Significant display artifacts + System Crashing! Please Help! (videos)

[thadman]

I recently built myself a custom computer for multiphysics simulations and light gaming.

System Specs

-Windows 7 Ultimate 64 bit

-Intel Core i7 930 (NO OC, Corsair H50 cooler)

-Gigabyte X58A-UD3R rev1.0

-EVGA nVidia GTX470 (NO OC, Stock cooler)

-(6x) Corsair TR3X6G1600C8D Dominator DDR3 2GB (12GB) in triple channel mode (modules are rated at 8-8-8-24-2, 1600MHz, 1.65V)

-Intel X25-M 80GB SSD (OS drive)

-(2x) Western Digital Caviar Black 1.5TB HDD

-Corsair 800D case

-Corsair TX950W power supply

-Corsair H50 CPU cooler

(BIOS set to load Optimized Defaults)

The system is highly unstable and significant display artifacts are present. In addition, the system will always crash during the introduction scene of Crysis (display=1920x1200). (I've been trying to use it to benchmark my system). I have tried very high settings, high settings, medium settings, and low settings at 0x,4x, and 8x AA. It doesn't seem to make a difference, the system always crashes during that scene. Also, the system consistently fails to boot. Whenever the system crashes or windows fails to boot, usually the GPU fan kicks into overdrive (insanely loud at this point compared to very quiet during normal operation).

I have been using the application "GPU Observer" to monitor the GTX470. The card operates around 45*C at idle and 55-65*C under load (I have been using Starcraft 2 for testing stability since Crysis always crashes). For reference, CPU is around 35*C with the H50 cooler. I once found the GPU at 83*C after playing SC2 for an hour (Corsair 800D with side on). I thought that may have been a problem with Crysis, so I turned off the computer, removed the side of my 800D and placed a LARGE box fan immediately in front of the system components. Idle temperatures dropped by about 10-15*C all around. After 6 hours with the side off and the fan blowing, the GPU was sitting around ~45*C. I loaded Crysis and once again, the system crashed. I rebooted the computer immediately after Crysis crashed. According to "GPU Observer" , the GTX470 core was at 59*C and the PCB was at 49*C. I don't feel like the GPU temp would've changed drastically in the time between system crash and temp measurement after reboot (I'm booting off of an Intel X25-M solid state drive).

I have reinstalled Windows 7 4 times. I have used the GTX470 drivers available through Guru3d, the drivers that came with the card (CD), and EVGAs drivers. I have been using the application "Driver Sweeper" to remove the old drivers. The system still fails. It doesn't seem to matter what drivers are used.

The system fails to load the application "EVGA OC Scanner". An error results.

The title of the error window reads:

"EVGA OC Scanner ERROR"

The text within the error window reads:

"Could not initialize ZoomGPU. EVGA OC Scanner startup failed. Bye!"

I loaded Sandra SiSoftware. My system crashed during the GPGPU test. I decided to run the environment sensors test

CPU 1 DC line = .93V (min .91V, avg .94V, max 1.2V)
CPU 2 Aux DC line = 1.52V (min 1.52V, avg 1.52V, max 1.52V)
+3.3V DC Line = 3.31V (min 3.31V, avg 3.31V, max 3.31V)
+5V DC Line = 5.00V (min 5.00V, avg 5.00V, max 5.00V)
+12V DC Line = 1.75V (min 1.75V, avg 1.84V, max 3.99V) = ?

However, immediately after I got this reading, I rebooted the computer and entered the BIOS. According to the X58A-UD3R, my 12V rail is at 12.365V and my 5V rail is at 4.99V.

I took videos of the system while it was exhibiting this behavior and uploaded them to youtube. Hopefully, these videos will convey the problem accurately.


http://www.youtube.com/watch?v=G35BD3QS1Po

^I've uploaded a video of the system booting. This was the fourth boot attempt and was finally successful (the system crashed at windows start-up the first three times).

(There is a few seconds of darkness at the beginning of the video, I do not enter the BIOS. This is a standard boot.) The display artifacts should be readily apparent.

At the conclusion of this video, I use the application "Display Sweeper" to remove the display drivers (I did not take a video of this).



Computer Problems Part 1

Prior to this video, I used the freeware application "driver sweeper" to remove nVidia's display drivers. The video shows the installation of EVGAs Windows 7 64 bit GTX470 display drivers and the beginning of reboot. You should be able to see the display artifacts clearly.

http://www.youtube.com/watch?v=HsCifaeyDL8



Computer Problems Part 2

This video immediately follows "Computer problems part 1". The video begins with a system restart after EVGAs nVidia GTX470 display drivers have been installed. Upon reboot, I enter the BIOS and show the system settings.

http://www.youtube.com/watch?v=J7ojBPzHM7s



Computer Problems Part 3

This video immediately follows "Computer Problems Part 2". The video shows the display artifacts on the BIOS enter screen and the system attempting to boot into Windows 7. The system freezes upon startup. The GPU fan kicks into overdrive (insanely loud).

