Associate
- Joined
- 7 Dec 2008
- Posts
- 79
- Location
- Newcastle, UK
Apologies for the length of the post; but pease bear with it and help if possible.. 
Spec:
Intel Corei7 920 C0 ~2.67Ghz @ 3.2Ghz
Cooler: Thermaltake Ultra-120 TruBlack with 1 x 120mm FAN.
G-Skill 6GB (3x2GB) DDR3-1600 @1604Mhz
Asus P6T Deluxe Motherboard
2 x GeForce GTX 460's in SLI
2 x Western Digital Black 512GB (Performance RAID)
Zalan Ultra Quiet Dual Pipe Cooled 850W PSU
The Problem
If I play a game; or run any sort of graphical stress test my computer tends to freeze roughly after 2 hours. Afterwards I have to hard reset the PC.
However after this has happened the computer starts freezing after just getting into windows; litterally seconds after logging in; or sometimes during the screen transistion from login to desktop and it can take a few resets for it to settle down and stop freezing. If I then don't then run any grpahical applications (just standard windows browsing) my computer stays stable for as long as required.
History
You'll see why I'm giving you the drawn out history in a moment..
Early 2009 - PC Built
Ran fine; CPU stock speeds; ATI Radeon 4870. However I've always get a noise problem when viewing certain large 2D animations, menus or scrolling up and down websites. I get a strange squeeking noise from my computer; it sounds somewhat electrical; which has always made me suspicous of my PSU; however it never caused me an issue. Its not coming from the speakers..
Mid August 2010 - Purchased 2 x GTX 460's & CPU Overclock
Graphics cards came first; when I first installed them I got a problem where there was a strange blue flashing interference effect when running in SLI mode. Both cards didn't have this problem when running individually or when the SLI bridge was removed. Video Here >
However after re-sitting the SLI bridge 3 or 4 times the problem seemed to disappear. Once that was sorted I overclocked my Core i7 as people claimed the stock speed was damaging my SLI scalablility. COnsidering how GPU dependant games are I doubted this but I thought "what the hell". I spent a number of weeks fiddling but settled on 3.2Ghz as that was the best balance of performance and acceptable temperatures.
What I've tried
For starters I've not ran a game for more than 30 mins since overclocking or putting in my new GTX-460s; I only ran them long enough to see what performance differences I was getting. So honestly; either the overclock or the graphics cards could be causing the problem. Basically from the history of my machine my suspicions are as follows:
Unstable CPU (Either Overclock or Temperatures)
Faulty Graphics Card\SLI Bridge? (Either Faulty or Overheating)
Faulty PSU (Due to strange noises)
with this in mind; I've tried the following things:
1) Unstable CPU? - CPU Tests
Removed my CPU overclock immediately after the PC locked (during some gameplay). It still locked a number of times on bootup (at windows login) before the system settled. Overclock not the issue as problem still remained after removal. Overclock re-instated back to 3.2Ghz.
Checked stability and temperatures of CPU with Realtemp and Prime95 -->
Ran Prime95 (Blend) for 13 Hours; all worker threads completed with zero errors; CPU temperatures:- ||Under Load:|| 79c (Max) \ 75c (Avg) ||Idle:|| 45c (Avg)
Result: Not the CPU causing my problems.
1) Problematic GPUs? - GPU Tests
Cards are running stock speeds; no overclocks and never have.
First test was to ascertain stability of current SLI setup and possible temperature problems. I ran FurMark (Multi-GPU) Stability test; it managed 2 Hours 6 mins until the video locked. This is roughly in line with my experience when playing games. Only the video locks; music I was playing in the background continued; as did HDD activity. Temperatures of the cards were acceptable:
GeForce GTX 460 (1): (80c) Max.
GeForce GTX 460 (2): (74c) Max.
This is a hard video lock; no amount of Ctrl-Alt-Delete or removing\plugging back in monitors makes any difference.
Before condemning SLI; I then decided to test each card individually for stability.
