[Gaming Rig] Unexplainable Crashing

You should get both the latest BIOS and the latest drivers from here:
Gigabyte

F5 was still an early BIOS and it's possible the board was not yet stable with that version. The newest version is F10. Also, if you just used the drivers from the disk, you might need updates to the chipset and audio drivers.

It's also possible that the video card or CPU are overheating.

How exactly was the RAM tested?
Memtest
The proper way is to download an ISO from that URL, then burn it to CD. Boot using that CD and then run memtest for several hours, or overnight.

You also said you upgraded the PSU, but you didn't list what model you have. A faulty PSU will cause symptoms like yours where the system runs fine in 2D mode, but crashes in 3D mode. The GPU doesn't draw much power in 2D mode, but a low-end PSU can't deliver clean power when the GPU goes into 3D mode causing a crash. Please list the PSU model so we can rule that out as the problem.

If you replaced a junk PSU with a junk PSU, that would explain it.

lol - That's what I was getting at with my above post.

Increase your VDimm, I had the same issues on the board.

You might want to list the exact model # of your RAM. It's always a good idea to manually set the RAM timings and voltage to the manufacturers specs in the BIOS. Gigabyte boards require you to press CTRL+F1 when you first get into the BIOS to open up the advanced settings including RAM timings.

I had a similar problem with x-plane, and it turned out to be the driver. They just released a new version, and as soon as I updated it, my problem was fixed. I have 8800gts and e6600, running great now.

they're right, probably the PSU or video card driver.
also possible is heat buildup on the card or the card is defective.
and loose power connectors or dirty pci connection of your card.

keep it simple. it sounds, at least to me, like you are not getting enough power to your ram, graphics card or cpu.

1) open case and check all power and card connections. sometimes an expansion card is not seated properly or all the way.
2)restart and go into bios. update bios to the most recent revision. make sure the frequency, timings and voltage for your RAM match manufacturers specs. Make sure the rest of your settings are on auto.
3)restart and go into windows, make sure (go to manufacturer web sites) all your drivers are up to date, run your game or whatever has been causing the crash.

if the problem persists you will need to test your ram 1 stick at a time, try a different graphics card and/or power supply. do each hardware item 1 at a time so you can isolate the problem.

The most important thing I have learned from this forum while building comps is that any problem can be fixed it just requires a methodical approach and can take a few hours to nail down the problem.

I hate to say this but if money is an issue go to your local big box store, buy the parts needed to run the test and if you do not need to keep them then take them back for a refund. I would normally not say that but I just paid $42 for 2 sata cables last night and I am still bitter.

Additional notes:

1) I know its not the PSU because Im using a very nice ThermalTake PSU thats actually more than I need to run the rig.

2) I tried updating the BIOS using Gigabytes recommended BIOS update program @Bios and it said that the most up to date driver for my motherboard was F9. That reaked serious havoc on stability and it took quite a while to downgrade back to the more stable F5 version.

3) On principle I dont mess with the BIOS screen other than to make sure its set to "Optimized Defaults" on the very first run.

4) The Ram Im using is OCZ - PC2 9200 1GB qty 2
part number OCZ2RPR11502GK

5) Considering all the fans I have in the case (5 total) and the fact that I never...repeat NEVER overclock anything (My opinion is that if you need to overclock something you should have gone with a faster model of the item in the first place) its definitly not a overheating problem.

---------------------------

Alright I believe I found the source.

I tried running MemTest 3 times testing only 500 out of 2000 megabytes and the computer locked up every time after only 1-5% progress.

I then removed one stick and am testing 500 out of 1000 and now that Im only 37.7% into the test Im already upwards of 87+ errors. The actual error message says that "Copying between (x) and (y) did not result in an accurate copy. MemTest has detected that your computer cannot accurately store data in Ram.

I have yet to try the other stick, but I have a good idea of the result.


The only thing Im skeptical on is if there truly are this many errors in the ram, then the computer shouldnt be able to detect it at all let alone run. I would think it would stop at the BIOS screen on startup and begin beeping all to hell like its trying to say "Hey! Hey! I cant find any RAM!"



Update:

Final Results for MemTest

Stick 1:
Tested 500/1000mb 201 errors @ 125% Coverage

Stick 2:
Tested 500/1000mb 159 errors @ 125% Coverage


Either this program is flawed or the motherboard is really well made for it to be able to run when ram has that many errors.





I think my most pressing question would be if all these crashings and using bad ram would cause damage to the motherboard or graphics card. Because this was a pretty expensive rig and its going to be bad enough having to chalk up more money to replace the ram only to find out I'll have to replace the MB, CPU or 8800GTS too.

That could very well be causing your RAM problems. That RAM is rated at insanely high voltage (2.1v - 2.3v). The motherboard will only supply 1.8v by default. If you refuse to manually set the RAM timings and voltage to the manufacturers specs you can't expect the RAM to work correctly. I'd try memtest again after manually setting the timings and voltage. The best way around this type of problem is choosing RAM that has tight timings at the standard 1.8v. Here's a couple examples:

mushkin 4GB (2 x 2GB) 240-Pin DDR2 SDRAM DDR2 800 (PC2 6400) Dual Channel Kit Desktop Memory - Retail
http://www.newegg.com/Product/Product.aspx?Item=N82E168... $45.99 - $20 MIR

or

G.SKILL PI Black 4GB (2 x 2GB) 240-Pin DDR2 SDRAM DDR2 800 (PC2 6400) Dual Channel Kit Desktop Memory - Retail
http://www.newegg.com/Product/Product.aspx?Item=N82E168... $49.99

If the BIOS seems too daunting for you, then you should have someone work on the computer that has more knowledge.

Some highlights from the page I first linked:

F10 2008/11/10 1. Fixed Dynamic energy saving tool can't run issue

F9 2008/09/22 1. Fixed Vista S3 resume sometimes failed

F8 2008/08/18 1. Support DDR2 speed to 1333MHz and beyond by overclocking !!!!! PC 9200 !!!!!!

F7 2008/07/23 1. Enhanced memory compatibility !

F6 2008/07/09 1. Enhanced memory compatibility
2. Update CPU ID (Support Intel Wolfdale/Yorkfield E0-stepping CPU) HEY! You have Wolfdale!

F5 2008/06/19 1. Fix VGA compatibility issues

F4 2008/06/12 1. Enhance system performance and overclock capability

So, three of the BIOS updates potentially deal with your problems, and the fact they exist should suggest to you that perhaps the BIOS is an issue.

As to the PSU suggestion, the brand Thermaltake by no means guarantees any sort of quality. In fact, I always avoid them because I find their marketing tactics reprehensible. They will usually have one model of a certain wattage reviewed, then actively push a different model number with the same name and wattage, but never get the inferior unit reviewed. Also, the supposed wattage rating of a given unit is not related to it's quality in any way.

Lastly, having 5, or even 12 case fans and not overclocking doesn't make you immune to overheating. In fact, since you are running an old BIOS and refuse to edit the settings, your BIOS could easily be overvolting the CPU causing it to heat up. Poor insertion of the push-pins when installing the CPU cooler would also cause overheating. Misapplication of the thermal interface material could cause overheating.