Constant EUE Problems

Moderators: slegrand, Site Moderators, PandeGroup

Constant EUE Problems

Postby burn.intel » Thu Feb 25, 2010 4:57 am

Okay, so I have had nothing but problems after putting together my new dedicated rig. here are the specs:
Q6600
Foxconn ELA P45 mobo
2x1gb Corsair XMS2 ddr2 800
1x xfx 9800gt
2x gigabyte 8800gt (not currently being used... was having issues)
PC Power and Cooling 750W silencer PSU
320gb seagate barracuda

---

First I tried running all 3 cards together, installed newest drivers, followed the multi gpu setup on stanford's site, all that went fine. So I try to run f@h on them, 1 card (middle pci-e slot) reaches the EUE limit after about an hour depending on the WU, and the other two will complete 1 WU and then EUE all the rest before even the first %, EVERY time. I've reinstalled with about 3 different driver versions, reinstalled f@h a bunch of times, bumped up the motherboard's voltages (was thinking 3 cards was straining the NB too much), nothing seems to work. the 9800gt ran PERFECTLY fine on my other rig (recently upgraded) for weeks, and now it won't even finish more than one WU.

I'm thinking this all points to a motherboard issue, but I figured I'd get some more opinions on it before I have to issue an RMA

Here's a log for the 1 good WU:
Code: Select all
[03:17:54] + Processing work unit
[03:17:54] Core required: FahCore_11.exe
[03:17:54] Core found.
[03:17:54] Working on queue slot 01 [February 25 03:17:54 UTC]
[03:17:54] + Working ...
[03:17:54]
[03:17:54] *------------------------------*
[03:17:54] Folding@Home GPU Core
[03:17:54] Version 1.31 (Tue Sep 15 10:57:42 PDT 2009)
[03:17:54]
[03:17:54] Compiler  : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86
[03:17:54] Build host: amoeba
[03:17:54] Board Type: Nvidia
[03:17:54] Core      :
[03:17:54] Preparing to commence simulation
[03:17:54] - Looking at optimizations...
[03:17:54] DeleteFrameFiles: successfully deleted file=work/wudata_01.ckp
[03:17:55] - Created dyn
[03:17:55] - Files status OK
[03:17:55] - Expanded 22183 -> 146631 (decompressed 661.0 percent)
[03:17:55] Called DecompressByteArray: compressed_data_size=22183 data_size=146631, decompressed_data_size=146631 diff=0
[03:17:55] - Digital signature verified
[03:17:55]
[03:17:55] Project: 5799 (Run 9, Clone 57, Gen 12)
[03:17:55]
[03:17:55] Assembly optimizations on if available.
[03:17:55] Entering M.D.
[03:18:01] Tpr hash work/wudata_01.tpr:  2683761015 1676872079 4008354383 1962280214 3511978832
[03:18:01]
[03:18:01] Calling fah_main args: 14 usage=100
[03:18:01]
[03:18:01] Working on Protein
[03:18:01] Client config found, loading data.
[03:18:01] Starting GUI Server
[03:18:52] Completed 1%
[03:19:42] Completed 2%
[03:20:32] Completed 3%
[03:21:22] Completed 4%
[03:22:12] Completed 5%
[03:23:03] Completed 6%
[03:23:56] Completed 7%
[03:24:42] Completed 8%
[03:25:26] Completed 9%
[03:26:10] Completed 10%
[03:26:53] Completed 11%
[03:27:38] Completed 12%
[03:28:24] Completed 13%
[03:29:08] Completed 14%
[03:29:52] Completed 15%
[03:30:36] Completed 16%
[03:31:21] Completed 17%
[03:32:05] Completed 18%
[03:32:48] Completed 19%
[03:33:31] Completed 20%
[03:34:14] Completed 21%
[03:34:57] Completed 22%
[03:35:39] Completed 23%
[03:36:22] Completed 24%
[03:37:05] Completed 25%
