GPU causing Computer Freeze

[xxlawman87xx]

I am folding on a GTX480, a standard 9800GT and an Energy Efficient 9800GT. I can start all three clients just fine. However, after about 5-15 minutes it causes my computer to freeze completely, requiring a reboot. This is with everything at stock speeds. I can run any single card individually with no problem, but shortly after I run them all it freezes. Suggestions? BTW, I am running the newest systray client. 6.40 i think.

[PantherX]

My guess would be that the PSU is being overloaded so can you give us your hardware details? Also, can you please provide us a copy of your FAHlogs (viewtopic.php?p=160884#p160884) so we can see if there is any error and how you have setup the Clients.

BTW, when you say "freeze" do you mean:
A) Desktop lagging i.e. when you move your mouse, it takes a significant time for it to move on the screen
B) The system will not respond at all? If that's the case, can you see your HDD LED blinking, continuously lit or not lit at all?

[xxlawman87xx]

I have a 1000w psu with a Q9550 stock so PSU should not be a problem


LOG:

Code: Select all

--- Opening Log file [February 5 04:49:41 UTC] 


# Windows GPU Systray Edition #################################################
###############################################################################

                       Folding@Home Client Version 6.41r2

                          http://folding.stanford.edu

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

Launch directory: C:\Users\Thomas\AppData\Roaming\Folding@home-gpu3
Arguments: -gpu 2 -forcegpu nvidia_g80 

