Need help with cpu/gpu overheating and shutdown

Nah dualcores can go much much hotter, have you checked that all your cables are in might have unplugged something when you cleaned it. GFX's dedicated power cable?

Else it can be any number of hardware issues, but most often when random reboot's and shutdown's the PSU is the culprit... then motherboard, RAM, gfx card.

Start testing if you have spare parts to switch with.

BTW is it only when you play SoaSE as this happens tried any other hardwaretaxing games? maybe do this first before you go wild on the hardware.

Hello, try removing case side panel and see if temperatures drop. There might be an air cycling problem if you don't have cooling fans mounted in your case.

One thing I did to fix a shutdown due to overheating was get a higher rated power supply.  As the power supply heats up, it looses efficency and may drop below the required output due to the draw from the processor/video card and thus brownout the computer.

Well I figured it was something temperature related because if I reboot and try to play immediately (or do anything immediately, really, like just go online) odds are it shuts down again. For me to avoid it shutting down, I have to leave the pc off for a while before turning it on again, which to me just sounds like letting the thing cool down.

I have an extra intake and exhaust fans running all the time, mounted in front of the tower and in back (respectively).

I'd try the PSU (i.e. swapping it out) but I dont have any extras. Any other ideas for what I can do? Or that I can do to check if something else is the problem?

Is there any way I can check the efficiency of my PSU? I have 550W, and a single 8800 GTS 320 mb. From what I read in specs before I bought it I was led to believe I had ample power.

How about it being a heatsink issue or the air-flow has been modified???

I myself have had a heatsink issue of overheating (never actually shut down because of it but you coul,d feel the heat), I opened the pc up and tried to reorganise the cables to allow a better airflow! Also, my pc now has half of its casing permanently down and since then I've been running flawlessly!

Hmm, maybe I'll try that next time and see if the temperature spikes like it has (was getting around 103 degrees a couple of times under load). So out of curiosity, what IS the ideal temp that you should see for a dual core CPU and a gpu? Both idle and under load?

I have Core 2 Duo E6600 @ 2.6 GHz and GeForce 8800GTS 512MB and i have around 34°C CPU, 45°C GPU idle; 60°C CPU, 65°C GPU under load. I have two case fans and original boxed fan on cpu.

Had the exact same problem with the same processor on a friends machine, and the culprit was the cpu heatsink, i took it off cleaned it up, and remounted it, i think it was never mounted properly to begin with, he has never had a problem since, although the pc made an audible siren sound after it shut down and he tried to reboot

@Spell
   I have almost the exact same system, same number of fans, etc. Your CPU and GPU are running under load at about what mine run idle. Do I just need to slap some more fans on? I already cleaned them both out, made sure the heatsinks are on correctly (and yeah it did help but the temps I'm describing are already after all of this, they used to be worse).

Any other ideas what I can do?

Spell what app are you using to check your temps?

coretemp and realtemp... they sometime show diffrent temps.. because the tjmax is diffrent.. when running coretemp my

E6850 ES (stock speed (3Ghz)with scythe infinity cooler) shows 43 idle and realtemp shows 48.... there is always a 5 degree difference between these apps and its becuase tjmax in realtemp is 90 and in coretemp 85.

My gpu is at 76 and its a 8800 gts 320mb.

Make sure your app knows what tjmax your cpu have and they differ from stepping, to get correct value.

srlarue check that your cooler is on flush and a reasonable amount of thermal grease/paste on it.

but be careful...to much thermal paste actually hinders cooling causing the cpu/gpu to not cool properly because the paste is too thick to correctly transfer heat to the HSF.

on a side note voidcore I am running the same 8800gts 320mb, according to ntune my temps are high 50's idle to mid-high 60's load.

@srlarue

"what IS the ideal temp that you should see for a dual core CPU/gpu? Both Idel and load?"

what CPU are you running?

CPUs...generally you should be looking to stay @ or below 40C Idle (25-30C is about the best you can expect without water cooling), and below 70C Full Load. really you don't need to worry too much about CPU temps so long as you don't allow them to reach 70+ that is where the danger zone of burning up a CPU begins.

GPUs...typically run hotter, it is not uncommon to run 60C idle and 70/75 Load, you should still avoid getting up to 80+ though. and if you can cool down to 40C idle 50 load with recent GPUs your doing pretty good. usually water cooling is need though to get that low.

excuse the back to back/dbl post but this needs its own.

the efficiency has NO bearing on whether or not the PSU is sufficiently powered.The ONLY thing as far as Power output that the efficiency rating will determine is that a poor efficiency will cause the PSU to run HOT, a prolonged period of running hot will degrade the PSU, and that will effect it's power output because PSUs lose output ability over time.

the efficiency rating simply means x percent of power coming into the PSU from the wall socket will be converted to Watts to supply the needed amps. the rest is lost to heat.

example...a 73% efficeint PSU will use 73% of incoming power to convert to watts, the remaining 27% is lost to heat in the conversion process.