http://www.youtube.com/watch?v=oqZ5qZfUczE



Computer Problems Part 4

This video immediately follows "Computer Problems Part 3". It begins with the frozen Win7 loading screen. I physically turn the computer off by pressing the power button and then turn it on by pressing the power button. Upon startup, I enter the BIOS to check the PC Health Status (rail voltage, temp). I then exit the BIOS and boot into Win7 Safe Mode with Networking. The system boots successfully and I enter "my computer" to view the display drivers. After checking the drivers, I restart the system.

http://www.youtube.com/watch?v=0rMiL_ayp_Y



Computer Problems Part 5

This Video immediately follows "Computer Problems Part 4". The video begins with the system restarting from Safe Mode with Networking. Upon startup, I enter the BIOS to check the PC Health Status (rail voltage, temp). I then exit the BIOS and allow the system to boot normally. The system freezes attempting to boot Win7. About when the system crashes, you should be able to hear the GPU fan kick into overdrive. I physically turn off the system by pressing the power button and restart it by pressing the power button. I do not enter the BIOS, I allow the system to boot normally. During startup, Windows recognizes it failed to boot. I enter Safe Mode. I attempt to enter the application "EVGA OC Scanner". An error results.

The title of the error window reads:

"EVGA OC Scanner ERROR"

The text within the error window reads:

"Could not initialize ZoomGPU. EVGA OC Scanner startup failed. Bye!"

I then enter my computer and search for the application "Driver Sweeper". I load "Driver Sweeper" and click the box next to "nVidia display". I then click clean and remove the display drivers. After the display drivers have been removed, I restart the system.

http://www.youtube.com/watch?v=GHSfmAmqqio



Computer Problems Part 6

This Video immediately follows "Computer Problems Part 5". During startup, I enter the BIOS to check the PC Health Status (rail voltage, temp). I then exit the BIOS and allow the system to boot up normally. Windows loads (finally). However, there are still significant display artifacts. As soon as Windows loads, I check the application "GPU Observer". The program doesn't recognize the card (GPU not found, 0*C GPU, 0*C PCB). I then attempt to load the application "EVGA OC Scanner". An error results.

The title of the error window reads:

"EVGA OC Scanner ERROR"

The text within the error window reads:

"Could not initialize ZoomGPU. EVGA OC Scanner startup failed. Bye!"

http://www.youtube.com/watch?v=QVGzLG_rTXc



Any ideas? I will be uploading more videos (loading Crysis, SC2, benchmarking applications, etc). Could you recommend some benchmarking applications that could properly elucidate the problem. I'm familiar with Sandra SiSoftware, Prime95, and Memsoft.

 

[thadman]

I completed the Sandra SiSoftware burn-in. Here are the results. (GPU drivers are not installed).

I think my PSU might be shot, any thoughts? It's interesting that my motherboard thinks the 12V rail is at 12.365V, while Sandra observed it below 3V.

 

[MagnusTheBrewer]

You left out the current barometric pressure and what you had for breakfast.
All of your problems seem video related but you think it's your PSU?
What monitor are you using and what resolution are you displaying?

 

[thadman]

I completed the Sandra SiSoftware Analysis and Advice. Here are the results. (GPU drivers are not installed)

I currently have CoreTemp running. Here's a screen shot.

 

[thadman]

You left out the current barometric pressure and what you had for breakfast.
All of your problems seem video related but you think it's your PSU?
What monitor are you using and what resolution are you displaying?

Dell U2410 @ 1920x1200

 

[thadman]

I also have GPU Observer running. Here's a screen shot.

I assume GPU Observer is not recognizing the GTX470, because no GPU drivers are installed.

 

[thadman]

I completed the Sandra SiSoftware Performance Index. Here are the results. (GPU drivers are not installed)

 

[MagnusTheBrewer]

Do you have both 6-pin PCI-E power supplies plugged in to the GPU?

 

[thadman]

Also, Windows 7 is not recognizing my 1.5TB Western Digital Caviar Black HDDs. However, they are recognized in the BIOS.

Here are some screen shots.

My Computer (showing only the Intel X25-M 80GB SSD)

Devices and Printers screen shots.

It appears the HDDs are recognized as hardware :hmm:

The HDDs are apparently working properly.

HDD volume data (all blank)

HDD driver details.

I think this discrepancy might be presenting itself because when the Gigabyte X58A-UD3R is set to optimized defaults, the HD controllers are set to IDE mode instead of AHCI. Any thoughts?

 

[thadman]

Do you have both 6-pin PCI-E power supplies plugged in to the GPU?

I believe so.

 

[thadman]

Any thoughts on why the 12V voltage reading in the BIOS would be significantly different from the voltage reading conducted during the Sandra SiSoftware Burn-In Test?