[03:37:47] Completed 26%
[03:38:31] Completed 27%
[03:39:16] Completed 28%
[03:40:01] Completed 29%
[03:40:46] Completed 30%
[03:41:31] Completed 31%
[03:42:16] Completed 32%
[03:43:01] Completed 33%
[03:43:46] Completed 34%
[03:44:30] Completed 35%
[03:45:15] Completed 36%
[03:46:00] Completed 37%
[03:46:45] Completed 38%
[03:47:30] Completed 39%
[03:48:14] Completed 40%
[03:48:57] Completed 41%
[03:49:39] Completed 42%
[03:50:23] Completed 43%
[03:51:05] Completed 44%
[03:51:48] Completed 45%
[03:52:31] Completed 46%
[03:53:13] Completed 47%
[03:53:56] Completed 48%
[03:54:39] Completed 49%
[03:55:22] Completed 50%
[03:56:04] Completed 51%
[03:56:47] Completed 52%
[03:57:30] Completed 53%
[03:58:12] Completed 54%
[03:58:55] Completed 55%
[03:59:38] Completed 56%
[04:00:21] Completed 57%
[04:01:03] Completed 58%
[04:01:46] Completed 59%
[04:02:29] Completed 60%
[04:03:12] Completed 61%
[04:03:54] Completed 62%
[04:04:37] Completed 63%
[04:05:20] Completed 64%
[04:06:03] Completed 65%
[04:06:46] Completed 66%
[04:07:28] Completed 67%
[04:08:11] Completed 68%
[04:08:54] Completed 69%
[04:09:37] Completed 70%
[04:10:19] Completed 71%
[04:11:02] Completed 72%
[04:11:45] Completed 73%
[04:12:27] Completed 74%
[04:13:10] Completed 75%
[04:13:53] Completed 76%
[04:14:36] Completed 77%
[04:15:18] Completed 78%
[04:16:01] Completed 79%
[04:16:43] Completed 80%
[04:17:26] Completed 81%
[04:18:09] Completed 82%
[04:18:52] Completed 83%
[04:19:34] Completed 84%
[04:20:17] Completed 85%
[04:21:00] Completed 86%
[04:21:43] Completed 87%
[04:22:25] Completed 88%
[04:23:08] Completed 89%
[04:23:51] Completed 90%
[04:24:33] Completed 91%
[04:25:16] Completed 92%
[04:25:59] Completed 93%
[04:26:42] Completed 94%
[04:27:24] Completed 95%
[04:28:07] Completed 96%
[04:28:50] Completed 97%
[04:29:32] Completed 98%
[04:30:15] Completed 99%
[04:30:58] Completed 100%
[04:30:58] Successful run
[04:30:58] DynamicWrapper: Finished Work Unit: sleep=10000
[04:31:08] Reserved 58852 bytes for xtc file; Cosm status=0
[04:31:08] Allocated 58852 bytes for xtc file
[04:31:08] - Reading up to 58852 from "work/wudata_01.xtc": Read 58852
[04:31:08] Read 58852 bytes from xtc file; available packet space=786371612
[04:31:08] xtc file hash check passed.
[04:31:08] Reserved 6936 6936 786371612 bytes for arc file=<work/wudata_01.trr> Cosm status=0
[04:31:08] Allocated 6936 bytes for arc file
[04:31:08] - Reading up to 6936 from "work/wudata_01.trr": Read 6936
[04:31:08] Read 6936 bytes from arc file; available packet space=786364676
[04:31:08] trr file hash check passed.
[04:31:08] Allocated 560 bytes for edr file
[04:31:08] Read bedfile
[04:31:08] edr file hash check passed.
[04:31:08] Logfile not read.
[04:31:08] GuardedRun: success in DynamicWrapper
[04:31:08] GuardedRun: done
[04:31:08] Run: GuardedRun completed.
[04:31:10] + Opened results file
[04:31:10] - Writing 66860 bytes of core data to disk...
[04:31:10] Done: 66348 -> 63501 (compressed to 95.7 percent)
[04:31:10]   ... Done.
[04:31:10] DeleteFrameFiles: successfully deleted file=work/wudata_01.ckp
[04:31:10] Shutting down core
[04:31:10]
[04:31:10] Folding@home Core Shutdown: FINISHED_UNIT
[04:31:13] CoreStatus = 64 (100)
[04:31:13] Sending work to server
[04:31:13] Project: 5799 (Run 9, Clone 57, Gen 12)


