Debugging System Crashes

I am running Windows 7 64bit at home.  Before that I was running Windows Vista 64.  I played a hodgepodge of games.  Too many to go into.  Lately I’ve been replaying the Supreme Commander and Forged Alliance campaigns.

Last time I played them, it was on the same system, but running Vista.  Now I am replaying them on Win 7.  It started crashing.  Not just crashing, but locking up my computer.  It starts with a screen flicker (like shadows vertically scrolling down the screen) and then a troubling static like thumping noise coming through my speakers.

I downloaded some temperature gauges.  From past experience this seemed like a thermal problem.  My case temperature was 54 C.  That’s a little on the warm side, but not horrible.  My CPU was running 65 C.  That also seems hot, but not beyond the allowed range.  I’m running an E8400.  Then, I looked at my GPU’s.

I’m running, in SLI mode, a Superclocked EVGA 8800 and a Superclocked EVGA 9800.  Both of these guys were over 80 C.  One was at 90 C.   This seemed excessive.  I started looking at cooling solutions.

It seems to be documented rather well that the EVGA video cards run hot.  Once, many moons ago, I tried to replace the heatsink and fan on a video card.  It was an unpleasant experience.  I would rather not do that.  I discovered that EVGA had a little utility to manage their video cards.  You can set things like “fan speed”.

I tested this and discovered that it has lots of settings.  I could even slow down my Superclocked cards.  However, I was interested in cooling so I set the fan speed to 100%.  This made the cards run cooler initially, but they’d still get hot.  I guess the default is a variable fan speed.  The hotter the card, the faster the fan.  They idle at about 35% which is quiet.  100% fan speed is pretty loud.  Alas this didn’t help me.  And there was no way (that I could see) to change the defaults.  You have to run the utility and change the settings every time you boot.  You can automate it to a certain extent, but it didn’t give me the warm fuzzies that I was hoping for and that game was still locking up.

I bought a Zalman CNPS7700 CPU heatsink and fan.  I did some of this in parallel.  This thing is huge.  I was hoping it would cool off my CPU which would keep the case cooler, which, in turn would keep the GPU’s cooler.  It did all of that.  And last night my computer locked up (again).  Back to the drawing board.

I realized that when it locks up, the vertical shadows always appear on the display and there is the noise from the speakers.  I did two things.  I have an ASUS P5N-D motherboard so I went to Asus and downloaded the Win 7 64bit audio drivers.  And I changed the graphics settings on the video game to not synchronize with the refresh rate of the monitor.  Voila, this fixed it.  I think it was the sound drivers all along.  The motherboard has an integrated Realtec system, but it was using Microsoft drivers.  The refresh was always causing flickering.  It was just more pronounced when the computer was locked up because of the audio driver.