connect to 171.67.108.25:80: No connection could be made because the target machine actively refused it.

that msg is what I get everytime the new client tries to send work back so far it's upto attempt #22

The Folding forums has this guide: Troubleshooting Server Connectivity Issues (Do This First)

 

[Athlonite]

i figured I'd just rather not bother with smp and stick with GPU Wu's the new ocl based ones are up and running here and 1835 points a piece i stand to make more with just those

plus I'd already tried the server con issues thing already didn't make jot of difference