[04:31:13] + Attempting to send results [February 25 04:31:13 UTC]
[04:31:13] + Results successfully sent
[04:31:13] Thank you for your contribution to Folding@Home.
[04:31:13] + Number of Units Completed: 6

[04:31:17] - Preparing to get new work unit...
[04:31:17] + Attempting to get work packet
[04:31:17] - Connecting to assignment server
[04:31:18] - Successful: assigned to (171.67.108.21).
[04:31:18] + News From Folding@Home: Welcome to Folding@Home
[04:31:18] Loaded queue successfully.
[04:31:19] + Closed connections
[04:31:19]


And here is one for 1 of the bad WU's:
Code: Select all
[04:31:19] + Processing work unit
[04:31:19] Core required: FahCore_11.exe
[04:31:19] Core found.
[04:31:19] Working on queue slot 02 [February 25 04:31:19 UTC]
[04:31:19] + Working ...
[04:31:19]
[04:31:19] *------------------------------*
[04:31:19] Folding@Home GPU Core
[04:31:19] Version 1.31 (Tue Sep 15 10:57:42 PDT 2009)
[04:31:19]
[04:31:19] Compiler  : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86
[04:31:19] Build host: amoeba
[04:31:19] Board Type: Nvidia
[04:31:19] Core      :
[04:31:19] Preparing to commence simulation
[04:31:19] - Looking at optimizations...
[04:31:19] DeleteFrameFiles: successfully deleted file=work/wudata_02.ckp
[04:31:19] - Created dyn
[04:31:19] - Files status OK
[04:31:19] - Expanded 65413 -> 344335 (decompressed 526.4 percent)
[04:31:19] Called DecompressByteArray: compressed_data_size=65413 data_size=344335, decompressed_data_size=344335 diff=0
[04:31:19] - Digital signature verified
[04:31:19]
[04:31:19] Project: 5782 (Run 7, Clone 75, Gen 40)
[04:31:19]
[04:31:19] Assembly optimizations on if available.
[04:31:19] Entering M.D.
[04:31:25] Tpr hash work/wudata_02.tpr:  1180181535 4208087869 3251397188 2968155575 4283936810
[04:31:25]
[04:31:25] Calling fah_main args: 14 usage=100
[04:31:25]
[04:31:25] mdrun_gpu returned
[04:31:25] Going to send back what have done -- stepsTotalG=0
[04:31:25] Work fraction=0.0000 steps=0.
[04:31:29] logfile size=4945 infoLength=4945 edr=0 trr=25
[04:31:29] + Opened results file
[04:31:29] - Writing 5483 bytes of core data to disk...
[04:31:29] Done: 4971 -> 1854 (compressed to 37.2 percent)
[04:31:29]   ... Done.
[04:31:29] DeleteFrameFiles: successfully deleted file=work/wudata_02.ckp
[04:31:29]
[04:31:29] Folding@home Core Shutdown: UNSTABLE_MACHINE
[04:31:33] CoreStatus = 7A (122)
[04:31:33] Sending work to server
[04:31:33] Project: 5782 (Run 7, Clone 75, Gen 40)
[04:31:33] - Read packet limit of 540015616... Set to 524286976.


[04:31:33] + Attempting to send results [February 25 04:31:33 UTC]
[04:31:33] + Results successfully sent
[04:31:33] Thank you for your contribution to Folding@Home.
[04:31:37] - Preparing to get new work unit...
[04:31:37] + Attempting to get work packet
[04:31:37] - Connecting to assignment server
[04:31:37] - Successful: assigned to (171.67.108.21).
[04:31:37] + News From Folding@Home: Welcome to Folding@Home
[04:31:37] Loaded queue successfully.
[04:31:38] + Closed connections
[04:31:43]
Image
burn.intel
 
Posts: 13
Joined: Thu Oct 08, 2009 9:52 pm

Re: Constant EUE Problems

Postby PantherX » Thu Feb 25, 2010 11:36 am

Welcome to the forum burn.intel

From my experience, I can suggest the following:

1) Did you check the forums for the project because some of them can be faulty.

2) If you have overclocked your GPUs, then i suggest that you return them to stock settings and then see if the EUE error occurs.

3) You mentioned that the middle GPU was the first to EUE so i am guessing that it is heat related so i suggest that you manually increas the fan speed of all your GPUs and provided sufficient cooling as GPU can really heat up if you have configured to accept large WUs.