[04:49:41] Gpu type=2 species=11.
[04:50:09] - Ask before connecting: No
[04:50:09] - User name: xxlawman87xx (Team 37726)
[04:50:09] - User ID: 72D8AF53120D9092
[04:50:09] - Machine ID: 4
[04:50:09] 
[04:50:09] Work directory not found. Creating...
[04:50:09] Could not open work queue, generating new queue...
[04:50:09] Initialization complete
[04:50:09] - Preparing to get new work unit...
[04:50:09] Cleaning up work directory
[04:50:09] + Attempting to get work packet
[04:50:09] Gpu type=2 species=11.
[04:50:09] - Connecting to assignment server
[04:50:09] - Successful: assigned to (171.64.65.61).
[04:50:09] + News From Folding@Home: Welcome to Folding@Home
[04:50:10] Loaded queue successfully.
[04:50:10] Gpu type=2 species=11.
[04:50:11] + Closed connections
[04:50:11] 
[04:50:11] + Processing work unit
[04:50:11] Core required: FahCore_11.exe
[04:50:11] Core not found.
[04:50:11] - Core is not present or corrupted.
[04:50:11] - Attempting to download new core...
[04:50:11] + Downloading new core: FahCore_11.exe
[04:50:11] + 10240 bytes downloaded
[04:50:11] + 20480 bytes downloaded
[04:50:11] + 30720 bytes downloaded
[04:50:11] + 40960 bytes downloaded
[04:50:11] + 51200 bytes downloaded
[04:50:11] + 61440 bytes downloaded
[04:50:11] + 71680 bytes downloaded
[04:50:11] + 81920 bytes downloaded
[04:50:11] + 92160 bytes downloaded
[04:50:11] + 102400 bytes downloaded
[04:50:11] + 112640 bytes downloaded
[04:50:11] + 122880 bytes downloaded
[04:50:11] + 133120 bytes downloaded
[04:50:11] + 143360 bytes downloaded
[04:50:11] + 153600 bytes downloaded
[04:50:11] + 163840 bytes downloaded
[04:50:11] + 174080 bytes downloaded
[04:50:11] + 184320 bytes downloaded
[04:50:11] + 194560 bytes downloaded
[04:50:11] + 204800 bytes downloaded
[04:50:11] + 215040 bytes downloaded
[04:50:11] + 225280 bytes downloaded
[04:50:11] + 235520 bytes downloaded
[04:50:11] + 245760 bytes downloaded
[04:50:12] + 256000 bytes downloaded
[04:50:12] + 266240 bytes downloaded
[04:50:12] + 276480 bytes downloaded
[04:50:12] + 286720 bytes downloaded
[04:50:12] + 296960 bytes downloaded
[04:50:12] + 307200 bytes downloaded
[04:50:12] + 317440 bytes downloaded
[04:50:12] + 327680 bytes downloaded
[04:50:12] + 337920 bytes downloaded
[04:50:12] + 348160 bytes downloaded
[04:50:12] + 358400 bytes downloaded
[04:50:12] + 368640 bytes downloaded
[04:50:12] + 378880 bytes downloaded
[04:50:12] + 389120 bytes downloaded
[04:50:12] + 399360 bytes downloaded
[04:50:12] + 409600 bytes downloaded
[04:50:12] + 419840 bytes downloaded
[04:50:12] + 430080 bytes downloaded
[04:50:12] + 440320 bytes downloaded
[04:50:12] + 450560 bytes downloaded
[04:50:12] + 460800 bytes downloaded
[04:50:12] + 471040 bytes downloaded
[04:50:12] + 481280 bytes downloaded
[04:50:12] + 491520 bytes downloaded
[04:50:12] + 501760 bytes downloaded
[04:50:12] + 512000 bytes downloaded
[04:50:12] + 522240 bytes downloaded
[04:50:12] + 532480 bytes downloaded
[04:50:12] + 542720 bytes downloaded
[04:50:12] + 552960 bytes downloaded
[04:50:12] + 563200 bytes downloaded
[04:50:12] + 573440 bytes downloaded
[04:50:12] + 583680 bytes downloaded
[04:50:12] + 593920 bytes downloaded
[04:50:12] + 604160 bytes downloaded
[04:50:13] + 614400 bytes downloaded
[04:50:13] + 624640 bytes downloaded
[04:50:13] + 634880 bytes downloaded
[04:50:13] + 645120 bytes downloaded
[04:50:13] + 655360 bytes downloaded
[04:50:13] + 665067 bytes downloaded
[04:50:13] Verifying core Core_11.fah...
[04:50:13] Signature is VALID
[04:50:13] 
[04:50:13] Trying to unzip core FahCore_11.exe
[04:50:13] Decompressed FahCore_11.exe (1908736 bytes) successfully
[04:50:18] + Core successfully engaged
[04:50:23] 
[04:50:23] + Processing work unit
[04:50:23] Core required: FahCore_11.exe
[04:50:23] Core found.
[04:50:23] Working on queue slot 01 [February 5 04:50:23 UTC]
[04:50:23] + Working ...
[04:50:23] 
[04:50:23] *------------------------------*
[04:50:23] Folding@Home GPU Core
[04:50:23] Version 1.31 (Tue Sep 15 10:57:42 PDT 2009)
[04:50:23] 
[04:50:23] Compiler  : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86 
[04:50:23] Build host: amoeba
[04:50:23] Board Type: Nvidia
[04:50:23] Core      : 
[04:50:23] Preparing to commence simulation
[04:50:23] - Looking at optimizations...
[04:50:23] DeleteFrameFiles: successfully deleted file=work/wudata_01.ckp
[04:50:23] - Created dyn
[04:50:23] - Files status OK
[04:50:23] - Expanded 73803 -> 383588 (decompressed 519.7 percent)
[04:50:23] Called DecompressByteArray: compressed_data_size=73803 data_size=383588, decompressed_data_size=383588 diff=0
[04:50:23] - Digital signature verified
[04:50:23] 
[04:50:23] Project: 6602 (Run 8, Clone 459, Gen 480)
[04:50:23] 
[04:50:23] Assembly optimizations on if available.
[04:50:23] Entering M.D.
[04:50:29] Tpr hash work/wudata_01.tpr:  894508150 2936055791 4039105231 736012801 223861453
[04:50:29] 
[04:50:29] Calling fah_main args: 14 usage=100
[04:50:29] 
[04:50:30] Working on Protein
[04:50:31] Client config found, loading data.
[04:50:31] Starting GUI Server
[04:52:00] Completed 1%
[04:53:28] Completed 2%
[04:54:57] Completed 3%
[04:56:26] Completed 4%
[04:57:55] Completed 5%
[04:59:24] Completed 6%
[05:00:53] Completed 7%
[05:02:21] Completed 8%

This is where the whole rig froze

Mod Edit: Added Code Tags - PantherX

[xxlawman87xx]

My guess would be that the PSU is being overloaded so can you give us your hardware details? Also, can you please provide us a copy of your FAHlogs (viewtopic.php?p=160884#p160884) so we can see if there is any error and how you have setup the Clients.

BTW, when you say "freeze" do you mean:
A) Desktop lagging i.e. when you move your mouse, it takes a significant time for it to move on the screen
B) The system will not respond at all? If that's the case, can you see your HDD LED blinking, continuously lit or not lit at all?

Option B. LOL....I will check the HDD light to see what it does.

[Fahrenheit451]

@PantherX: Do you think it's a lack of memory issue? How much is required if you run three gfx cards in one system?