GeForce GTX 460 (1): 6 Hours, 8 Mins (75c) Success (I stopped it at 6 hours)
GeForce GTX 460 (2): 3 Hours, 19 Mins (73c) Locked! Failed. (Music playlist continued to play in the background; CTRL-ALT-DLT\ALT-TAB does nothing)
at this point I'm assuming the fault must lie with either the second graphics card or the second PCI-E port on the motherboard. Upon Removing the second graphics card I noticed some dust on the graphic card's contacts. Now whether these were attracted to the bottom of the card upon removal (thus not the problem) or if they were in the motherboard's PCI-E slot the whole time is unknown. To rule this out I used compressed air to clean the PCI-E slot and made sure the card's contacts were dust free before reseating the graphics card.
I then commenced the Furmark test on the second card (again)....
GeForce GTX 460 (2): 0 Hours, 8 Mins (71c) Locked! Failed. (Music playlist continued to play in the background; CTRL-ALT-DLT\ALT-TAB does nothing)
Okay; it wasn't the dust. To further try and rule out the second PCI-E slot I swapped the cards around and ran another test on the suspected card.
GeForce GTX 460 (2): 0 Hours, 8 Mins (70c) Locked! Failed. (Music playlist continued to play in the background; CTRL-ALT-DLT\ALT-TAB does nothing)
Definitely looking like the second card is faulty; however just to be sure the first card's 6 hour successful test wasn't a fluke; I gave it another stability test but obviously now in the other PCI-E slot.
GeForce GTX 460 (1): 11 hours, 24 mins (73c) Success (11 Hours is more than enough testing).
Result
The other card; as expected ran perfectly fine in the other slot as well; this time I ran it overnight for 11 hours; so it definitely appears to be the second card that’s at fault as it won't run for longer than 3 hours (usually much less) in either PCI-E slot.
I'm within my first 28 days of purchase so I should get an RMA/Swap out from overclockers hassle free? Unless someone’s thinks I'm missing something obvious in my testing? Everyone agree with my final result?
Spec:
Intel Corei7 920 C0 ~2.67Ghz @ 3.2Ghz
Cooler: Thermaltake Ultra-120 TruBlack with 1 x 120mm FAN.
G-Skill 6GB (3x2GB) DDR3-1600 @1604Mhz
Asus P6T Deluxe Motherboard
2 x GeForce GTX 460's in SLI
2 x Western Digital Black 512GB (Performance RAID)
Zalan Ultra Quiet Dual Pipe Cooled 850W PSU
The Problem
If I play a game; or run any sort of graphical stress test my computer tends to freeze roughly after 2 hours. Afterwards I have to hard reset the PC.
However after this has happened the computer starts freezing after just getting into windows; litterally seconds after logging in; or sometimes during the screen transistion from login to desktop and it can take a few resets for it to settle down and stop freezing. If I then don't then run any grpahical applications (just standard windows browsing) my computer stays stable for as long as required.
History
You'll see why I'm giving you the drawn out history in a moment..
Early 2009 - PC Built
Ran fine; CPU stock speeds; ATI Radeon 4870. However I've always get a noise problem when viewing certain large 2D animations, menus or scrolling up and down websites. I get a strange squeeking noise from my computer; it sounds somewhat electrical; which has always made me suspicous of my PSU; however it never caused me an issue. Its not coming from the speakers..
Mid August 2010 - Purchased 2 x GTX 460's & CPU Overclock
Graphics cards came first; when I first installed them I got a problem where there was a strange blue flashing interference effect when running in SLI mode. Both cards didn't have this problem when running individually or when the SLI bridge was removed. Video Here >
However after re-sitting the SLI bridge 3 or 4 times the problem seemed to disappear. Once that was sorted I overclocked my Core i7 as people claimed the stock speed was damaging my SLI scalablility. COnsidering how GPU dependant games are I doubted this but I thought "what the hell". I spent a number of weeks fiddling but settled on 3.2Ghz as that was the best balance of performance and acceptable temperatures.