4) If you are worried that it is your motherboard fault, i suggest that you run F@H on your 9800GT without any other cards and monitor its progress, if it EUEs then replace it with our other GPU and monitor. If on all three of your GPU EUE occurs then you can assume it is the motherboard. (you need to switch the GPU to insure that it is the motherboard problem and not your GPU)

5) What drivers are you using? i am using GeForce ION Driver Release 195.62 WHQL (i read that the latest driver causes problems with manual overclocking so i avoided them) and i downloaded and copied the cudart.dll file into the GPU folder. I read somewhere in the forum that it may cause problems so you can download the CUDA drivers and install them instead.

6) Is your power supply working properly as a faulty PSU can cause hardware failure.

7) Which motherboard voltages did you increase? the power load of GPU is independent of the motherboard unless it uses less than 75W which PCI slot supports and in your case, the GPU has its own power-in line so you don't have to tweak the motherboard voltage unless you are overclocking the processor.

hope this is useful to you.
ETA:
Now ↞ Very Soon ↔ Soon ↔ Soon-ish ↔ Not Soon ↠ End Of Time

Welcome To The F@H Support Forum Ӂ Chrome Folding App (Beta) Ӂ Troubleshooting "Bad WUs" Ӂ Troubleshooting Server Connectivity Issues
User avatar
PantherX
Site Moderator
 
Posts: 6321
Joined: Wed Dec 23, 2009 9:33 am

Re: Constant EUE Problems

Postby toTOW » Thu Feb 25, 2010 2:05 pm

I've checked Project: 5782 (Run 7, Clone 75, Gen 40), and someone else completed it fine.

Panther-X's suggestions are good ... I'd add the you should use MemtestG80 to check if your boards are not defective.
Folding@Home beta tester since 2002. Folding Forum moderator since July 2008.

FAH-Addict : latest news, tests and reviews about Folding@Home project.

Image
User avatar
toTOW
Site Moderator
 
Posts: 8776
Joined: Sun Dec 02, 2007 10:38 am
Location: Bordeaux, France

Re: Constant EUE Problems

Postby JimF » Thu Feb 25, 2010 7:01 pm

I have the same Foxconn ELA motherboard and three GT240s dedicated only to Folding. What you need to do is install "dummy plugs" in the VGA port of each card that does not have a monitor attached. The conventional wisdom is that you don't need to do that, but you do. Maybe it is the three card setup. But I have been there, done that.
viewtopic.php?f=50&t=13505

The resistors are nominally 75 ohms; I used 100 ohms with no problem, and 68 or 82 ohms are OK also, but I did not have any luck with 150 ohms, though that may have been just due to a bad connection.
http://www.overclock.net/overclock-net- ... -plug.html
http://www.xtremesystems.org/forums/sho ... p?t=200444
GTX 970 (i5-3550), GTX 980 (i7-3770); Win10 64-bit; FAH 7.4.4
JimF
 
Posts: 490
Joined: Thu Jan 21, 2010 2:03 pm

Re: Constant EUE Problems

Postby burn.intel » Thu Feb 25, 2010 9:29 pm

Yeah I have (had) dummy plugs plugged in when I was running the 3 cards.

@Panther
1) toTow covered the bad WU issue
2) Nope, all are at stock.
3) I'm almost 100% sure its not heat related. I have 3 120mm fans pointing right at the cards and temps across all cards are <70C
4) Currently running the 9800gt alone, and it still EUE's after the first WU. However I haven't tried running the other two 8800GT's alone. Just got busy and never had the time; I'll do that next.
5) Newest ones, 196.xx... I've tried several, back as far as 185.xx
6) PSU is fine, tested each rail with a DMM.
7) All of them lol... Not by much, just a notch or two.

