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

[mstenholm]

Update: Installed win7 again as well as all drivers. Played around with the affinity's of the clients and now she is producing 26.5K ppd with the 3x GPU2 clients. Total for the rig is 49K+/-. I'm satisfied for now, but feel there is more O/C left in this chip. Currently @ 3.8Ghz/1.35V/190bclk/x20multi running 68C on air. Gonna shoot for 3.9/4.0 next week.

49 k is not bad. Did you end up staying with the 7 treads for the CPU?

All you need is to put the rig in a Danish open window and then shoot for 4.2 GHz. I open one from time to time and see a 10 degree C drop on my GPU's. My hardworking hex still struggles go higher then 40 C @ 4.1 GHZ in its 0-5 C environment. I did have an EUE (core communication something) @ 84 % but completed it in the second run, so I will stay at 4.1 GHz/1.26 V but it is tempting to go higher when temps are that low.

 

[BUCK NASTY]

49 k is not bad. Did you end up staying with the 7 treads for the CPU?

All you need is to put the rig in a Danish open window and then shoot for 4.2 GHz. I open one from time to time and see a 10 degree C drop on my GPU's. My hardworking hex still struggles go higher then 40 C @ 4.1 GHZ in its 0-5 C environment. I did have an EUE (core communication something) @ 84 % but completed it in the second run, so I will stay at 4.1 GHz/1.26 V but it is tempting to go higher when temps are that low.

I'm running smp-7, but I'm allowing -bigadv access to all cores. The 3x GPU's are locked to Core 0. I've tried several configurations and this one is the most productive. Now i just need to check the wattage that it is using. I have been lucky and have not lost a WU in about 2 months.

 

[mstenholm]

I'm running smp-7, but I'm allowing -bigadv access to all cores. The 3x GPU's are locked to Core 0. I've tried several configurations and this one is the most productive. Now i just need to check the wattage that it is using. I have been lucky and have not lost a WU in about 2 months.

Seems to be the way to go. I never tried with more then one GPU(2) on either on my -bigadv rigs but I hate to see the 2-2½ minutes penalty I get per frame. If/when I have a 2684 I run my GPU. When the TPF gets that long the bonus is less affected anyway.

Well, this one lost WU was the first one in a long time too and since that is the case I will no ease on my OC.

 

[Ross211]

I just installed a 9800 GTX+ into a secondary rig of mine and want to run the GPU2 client on it. I've tried the stock GPU2 client and also the FAH GPU Tracker V2 client and I end up getting the same error using both.

This system with the 9800 GTX+ does have an onboard Radeon Xpress 200 VGA adapter. I had to go into the BIOS on this board and set the PCI-E as the primary display adapter. I removed the ATI catalyst software and removed all related ATI software. I even ran Driver Sweeper afterwards. I'm using the 260.99 Nvidia driver on Windows Server 2008 32 bit. The 9800 GTX+ can withstand 10 minutes of Kombuster and temps don't get out of hand so I'm pretty certain the card is fine (I do have an aggressive fan profile set on the card via Afterburner).

Here is what I get in the log from both the clients when trying to run them. (This one below is from the WU Failures directory on the GPU Tracker client)

--- Opening Log file [January 1 07:36:23 UTC]


# Windows GPU Console Edition #################################################
###############################################################################

Folding@Home Client Version 6.23

http://folding.stanford.edu

###############################################################################
###############################################################################

Launch directory: C:\Users\Ross\Documents\FAH GPU Tracker V2\GPU0
Executable: C:\Users\Ross\Documents\FAH GPU Tracker V2\FAH_GPU2.exe
Arguments: -oneunit -gpu 0