What I've tried
For starters I've not ran a game for more than 30 mins since overclocking or putting in my new GTX-460s; I only ran them long enough to see what performance differences I was getting. So honestly; either the overclock or the graphics cards could be causing the problem. Basically from the history of my machine my suspicions are as follows:
Unstable CPU (Either Overclock or Temperatures)
Faulty Graphics Card\SLI Bridge? (Either Faulty or Overheating)
Faulty PSU (Due to strange noises)
with this in mind; I've tried the following things:
1) Unstable CPU? - CPU Tests
Removed my CPU overclock immediately after the PC locked (during some gameplay). It still locked a number of times on bootup (at windows login) before the system settled. Overclock not the issue as problem still remained after removal. Overclock re-instated back to 3.2Ghz.
Checked stability and temperatures of CPU with Realtemp and Prime95 -->
Ran Prime95 (Blend) for 13 Hours; all worker threads completed with zero errors; CPU temperatures:- ||Under Load:|| 79c (Max) \ 75c (Avg) ||Idle:|| 45c (Avg)
Result: Not the CPU causing my problems.
1) Problematic GPUs? - GPU Tests
Cards are running stock speeds; no overclocks and never have.
First test was to ascertain stability of current SLI setup and possible temperature problems. I ran FurMark (Multi-GPU) Stability test; it managed 2 Hours 6 mins until the video locked. This is roughly in line with my experience when playing games. Only the video locks; music I was playing in the background continued; as did HDD activity. Temperatures of the cards were acceptable:
GeForce GTX 460 (1): (80c) Max.
GeForce GTX 460 (2): (74c) Max.
This is a hard video lock; no amount of Ctrl-Alt-Delete or removing\plugging back in monitors makes any difference.
Before condemning SLI; I then decided to test each card individually for stability.
GeForce GTX 460 (1): 6 Hours, 8 Mins (75c) Success (I stopped it at 6 hours)
GeForce GTX 460 (2): 3 Hours, 19 Mins (73c) Locked! Failed. (Music playlist continued to play in the background; CTRL-ALT-DLT\ALT-TAB does nothing)
at this point I'm assuming the fault must lie with either the second graphics card or the second PCI-E port on the motherboard. Upon Removing the second graphics card I noticed some dust on the graphic card's contacts. Now whether these were attracted to the bottom of the card upon removal (thus not the problem) or if they were in the motherboard's PCI-E slot the whole time is unknown. To rule this out I used compressed air to clean the PCI-E slot and made sure the card's contacts were dust free before reseating the graphics card.
I then commenced the Furmark test on the second card (again)....
GeForce GTX 460 (2): 0 Hours, 8 Mins (71c) Locked! Failed. (Music playlist continued to play in the background; CTRL-ALT-DLT\ALT-TAB does nothing)
Okay; it wasn't the dust. To further try and rule out the second PCI-E slot I swapped the cards around and ran another test on the suspected card.
GeForce GTX 460 (2): 0 Hours, 8 Mins (70c) Locked! Failed. (Music playlist continued to play in the background; CTRL-ALT-DLT\ALT-TAB does nothing)
Definitely looking like the second card is faulty; however just to be sure the first card's 6 hour successful test wasn't a fluke; I gave it another stability test but obviously now in the other PCI-E slot.
GeForce GTX 460 (1): 11 hours, 24 mins (73c) Success (11 Hours is more than enough testing).
Result
The other card; as expected ran perfectly fine in the other slot as well; this time I ran it overnight for 11 hours; so it definitely appears to be the second card that’s at fault as it won't run for longer than 3 hours (usually much less) in either PCI-E slot.
I'm within my first 28 days of purchase so I should get an RMA/Swap out from overclockers hassle free? Unless someone’s thinks I'm missing something obvious in my testing? Everyone agree with my final result?
. Still, I cant argue with the results of your testing and it does indeed sound like the 2nd GPU aint up to it.