@toTow: I've already run memtestg80 (ran it simultaneously on the 3 cards) and got errors on one of them over 500 iterations. BUT, I ran it again on that GPU only, and got 0 errors over 1000 iterations. Again, to me that points to a motherboard issue (can't handle the 3 cards).
burn.intel
 
Posts: 13
Joined: Thu Oct 08, 2009 9:52 pm

Re: Constant EUE Problems

Postby P5-133XL » Thu Feb 25, 2010 10:32 pm

From your description of the memory test, it may be a motherboard issue and it may be a ram (motherboard) issue, but it screams power issues to me. It may be that you are trying to pull too much power (with all three cards) from the PCI-E bus (75W max for PCI-e 1.0 and 150W max for PCI-e 2.0) or your power supply isn't supplying enough power to power everything but when you have only one card there is power a plenty.

Just to be clear about your description. The card gets the first WU; completes that WU; Sends the data back; gets a second WU and that one and every successive WU EUE's. That's a peculiar failure mode: It indicates that something isn't being reinitialized that is supposed to be so that it isn't starting from scratch.

I'd first make sure that the card being tested is using a PCI-e power plug to eliminate the potential power issues with the PCI-e bus.

I think you need to test each card individually to make sure that it isn't the card. Just change one variable at a time, so keep everything the same. So put the card(s) in the same slot. If the same symptoms occur, then next would be to do the same with a different slot.

I'd also check the system RAM. Effectively, everything goes through RAM so when you get peculiar issues that don't fit normal failure modes then RAM is always a potential suspect. RAM can also be very sensitive to power issues, as well as motherboard timing issues.

P.S. how did you test each rail on your PS. It indicates that there are more than one and on a PC Power and Cooling 750W there is only one rail. As to testing it, you do know that you need to test it with the DMM while it is under load because the voltages can drop significantly as compared to no-load. It may be sometimes inaccurate, but it often helps to use a motherboard monitor and then set alarms for the voltages to test potential power issues while under load.

With the newest Nvidia drivers you shouldn't need to dummy plugs for multiple cards. Just make sure that you extend the monitor to the other cards and use the -gpu # parameter for each additional client beyond -gpu 0. I've also seen situations that require the -forcegpu Nvidia_G80 on cards beyond the first but theoretically that shouldn't be necessary.
Image
P5-133XL
 
Posts: 4034
Joined: Sun Dec 02, 2007 4:36 am
Location: Salem. OR USA

Re: Constant EUE Problems

Postby JimF » Thu Feb 25, 2010 11:17 pm

P5-133XL wrote: I've also seen situations that require the -forcegpu Nvidia_G80 on cards beyond the first but theoretically that shouldn't be necessary.

I forgot to mention that I have to do that as well.
So it is -forcegpu Nvidia_G80 -gpu 0 for the first card set up with the monitor, and then -forcegpu Nvidia_G80 -gpu 1 and finally -forcegpu Nvidia_G80 -gpu 2 for the others. I think it is easiest to insert the cards one at a time and set them up to get the proper order so that you can keep track of them.
JimF
 
Posts: 490
Joined: Thu Jan 21, 2010 2:03 pm

Re: Constant EUE Problems

Postby P5-133XL » Thu Feb 25, 2010 11:25 pm

Yes and no. yes, you have the parameters correct, if you have multiple cards in at the same time. However, if all you are doing is a swapping out a single card then it will always be -gpu 0 regardless of the card. The -gpu # parameter is not picking the slot, but is actually using an enumerated card. If you do -gpu 1 with only one card installed, then the client will immediately fail because it won't see a second card.
P5-133XL
 
Posts: 4034
Joined: Sun Dec 02, 2007 4:36 am
Location: Salem. OR USA

Re: Constant EUE Problems

Postby P5-133XL » Thu Feb 25, 2010 11:34 pm

One other thing, if that machine has ever had an ATI card installed make sure you run the ATI uninstall program. Remnants of the ATI drivers can cause the client to download an ATI WU and attempt to run them on an Nvidia card. That can cause repeated EUE's but the first one didn't because it got an Nvidia WU ...
P5-133XL
 
Posts: 4034
Joined: Sun Dec 02, 2007 4:36 am
Location: Salem. OR USA

Re: Constant EUE Problems

Postby JimF » Fri Feb 26, 2010 12:22 am

Another thing I found was that extending the desktop was neither necessary to get Folding to work (if the parameters are set up as above at any rate), nor is it sufficient to eliminate the need for the dummy plugs. But it did have one subtle benefit: I could see the amount of GPU memory used in GPU-Z only if the desktop was extended to that card. Otherwise, I could see the temperature and other values of the card without it. I think the three-card situation is not a simple extension of the two-card case, at least with this setup with WinXP.
JimF
 
