Computer Lockups and Odd Graphics Rendering

[Tidane]

I posted once before about my computer randomly locking up and requiring a hard reset, but was unable to solve that update. After changing operating systems and upgrading drivers, the issue was still occurring. It happened again a few days ago and I noticed that the clock in the BIOS had been reset (the day was correct, but the year was wrong). I also noticed that the BIOS screen looked different. Today, while playing a game, part of the screen disappeared (about halfway up on the left side, extending diagonally about 3/4 of the way across the top) stopped rendering in the game. The rest of the screen was rendering ok in some places, but incorrectly in others. When I closed the game, the entire screen started to render awkwardly. Some of the desktop looked like the text from the beginning of the Star Wars movies (that perspective) and other areas were completely unreadable.

My first thought was that the graphics card was overheating, but the artifacts weren't like I'm used to seeing for overheating graphics cards. My second thought was that the driver was the issue. I'm going to go to an older version of the driver to see if that fixes the issue, but I'm wondering if there's any good ways to diagnose overheating/faulty graphics cards?

[Olo401]

Need your computer specs..

[Tidane]

Oh balls, I thought I linked to the old thread with the specs.

OS
Windows 7 64-Bit

Mainboard
Motherboard: Asus P5N32-E SLI, nForce 680i SLI chipset
BIOS: ASUS P5N32-E SLI ACPI BIOS Revision 1103 dated 06/04/2007.

CPU
Intel Core 2 Extreme QX6850 (Quad Core), Socket 775 LGA, 65nm chipset with a core voltage of 1.264 volts.
Core Speed: 3.00GHz
Multiplier: 9x
Bus Speed: 333.3 MHz
Rated FSB: 1333.3 MHz

Memory
Size: 1024MB x 2
Bandwidth: PC2-6400 (400 MHz)
Manufacturer: Corsair
Type: DDR2
Channels: Dual
DRAM Frequency: 400MHz
FSB:DRAM: 5:6
CAS# Latency: 5 clocks
RAS# to CAS# Delay: 5 clocks
RAS# Precharge: 5 clocks
Cycle Time: 18 clocks
Bank Cylce Time: 22 clocks
Command Rate: 2T
Voltage: 1.8

(Side note, does it matter how the memory is installed? I currently have them in non-sequential slots and I planned on adding two more 1GB sticks, but I wanted to fix all the problems before I started upgrading.)

Graphics
Card: NVIDIA GeForce 8800GTX
Drivers: Version 196.21
Size: 768MB
Type: GDDR3
Bus Width: 384 bits
90 nm chip

Sound
Sound Blaster Audigy SE

That's all the relevant information I could think to include, but if anything is missing, I'll be glad to add it. Additionally, I did a bit more research after I started the post last night and found that there have been some reports of 196.21 drivers causing issues with the game I was playing, but nothing like what I experienced. If it doesn't happen again with older drivers (I won't have time to test it until later this night), I'm back to the only issue being the computer locking up.

[Lhyet]

How old is the machine?

Has the clock had any more issues since the year being changed? If the clock keeps resetting itself then the CMOS battery could need replacing, but that shouldn't affect the graphics.

The driver could be what's causing the issue and rolling back to a previous version is indeed the first step. If you're still having problems then I do wonder if it could be a motherboard problem =/

[Tidane]

I rolled back to an older version and stressed the machine a bit. The graphics never went wonky, but it did lock up (I didn't get any decent information from it, but I enabled what I hope will be better logging). I'm also monitoring and logging CPU Usage, Commit, harddrive access, and actual RAM usage. Hopefully, one of these will spike when it crashes and that will help diagnose the issue.

The clock reset that one time and has been fine since then (even after the computer locked up). I thought the battery might be going, but I assume that it would keep losing time.

As for the age of the computer, it's a little over 2 years old, but it has been locking up like this for a very long time.

[Lhyet]

Yeah, if the clock only messed up once and not again after that, the battery is probably fine.
But how did the BIOS screen change? And is it still like that? Have you tried updating the BIOS' firmware?

How is the computer locking up exactly? Freezing and then forcing you to do the hard reset, or anything fancy?

I used to use a program call Everest by a company called Lavalys that could monitor the temperatures of different components, which is really good for finding out if something like the graphics card is overheating. I'm not sure if they still have the free version for download, but I'm sure it could be found somewhere. It's a great tool.

[Tidane]

Whenever the computer locks up, the screen will freeze with whatever was on it (no input, no updates) and a jarring repeating sound will come out of the speakers (I've never had it freeze when the sound was off, so I'm unsure what, exactly, that sound is). I've been trying to get it to lock up again, but I haven't been able to get it to since it locked up last night.

I'm not sure if the BIOS screen changing is actually a symptom of anything. After one of the lockups, I changed a setting to test all the components on boot up, then changed it back. I may have clicked another setting, but I'm not sure. All I know is that the BIOS boots faster and displays more information when booting. I wanted to make sure that I left nothing that happened without me explicitly remembering the change, just in case. I thought about updating the BIOS, but finding BIOS drivers is rather difficult.