[07:36:23] - Ask before connecting: No
[07:36:23] - User name: Ross211 (Team 50711)
[07:36:23] - User ID: 72663AEC695CEB6A
[07:36:23] - Machine ID: 3
[07:36:23]
[07:36:23] Work directory not found. Creating...
[07:36:23] Could not open work queue, generating new queue...
[07:36:23] - Preparing to get new work unit...
[07:36:23] + Attempting to get work packet
[07:36:23] - Connecting to assignment server
[07:36:24] - Successful: assigned to (171.67.108.11).
[07:36:24] + News From Folding@Home: Welcome to Folding@Home
[07:36:24] Loaded queue successfully.
[07:36:25] + Closed connections
[07:36:25]
[07:36:25] + Processing work unit
[07:36:25] Core required: FahCore_11.exe
[07:36:25] Core found.
[07:36:25] Working on queue slot 01 [January 1 07:36:25 UTC]
[07:36:25] + Working ...
[07:36:25]
[07:36:25] *------------------------------*
[07:36:25] Folding@Home GPU Core
[07:36:25] Version 1.31 (Tue Sep 15 10:57:42 PDT 2009)
[07:36:25]
[07:36:25] Compiler : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86
[07:36:25] Build host: amoeba
[07:36:25] Board Type: Nvidia
[07:36:25] Core :
[07:36:25] Preparing to commence simulation
[07:36:25] - Looking at optimizations...
[07:36:25] DeleteFrameFiles: successfully deleted file=work/wudata_01.ckp
[07:36:25] - Created dyn
[07:36:25] - Files status OK
[07:36:25] - Expanded 45411 -> 251112 (decompressed 552.9 percent)
[07:36:25] Called DecompressByteArray: compressed_data_size=45411 data_size=251112, decompressed_data_size=251112 diff=0
[07:36:25] - Digital signature verified
[07:36:25]
[07:36:25] Project: 5769 (Run 0, Clone 219, Gen 121)
[07:36:25]
[07:36:25] Assembly optimizations on if available.
[07:36:25] Entering M.D.
[07:36:31] Tpr hash work/wudata_01.tpr: 3281136117 2544728572 4217056393 3144557638 4075114056
[07:36:31]
[07:36:31] Calling fah_main args: 14 usage=100
[07:36:31]
[07:36:32] Working on Protein
[07:36:32] mdrun_gpu returned
[07:36:32] Self-test failure
[07:36:32]
[07:36:32] Folding@home Core Shutdown: UNSTABLE_MACHINE
[07:36:35] CoreStatus = 7A (122)
[07:36:35] Sending work to server
[07:36:35] Project: 5769 (Run 0, Clone 219, Gen 121)
[07:36:35] - Read packet limit of 540015616... Set to 524286976.
[07:36:35] - Error: Could not get length of results file work/wuresults_01.dat
[07:36:35] - Error: Could not read unit 01 file. Removing from queue.
[07:36:35] + -oneunit flag given and have now finished a unit. Exiting.
Folding@Home Client Shutdown.

 

[thebluebumblebee]

Are you running it as administrator?

 

[Ross211]

Nope but I just gave it a shot and it fails and gives me the same error in the log. Thanks hopefully we can figure this out.

 

[oily_17]

Take a look at this thread, seems people are having trouble with their cards when folding 57** projects.

Only thing I can think of is to let it run without -oneunit flag. It will probably fail 6 units in a row then pause, when it pauses restart the client and see if it will pick up a different WU project.

 

[hat]

//nevermind, worked it out

 

[bogmali]

Ross11-is your onboard graphics card enabled? It could be that you're folding on that ATI Xpress200 and that's why it's crashing on you Disable it completely from that BIOS and check your display settings on the control panel and make sure that the 9800GTX is your default display adapter. Hope it helps

 

[Ross211]

I have an option in my ASUS P5RC-LA that is "Primary Video Adapter" and I have it set to PCI-E. I actually had to go into BIOS with my onboard after putting a PCI-E card in, the board won't automatically change output. It gives me the option of choosing PCI, Onboard, and PCI-E. POS...

 

[msgclb]

I have an option in my ASUS P5RC-LA that is "Primary Video Adapter" and I have it set to PCI-E. I actually had to go into BIOS with my onboard after putting a PCI-E card in, the board won't automatically change output. It gives me the option of choosing PCI, Onboard, and PCI-E. POS...

Maybe your onboard video is seen as -GPU 0 so you might be able to use -GPU 1 to resolve the conflict.