Posts: 490
Joined: Thu Jan 21, 2010 2:03 pm

Re: Constant EUE Problems

Postby JimF » Fri Feb 26, 2010 3:30 am

burn.intel wrote:Foxconn ELA P45 mobo

By the way, you didn't enable "Auto select the PCI clock" or something like that in the motherboard BIOS did you? That will cause EUEs also. In fact, almost any setting different than the defaults should be checked. And don't even THINK of over-clocking.
JimF
 
Posts: 490
Joined: Thu Jan 21, 2010 2:03 pm

Re: Constant EUE Problems

Postby burn.intel » Fri Feb 26, 2010 4:16 am

P5-133XL wrote:From your description of the memory test, it may be a motherboard issue and it may be a ram (motherboard) issue, but it screams power issues to me. It may be that you are trying to pull too much power (with all three cards) from the PCI-E bus (75W max for PCI-e 1.0 and 150W max for PCI-e 2.0) or your power supply isn't supplying enough power to power everything but when you have only one card there is power a plenty.

I don't think it's power related... The person I bought the PSU from crunched also and never had any problems. And if it is power related, that still doesn't explain why my card would EUE when it is the only one plugged in, unless the PSU is pretty screwed up.

P5-133XL wrote:Just to be clear about your description. The card gets the first WU; completes that WU; Sends the data back; gets a second WU and that one and every successive WU EUE's. That's a peculiar failure mode: It indicates that something isn't being reinitialized that is supposed to be so that it isn't starting from scratch.

Yep, that's exactly what happens. Finishes one, then each one after it fails.


P5-133XL wrote:I think you need to test each card individually to make sure that it isn't the card. Just change one variable at a time, so keep everything the same. So put the card(s) in the same slot. If the same symptoms occur, then next would be to do the same with a different slot.

In the process of doing that.

P5-133XL wrote:I'd also check the system RAM. Effectively, everything goes through RAM so when you get peculiar issues that don't fit normal failure modes then RAM is always a potential suspect. RAM can also be very sensitive to power issues, as well as motherboard timing issues.

Good suggestion, I'll run memtest on it...

P5-133XL wrote:P.S. how did you test each rail on your PS. It indicates that there are more than one and on a PC Power and Cooling 750W there is only one rail. As to testing it, you do know that you need to test it with the DMM while it is under load because the voltages can drop significantly as compared to no-load. It may be sometimes inaccurate, but it often helps to use a motherboard monitor and then set alarms for the voltages to test potential power issues while under load.

The PSU has a single +12V rail, however I tested the +3.3V, +5V, and +12V rails all with a DMM. And yes, I know I have to test it at idle and load... The voltages of each rail at idle and load were well within their spec'd limits; which if IIRC should be +-5% of what the rail should be at.

P5-133XL wrote:With the newest Nvidia drivers you shouldn't need to dummy plugs for multiple cards. Just make sure that you extend the monitor to the other cards and use the -gpu # parameter for each additional client beyond -gpu 0. I've also seen situations that require the -forcegpu Nvidia_G80 on cards beyond the first but theoretically that shouldn't be necessary.

Didn't know that about the new drivers. Sweet. I'll try the -force gpu flag too.

P5-133XL wrote:One other thing, if that machine has ever had an ATI card installed make sure you run the ATI uninstall program. Remnants of the ATI drivers can cause the client to download an ATI WU and attempt to run them on an Nvidia card. That can cause repeated EUE's but the first one didn't because it got an Nvidia WU ...

Nope, no ATI drivers ever. It's been nvidia based from day 1.

JimF wrote:By the way, you didn't enable "Auto select the PCI clock" or something like that in the motherboard BIOS did you? That will cause EUEs also. In fact, almost any setting different than the defaults should be checked. And don't even THINK of over-clocking.