The BIOS is showing 12.365V for the +12V rail,

while Sandra is showing a dynamic voltage of 2.00V to 2.81V (2.40V avg)

I have absolutely no idea. I find it peculiar that the 12V rail readings differ significantly, while the 5V, 3.3V, and 1.52V rails are in quite close agreement.

I just assumed it was a PSU problem because of the significant voltage drop. What other hardware components could be causing that? (GFX?)

 

[RebateMonger]

Re: your disks.

If these are new disks, did you initialize them and partition them and format the partitions in Disk Management?

 

[MagnusTheBrewer]

I don't know how the Sandra software takes it's readings but, it looks like it is measuring the variance not, the total voltage.

 

[thadman]

I don't know how the Sandra software takes it's readings but, it looks like it is measuring the variance not, the total voltage.

I'm pretty sure its measuring the voltage.

The Orange Rail (+5V) is almost exactly at +5V on the graph and appears invariant. Likewise, the darker Green Rail (+3.3V) is almost exactly at +3.3V on the graph and the Red Rail (+1.52V) is almost exactly at +1.5V. They both appear invariant. However, the Pink Rail (+12V) begins at +2.8V and linearly drops to +2V. Also, the Aqua Rail (-12V) is at -.9V and appears to be invariant.

Maybe I'm reading it wrong...

 

[MagnusTheBrewer]

I'm pretty sure its measuring the voltage.

The Orange Rail (+5V) is almost exactly at +5V on the graph and appears invariant. Likewise, the darker Green Rail (+3.3V) is almost exactly at +3.3V on the graph and the Red Rail (+1.52V) is almost exactly at +1.5V. They both appear invariant. However, the Pink Rail (+12V) begins at +2.8V and linearly drops to +2V. Also, the Aqua Rail (-12V) is at -.9V and appears to be invariant.

Maybe I'm reading it wrong...

I guess I can't believe that anything would be running at only ~2 V.

 

[thadman]

I made a mistake earlier, I do not have Corsair XMS3. I have two kits of Corsair Dominator TR3X6G1600C8D, which is rated for 8-8-8-24-2 @ 1600MHz @ 1.65V.

I cleared my cmos and removed all of my DDR3, except for the module corresponding to DDR3_1 in Channel 0. I went into my BIOS and changed the default RAM timings from 8-8-8-20-2 to 8-8-8-24-2. I also upped the multiplier to 12 (133*12=~1600MHz), increased the DDR voltage to 1.640V, and increased the QPI voltage to 1.335V.

All other system settings are set to default.

I booted into the desktop without any display artifacts. I then attempted to run Crysis (1920x1200, High settings, 0x AA) and the system crashes. Display artifacts were present (even in the BIOS after reboot). However, if I enter the BIOS right after reboot and exit, the display artifacts are gone.

I cleared my cmos and used all the same settings, except I relaxed my RAM timings to 9-9-9-24-2. System crashes.

I then tried 8-8-8-24-2 timings with 8x and 10x memory multipliers. System crashes.

I then tried 9-9-9-24-2 timings with 8x and 10x memory multipliers. System crashes.

Here's a very short video showing the display artifacts in the BIOS after system crash. The artifacts are not static.

http://www.youtube.com/watch?v=zoiCZ73FpqA

I will get memtest ASAP

 

[thadman]

The desktop just crashed. The screen became "checkered". I wasn't doing anything intensive. I transferred a ~40,000kb video of Crysis crashing from my Droid to my uploads folder. I began watching the video and the system crashed after 5 seconds.

I had to physically turn off the system since it was unresponsive. The system then crashed trying to load Windows 7.

Here are photos of the system attempting to load Windows 7 and ultimately crashing.

After the system crashed attempting to load Windows 7, I physically turned off the system and rebooted. Upon rebooting, there were no display artifacts and the system loaded Windows 7 without a problem.

All system settings are stock, except the RAM.

I have one Corsair TR3X6G1600C8D Dominator module in the DDR3_1 Channel 0 slot running at 1066MHz (base clock = 133MHz, memory multiplier = 8) with 9-9-9-24-2 timings. DDR voltage is set to 1.640V and QPI voltage is set to 1.335V.

Here is the video of Crysis crashing.

http://www.youtube.com/watch?v=fx121sGRFJQ

The errors at (0:15 - 0:22) and at :29 are from my Droid, not the system. The glitching at 1:00 - 1:03 is from my system. After Crysis crashes, the screen becomes corrupted. There are squares all over the screen. The color does not appear to be effected significantly. However, the video may not show these effects because it is very low resolution (max 480p, taken via droid and uploaded to youtube) compared to my monitor which is at 1920x1200.

Here is a video of me entering the BIOS after Crysis crashed.

http://www.youtube.com/watch?v=YpLMdZ57E5g

Notice the pixelated lines running across the screen where you enter the BIOS (they might not show up in the youtube video due to the poor capture quality of my droid and the small file size of the youtube video). Also notice the changing squares (multicolor) at the bottom center of the screen within the BIOS. I have no idea why there are flashing symbols (letters and sometimes numbers) in those squares.