 

[Ross211]

Yeah I did try that in the console version and also in FAH GPU Tracker V2. When I run autodetect in FAH GPU Tracker V2 it finds GPU0 as the 9800 GTX+ and sets GPU0 as Nvidia G80. I tried manually setting it to GPU1 and it fails.

 

[PhysXerror]

Is it possible that my ip has been blocked by Stanford? I cant download any WUs nor can i even get onto the home page. Its been like this for near a day now, Ive tried everything i can think of but ive had no luck finding the problem

[02:01:00] --------------------------------------------------
[02:01:00] FAH GPU Tracker v3.43 Startup
[02:01:00] --------------------------------------------------
[02:01:00] Detected OS: Windows 7
[02:01:02] Detected CPU: AMD Phenom(tm) II X4 925 Processor
[02:01:02] 4 Cores, 4 Threads, 3,360 MHz
[02:01:02] Creating client folders...
[02:01:02] Downloading FAH clients...
[02:02:43] FATAL ERROR: Could not download clients
[02:02:43] Check your internet connection, then try Setup -> Download FAH Clients
[02:02:43] User cancelled FAH client download
[02:02:43] FATAL ERROR: Could not download clients
[02:02:43] Check your internet connection, then try Setup -> Download FAH Clients
[02:02:43] User cancelled FAH client download
[02:02:43] FATAL ERROR: Could not download clients
[02:02:43] Check your internet connection, then try Setup -> Download FAH Clients
[02:02:43] User cancelled FAH client download
[02:02:43] FATAL ERROR: Could not download clients
[02:02:43] Check your internet connection, then try Setup -> Download FAH Clients
[02:02:43] User cancelled FAH client download
[02:02:43] FATAL ERROR: Could not download clients
[02:02:43] Check your internet connection, then try Setup -> Download FAH Clients

This has happened on both my main and backup rigs on all clients

 

[thebluebumblebee]

Unless you've installed a new firewall, anti-virus or anti-spyware, that sounds like a question for your ISP.

 

[PhysXerror]

Unless you've installed a new firewall, anti-virus or anti-spyware, that sounds like a question for your ISP.

Hmm it seems they, for whatever reason, are blocking it. I can get in via proxy so that sorta says it all :shadedshu

 

[PhysXerror]

Just got a reply email from my ISP, apparently its a known issue and they are looking into it. Damn, looks like i wont be folding for awhile

Edit: Thankfully they fixed it up fairly fast!

 

[Athlonite]

physXerror who's ur ISP not Orcon is it as thats the same error msg I got

 

[erocker]

I need the non-systray GPU3 client for my GTX 580. Also, can someone just give me the quick rundown 1.2.3.. (or provide a link) on setting it up for the best output. I did have it set up correctly previously, but I kinda forgot the process.

Thanks!

 

[thebluebumblebee]

I need the non-systray GPU3 client for my GTX 580. Also, can someone just give me the quick rundown 1.2.3.. (or provide a link) on setting it up for the best output. I did have it set up correctly previously, but I kinda forgot the process.

Thanks!

Just use this:
http://fahtracker.com/

EDIT: Walk through here: Setting-up Folding Clients With FAH GPU Tracker V...

 

[Sinzia]

So, I have 2 9800GT's and a GTX470 running with a 4.5GHz core 2 duo.

Should I be running SMP on both cores? Just one? Or skip SMP all together?

I ask because when I added the other 9800GT, my PPD droped from ~17k to ~13k

 

[bogmali]

I wouldn't run SMP on it all since you have 3 GPU's folding, it will just bog down the system thus decreasing your PPD.

 

[PhysXerror]

physXerror who's ur ISP not Orcon is it as thats the same error msg I got

Hey man, Nope im with Slingshot

 

[Sinzia]

I wouldn't run SMP on it all since you have 3 GPU's folding, it will just bog down the system thus decreasing your PPD.

Thats what I was thinking as well, thanks!

 

[Athlonite]

Hey man, Nope im with Slingshot

ah ok still sounds like something that Orcon would do

 

[qu4k3r]

I have a strange read on FAH tracker... 850K ppd on gpu0 lol