Nope, disabled that. And why do you say don't think of OC'ing? What BIOS are you using? The default that came with the board flat out SUCKED, but the newest revision from foxconn is actually quite nice. I've run stock (2.4ghz) and OC'ed (3.15ghz) and that doesn't affect the GPU errors at all. Same EUE problems occur regardless.
burn.intel
 
Posts: 13
Joined: Thu Oct 08, 2009 9:52 pm

Re: Constant EUE Problems

Postby PantherX » Fri Feb 26, 2010 7:16 am

If you want to check your system for stability then i recommend the following:

CPU - IntelBurnTest 2.4 (http://www.softpedia.com/get/System/Benchmarks/IntelBurnTest.shtml)
it will stress the CPU to its limits and in the latest version, will stress the RAM too. The plus side is that it takes around 8 minutes to find out if the system is stable or not.

GPU - FurMark 1.8.0 (http://www.softpedia.com/get/System/Benchmarks/FurMark.shtml)
It will stress the GPU and it works with multiple GPUs and is extremely reliable.

Your above post states that you have OC your CPU to 3.15, in my motherboard the RAM too would automatically be overclocked so it might be possible that the RAM cant handle this increased frequency at high load levels. My system is Intel Core 2 Duo E8190 @ 3.00 and RAM @ 450MHz (effective is 900MHz). Although I can overclock my CPU higher, my RAM can't handle it. So i think that you many need to fine-tune your system a little more.

I too had a similar problem on my laptop and in the end the only solution that worked for me was a complete format of the system with a fresh copy of windows and the problem simply vanished. Hope you can solve it soon.
User avatar
PantherX
Site Moderator
 
Posts: 6321
Joined: Wed Dec 23, 2009 9:33 am

Re: Constant EUE Problems

Postby toTOW » Fri Feb 26, 2010 9:33 am

burn.intel wrote:@toTow: I've already run memtestg80 (ran it simultaneously on the 3 cards) and got errors on one of them over 500 iterations. BUT, I ran it again on that GPU only, and got 0 errors over 1000 iterations. Again, to me that points to a motherboard issue (can't handle the 3 cards).


Is it the same board that keeps failing on FAH when used alone in the machine ? If yes, I strongly advise you to get the card replaced.
User avatar
toTOW
Site Moderator
 
Posts: 8776
Joined: Sun Dec 02, 2007 10:38 am
Location: Bordeaux, France

Re: Constant EUE Problems

Postby burn.intel » Fri Feb 26, 2010 7:41 pm

Panther-X wrote:If you want to check your system for stability then i recommend the following:

CPU - IntelBurnTest 2.4 (http://www.softpedia.com/get/System/Benchmarks/IntelBurnTest.shtml)
it will stress the CPU to its limits and in the latest version, will stress the RAM too. The plus side is that it takes around 8 minutes to find out if the system is stable or not.

GPU - FurMark 1.8.0 (http://www.softpedia.com/get/System/Benchmarks/FurMark.shtml)
It will stress the GPU and it works with multiple GPUs and is extremely reliable.

Your above post states that you have OC your CPU to 3.15, in my motherboard the RAM too would automatically be overclocked so it might be possible that the RAM cant handle this increased frequency at high load levels. My system is Intel Core 2 Duo E8190 @ 3.00 and RAM @ 450MHz (effective is 900MHz). Although I can overclock my CPU higher, my RAM can't handle it. So i think that you many need to fine-tune your system a little more.

I too had a similar problem on my laptop and in the end the only solution that worked for me was a complete format of the system with a fresh copy of windows and the problem simply vanished. Hope you can solve it soon.

I've run Prime95 and LinX (basically the same as IBT, they both use linpack) and they both run fine. Ran furmark on the only card I have in at the moment and got no issues.

As for the RAM, I have the FSB:RAM frequency ratio set to something low, so the actual effective rate is 700mhz (350 in bios).

toTOW wrote:
burn.intel wrote:@toTow: I've already run memtestg80 (ran it simultaneously on the 3 cards) and got errors on one of them over 500 iterations. BUT, I ran it again on that GPU only, and got 0 errors over 1000 iterations. Again, to me that points to a motherboard issue (can't handle the 3 cards).


Is it the same board that keeps failing on FAH when used alone in the machine? If yes, I strongly advise you to get the card replaced.

No, different card.
burn.intel
 
Posts: 13
Joined: Thu Oct 08, 2009 9:52 pm

Next

Return to NVIDIA specific issues

Who is online

Users browsing this forum: No registered users and 2 guests

cron