However an equally IMPORTANT factor of efficiency is the tempurature that the rating was given at. If that rating was given at 25C the PSU is GARBAGE!!! and there are A LOT of junk PSUs out there that use this exact BS....what they want you to see is the 70%+ rating but not the temp is was rated at....

the PSU will be 25C (77F) an hour after complete shutdown and even longer in hot climates.this temp plays into efficiency because the efficiency rating is given at a particular temp saying that at that temp the PSU will be x efficient. so to say 70% efficient at 25C means you PSU will operate at 70% efficiency for the first few SECONDS it is powered on, after that your efficiency will plumet to more like 40 or 50% at normal operating temps which are closer to 40 or 50C.

Thus make sure the efficiency rating was given to the PSU at 40C+.

I could go on this topic with a book.....but instead I will link below to that info.

I would suggest a few things.....

1. visit  jonnyguru.com   the guy knows his shit about PSUs, browse his ratings of different PSU and use his site to make a WELL educated purchase on PSU's

2. the book I could write is here....I would suggest reading it. It may enlighten you to Just how important your PSU is. the more I read of this thread the more I lean toward your problem being a lack of power. If that is the case sooner or later, (sooner in your case, if indeed your PSU is causing your problems) that PSU will fry that video card, and maybe the motherboard with it due to lack of power.

3. remember it is not ALL about the watts, but rather about the AMP output of the PSU.

if your lucky your problem could also be from driver conflict or corruption. files can become corrupt causing conflict and crashes. or it may be that you have installed some software recently that conflicts with other software.though possible I really don't see this as your problem with it only happening during gameplay unless it's due to game files becoming corrupt.

in any case if it were me I would troubleshoot both the PSU and driver/spp conflict ASAP before you give the PSU a chance to fry any hardware.

good luck

blackcurtain is telling you just about exactly what I've seen around the internet...my pc runs a bit hot but it dosnt shut down.That dosnt mean I'm not doing slight damage.Heat is a pc killer...cooler is better so there can never be too much airflow,only crappy airflow from misplaced fans and jumbled cables etc...

Keep in mind,I run with the side off...I have a really quiet pc.

Cleaning got me 10C cooler...velcroing a leftover fan pointing in from the left(directly towards the MB),just under the graphics card got another 5C for the GPU...next step is a proper heatsink and fan...mine was originally for a sempron and its too wimpy for an AMD64(overclocked)...cpu fan rpm's are a measly 2500.

Any temps over 80C would worry me.Your power supply sounds adequate.But...I have lost more power supplies to uneven wall voltages and current than I care to remember.

Yeah psu's is what break's down most often for me and causes the most mysterious crashes when faulty, after that RAM following mobo as nr3 never had to switch a CPU

I have my side up(atm) and fresh airflow is not to good in my case haven't bothered buying the extra 5.25 inlet fan for my lian li but it have bugged me. Outlet flow is good with 2x120 and separate chamber for the psu to handle. Gpu is not getting all the fresh air it needs. But ive had my comp at 3.6Ghz (3.8Ghz is max on air on this sample stable) 24/7 without crashes and running benchmarks for whole nights (fried some ram's after some months running hot and high but hey lifetime warranty is good heh).

Then again what i expect of my comp is 2-3years hard service then its time for a totally new one and hand of the old one to my wife.

Great read, thanks all.  Got some good tips and will apply them.

I mainly use Core Temp or Asus PC Probe II which show quite similar temperatures. I've downloaded real temp and it showed about +5°C as you mentioned. Also i have a utility for showing system info named SIW,  but it shows temperature about +15°C, which i don't think is correct.

For gpu temperature i use Smart Doctor.

I've had very similar problem with my computer after replacing PSU and GPU for newer ones. I didn't have case fans and there wasn't enough air flow to cool components. After about 30 minutes you could place your hand on case and feel that it is very warm. The temperatures were also very similar to yours. After removing side panel the temperatures dropped. Eventually I've solved this problem by buying a better case with mounted fans.

Because you already have case fans then i'd guess that something is affecting an airflow or the fans are malfunctioning. You could also look on cpu and gpu fan speed if something is not slowing them down.

Basics

1. Are the fans spinning

2. Are the fans full of muck.

3. Are any of the fans making odd grinding noises (sign of the bearings going).

4. How long since you refreshed the thermal gunk?

5. Google 'thermal paste'.

7. Sniff around the power supply for the smell of ozone.

8. Ignore 'nifty utilities which show you temprature' they habitually lie; apparently my Semperon 2800 is BOTH 80dC and 43dC according to lm_sensors (two different sensors both allegedly CPU).  Touch the damn heat sink.  If it burns your finger you've got an issue; probably VRM related.

9. Look for buldging capacitors on the board.  No, that bad batch never quite vanished... even modern boards have had bad caps.

6. Swipe a paint brush over the heat sinks and blow them out with air duster.