[P5-133XL]

When you run into issues where each individual client runs fine but multiple clients cause the computer to bluescreen, freeze, or otherwise totally die then I would start by looking at temperatures all the cards as well as the CPU. Because it takes a few minutes for your machine to die and it normally takes a few minutes for things to heat up I would make sure that nothing is overheating.

Secondly, I understand that you have confidence in your PS but that really is the most likely cause. Even if you have lots of total power capability, it is quite common for there to be limitations such as 1000W total but only 500W combing all the 12V rails. You also need to make sure that you are not accidentally sharing 12V rails. Next you must look at the max amperage allowed on each of your 12V rails and compare them to the maximum current draw for each of the cards i.e. make sure that no cards can exceed the current allowed on your specific 12V rails.

[Nathan_P]

Secondly, I understand that you have confidence in your PS but that really is the most likely cause. Even if you have lots of total power capability, it is quite common for there to be limitations such as 1000W total but only 500W combing all the 12V rails. You also need to make sure that you are not accidentally sharing 12V rails. Next you must look at the max amperage allowed on each of your 12V rails and compare them to the maximum current draw for each of the cards i.e. make sure that no cards can exceed the current allowed on your specific 12V rails.

^^^This, there are 1000w psu's and there are 1000w psu's that actually give you 1000w. Brand and model name please and we can soon give you an answer

[Fahrenheit451]

Yes, power consumption could be also a reason.

@xxlawman87xx: Do you tested it with two cards only?

[xxlawman87xx]

Hey guys sorry I just woke up. I have a BFG 1000watt PSU and it allow 28amps per rail. I have two rails running to the 480 and one to the 9800gt. The 9800GT EE does not require an auxillery power source. I did some testing and the problem seems to be the regular 9800gt even when i run it by itself the comp freezes after a while. after look at the temps on the card, it hits 83c then it freezes. I have an after market cooler that I am gonna put on it and see what that does. I will keep you guys informed. thanks for your help so far.

EDIT: the rails are also linked.

[xxlawman87xx]

Good news, the problem seems to be heat. I put the aftermarket cooler on it and so far (30+ min) it hasnt locked up.

[xxlawman87xx]

When you run into issues where each individual client runs fine but multiple clients cause the computer to bluescreen, freeze, or otherwise totally die then I would start by looking at temperatures all the cards as well as the CPU. Because it takes a few minutes for your machine to die and it normally takes a few minutes for things to heat up I would make sure that nothing is overheating.

Secondly, I understand that you have confidence in your PS but that really is the most likely cause. Even if you have lots of total power capability, it is quite common for there to be limitations such as 1000W total but only 500W combing all the 12V rails. You also need to make sure that you are not accidentally sharing 12V rails. Next you must look at the max amperage allowed on each of your 12V rails and compare them to the maximum current draw for each of the cards i.e. make sure that no cards can exceed the current allowed on your specific 12V rails.

Your overheating intuition was correct. LOL....with the aftermarket cooler its been running for several hours with no problem. Its topping out at 57c...much better than 83c.

[Leonardo]

I put the aftermarket cooler on it

"It," meaning the CPU?

You really should start monitoring the temperatures. Core Temp is a good monitor for CPUs and MSI Afterburner is good for GPUs. There are others that are good as well.

[xxlawman87xx]

I put the aftermarket cooler on it

"It," meaning the CPU?

You really should start monitoring the temperatures. Core Temp is a good monitor for CPUs and MSI Afterburner is good for GPUs. There are others that are good as well.

"it" is the GPU. Secondly, I manage/monitor temps to the hilt, i run both temp monitoring software and temp monitoring hardware as any hardcore enthusiast would. Everything is liquid cooled except these auxiliary 9800gt's I know 83c is pushing it for a GPU but I did not know it would cause the system to freeze. My temps never get above 50c on any part of my system aside from those folding gpus. no matter, problem solved.

[Leonardo]

Sorry, I didn't mean to appear condescending.

I run several systems with 3 X GPU each. I really had to work on case airflow to feed the GPUs enough cool air. It's a challenge.

[xxlawman87xx]

Sorry, I didn't mean to appear condescending.

I run several systems with 3 X GPU each. I really had to work on case airflow to feed the GPUs enough cool air. It's a challenge.

Its okay, its been a rough day, i am a little touchy right now. I feel your pain on airflow, that why I went to water cooling. sick of hearing WHIRRRRRRRRRRRRRRRRRRR......lol