My computer has recently started crashing on me after roughly 20 minutes in any game. I've had this rig for over three months and have gamed a lot on it prior to the recent crashing issues and I've run numerous tests on the memory and power supply and they both check out so I'm thinking it's heat related. It has been quite warm where I live lately and these crashes have coincided with the warmest of the weather, not to mention I have been able to game for longer periods of time by cranking the a/c all the way, putting the computer closer (live in an apartment with just a window mounted a/c) to the a/c itself and turning all the fans to high.... this of course is a pain but is all evidence towards the problem being heat... so with that in mind how do I cool my system down?
-Coolermaster Cosmos S with three 120mm top mounted exhaust fans, one 120mm rear mounted exhaust, one side 200mm intake, one front 120mm intake, one bottom 120mm intake (so I'm getting lots of air circulation) SpeedFan puts the "Aux" idling at 34C.
-Intel E8400 (reduced to stock speed of 3GHz in my journey to find the root of the problem) with a Noctua NH-U12P cpu cooler in a push pull config, SpeedFan puts the CPU idling at 33C with Core 0 at 43C and Core 1 at 23C
-EVGA 750i FTW Motherboard (Thinking this is the culprit) known for it's hot north and southbridge, SpeedFan says the "System" is running at 120C from start up and it doesn't waver (so I'm thinking this is a glitch) also it has the "Temp" (I'm thinking northbridge) idling at 54C, I can get this number as low as 50C using the complicated air conditioning method mentioned above but it just ramps up during gameplay and after an instant alt tab from 10 minutes of TF2 the "Temp" was at 59C. Temp 1 in SpeedFan has never been too high hovering under 50C with an idle of 40C.
-Geforce 9800GTX overclocked graphics card (usually use 2 in SLI, took one out when during the investigation faze), there is warm air coming from the back of the card (as normal) and it idles at 54C according to Riva Tuner which I don't think is so bad for a graphics card.
Not sure if the rest are related to heat:
-Corsair TX750W PSU bottom mounted with fresh air intake
-4GB Crucial Ballistix memory (set it to 4-4-4-12 timings, everything else not overclocked) I'm using a Antec Spot Cool aimed at the memory as well.
-500GB Seagate Barracuda 7200.11, one of the intake fans blows over the HD cage.
-Scythe Fan controller
So do you guys agree with me that it's heat related? How should I fix the problem? Swap in different chipset heat sinks? Would that be a big job? What would you recommend for Northbridge and Southbridge heat sinks (keeping in mind I live in Canada and not everything is available here)? Or should I wait for winter?
With some help I determined the problem to be whichever program controlled my video card fan speed (I have 3 suspects) had it locked at 35%... that's right, it was my video card all along and given the hot weather we've been having 35% just wasn't enough for in game. It turns out also that the "Temp" reading I was getting from SpeedFan was actually the GPU temperature, not the north bridge at all which leads me to believe I just don't know how to read SpeedFan. The temperatures SpeedFan is currently giving me are
CPU: 33C <--- Guessing that's the CPU
Temp: 44C <--- Determined this was the GPU
Note I only have one video card currently installed so "Temp1" could not be the second, also "Temp1" stays at 40C no matter how hot "Temp" (the GPU) gets, so perhaps they aren't even related. That also makes me wonder if Core0 and Core1 are actually the two cores on my CPU (which gets its own reading) or perhaps they are actually something different, like the North and South bridge maybe? That would account for the massive difference in temperature. My thermal probes have the heat sink on the NB at 39.5C and the heat sink on the SB at 38.3C... but that's the temperature of the actual heat sinks themselves and they are connected so one reading could influence the other. Does anybody know how to read these SpeedFan temperatures? I went to their website and that didn't help any, I'll check the bios readings after work so hopefully the descriptions there are more specific.
So ya, sorry my other post was deleted (I posted during the scheduled maintenance and it must have been up for only a few hours)...
I've done a little research and apparently on the E8400 there is a known stuck sensor bug which causes a variance in the core temps. I also found some people recommending to just re-lap my heat sink and re-seat in as there is probably a concave over one of the cores. I will do this when I replace the chipset heat sinks as neither of the cores are anywhere close to critical levels (I ran Prime95 and the hot core got to a whopping 47C!)
Thanks for the warning though, RealTemp confirmed the temperature variance.