I'll see if I can find that information. Right now, I'm focused on memory related things using Microsoft's System Information tool. If it locks up and this doesn't help, I'm going to move onto video card/overheating.

[Tidane]

I finally got the computer to lock up (it only took two days, instead of the usual every night). I haven't had time to look through the information and see if any of it stands out, but all of the information seems within normal limits (except the Commit Charge). At the time of the crash, here is the last reported information:

CPU Usage
54.6%

Commit
Current: 2,711,244 KB
Limit: 5,598,056 KB
Peak: 5,840,608 KB

I/O Bytes
192.9KB

RAM
Total: 2,095,672 KB
Available: 455,744 KB
System Cache: 166,716

Kernal Memory
Paged Physical: 39,336
Paged Virtual: 159,388

One thing to not is the commit charge. I raised the amount of memory available to the page file because it was peaking at about 4GB before, which was a little more than I had. I upped the page file and the peak value went up (higher than the limit, again).

I'm going to look for that program you suggested and try to see what the card temperature is at when it crashes. What are the normal operating limits for an 8800GTX? I've had some people tell me 85C is normal for this card, but others say that if it is over 75C you probably are having cooling issues.

Edit: I just remembered something else. Sometimes, after the computer locks up and I do a hard reset, the computer will reboot and install a driver for a floppy disk drive. I don't have a floppy drive. If I restart it and go into the BIOS, the setting for legacy floppy will be switched on. I'm not sure if that's causes by the same thing that causes it to lock up or if that's causes by the hard reset. Since it doesn't happen everytime, it feels more like a side-effect rather than a symptom of the issue, but if I crash with normal temperatures, I'm going to try to look up BIOS information. Does anyone have suggestions for updating a BIOS?

[Killgannon]

sounds like a BIOS issue really, i would look into updating it... check ASUS's website they should have an ISO so you can burn a boot CD to do it with, then if that doesnt fix your issue, your motherboard maybe dying

if your computer is still plugged in, like physically to the power, your CMOS battery wont do anything, those only take effect when its COMPLETELY separated from power (it used to be different, but these days its not like that) so you can rule that out.

try updating the BIOS like I said, after that i would start checking your motherboard. the 680i is a nice chipset (i have the 780i) but everything is prone to death in technology

[Tidane]

I started looking around for BIOS updates, but the ASUS website seems to lack anything for the P5N32-E SLI and Windows 7 64bit. I did a bit of searching and it appears (maybe) that they're not supporting Windows 7 with this motherboard, but some people do talk about using Windows 7 and having a later BIOS revision than I, so I'm not entirely sure. It may be that ASUS is grouping Win7 and Vista support, but my last BIOS flashing experience has kept me from making those assumptions.

While I was looking through this, I found something that mentioned that the default voltage for memory on this board was a bit low. It said the default was 1.8V and suggested values online were between 1.9V and 2.1V. I'm a computer programmer, but my knowledge of RAM timings, command rate, and voltage is spotty at best. Is it possible for a low RAM voltage to cause the PC to lock up like this?

Edit: Well, it looks like RAM voltage can lock a machine up like that. I changed the voltage to 2.1V and the issue started to occur much more frequently (once I got into a game and it locked up, the second time it locked up while the game was loading). Given this new information, I'm going to search for more information about setting RAM voltage properly. Up until this point, I've just been using the auto setting in my BIOS. The auto setting appears to set it at 1.86V.

[Killgannon]

DO NOT change your RAM voltages past what auto gives unless you are ready and willing to buy new RAM

also BIOS doesnt care about OS... BIOS is your motherboard telling your HDW what to do on the most basic level and doesnt give a shit what OS you are running, all BIOS updates are the same for every OS you will find on Asus's site (their site sucks...) you will likely need a floppy drive unless you can find info on making a boot CD for it, when updating BIOS ignore your OS as it doesnt matter...

[Tidane]

I'm still not certain the issue is with the BIOS. I ran memtest86+ again and it started giving me errors on test 5, which suggests there could be issues with the timing settings. All of my memory timings are set automatically, so there's either an error in the test, faulty memory, or an error with the timings. I'm going to do some more research into what the manufacturer settings for this memory are and, if that fails, I'm going to test each stick individually. Of course, that could also be caused by a faulty memory controller.

[Killgannon]

one thing to check before you research more is to make sure your RAM sticks are plugged in correctly going by what you listed as RAM and this: http://dlsvr04.asus.com/pub/ASUS/mb/...-E_SLI_QVL.pdf you need to have both sticks in blue or white, and not one in blue and one in white, something simple that may have been over looked that you could check real quick

[Tidane]

It looks like it was an incompatibility with something in the BIOS. I found a BIOS revision from this year and applied it to the board and it hasn't locked up yet. Hopefully it has been fixed and isn't just fucking with me.