2

I have this computer that I did assemple, since 2013.

It is an i5 4690 @ 3.50 (not overclocked; I use the standard ASUS OC setup, which is set in the motherboard automatically). The video card is a GTX 9700, also from Asus. I did buy all from the same brand to minimize issues.

Since few weeks ago; I did notice that the computer fan would crank pretty hard; it is hot here, but has been hot also in the past 3 years after all. The GPU stay at idle at 39C, with fan at max, the CPU is at 40C and the mobo at 39C. I use the standard CPU cooler that came with the CPU, and the case has a front fan.

I did play any sort of game, cranking up the settings until the computer would scream; and never had issues (Guild wars 2, star citizen, Elite dangerous, Black desert online, crysis 2, among the ones that I play often). But recently, I did notice that the CPU overheat quite a lot, going at 70C and beyond; in some case the game crash; in other instance teh computer just forcefully reboot.

Do I have to replace the CPU fan? Thermal paste only? Is this a software issue? I do recall that I did update Windows 10, few weeks ago, and the Nvidia drivers, via Geforce experience, which is when the issue started.

Is there anything in the windows logs that would show software issues? Before take apart the CPU fan, I want to try anything less radical first.

UPDATE---------------------

I did remove the CPU cooler, nothing out of the ordinary, beside the paste that was quite dry; so I did clean up both the CPU and the fan surface; and re-applied Artic Silver 5. The temperature seems to be more stable, but still the CPU is going above 73; although now it just cause the game to hang; the computer does not shut down by itself.

This is the first time that I see something similar; I've been in the same place for 12 years; and none of my computers ever did overheat; staying on night and day. Could it be that the CPU is damaged? Maybe I am still under Intel Warranty.

  • You really won't know until you look. Is the fan spinning? Is it clogged with dust? Did something break or fall off? – David Schwartz Jun 12 '16 at 8:05
  • Use something like SpeedFan to monitor temperatures and fan speed. Then look at task manager's CPU utilization graph and see if anything is maxing it out or pegging it to 100%. – TheKB Jun 12 '16 at 8:58
  • The fan is spinning; the RPM are displayed in the fan monitor utility that came with the motherboard; and the fan kick in full throttle at the correct threshold. There is no dust; the computer is a small cube; on my desk; I use a can of compressed air to clean the dust off every 4-5 months. Nothing broke or fell of, from what I can tell. Task manager does not show any process taking more than 20% of CPU; memory stay around 22% idle, and when I use applications it goes up to 60% but never go beyond that (got 16 GB BTW). I did rule out usual suspect; which is why I don't get why is overheating – rataplan Jun 13 '16 at 2:55
  • "but still the CPU is going above 73" if you are referring to Tcase but not Core temperature, then I'm afraid your CPU is permanently damaged, as it exceeds Intel spec of 72.72°C, and you report a constant environment and usage pattern for the past 3 years. (hot weather and pushing performance to the extreme) – guest-vm Jun 14 '16 at 2:43
  • @guest: the software measure the die temperature; this is where the sensor is. There is one sensor in the die, and one on the motherboard; close to the mobo chipset. Not sure how you call each :) – rataplan Jun 14 '16 at 6:11
0

CPU would thermal throttle to 1.6GHz when overheated, which would lag the game but shouldn't crash it. From your description, it is not clear if your CPU failed to avoid overheat even if throttled and eventually shut itself down. Run CPU stress test (eg. OCCT, AIDA64), and monitor temperature and clock speed change. If it quickly throttles or crashes, then you should check your cooling (dust off fan, reapply thermal paste, etc). If it wasn't cooling and you suspect windows/driver update at fault, you would have to do your own testing (eg. system restore).

  • CPU does not throttle when under stress in my case. Also would be counterproductive if you are using an application that require 100% of each core, and the software would reduce your performance. I believe it throttle down only when using windows applications, not games. Also, this started recently; for 3 years it never gave any issue. – rataplan Jun 13 '16 at 2:59
  • The throttle for windows applications is due to speedstep, not overheat. So it seems either your CPU never overheated or the protection mechanism is broken or overrode. You mentioned "standard ASUS OC setup", is it a Z87? If so, can you disable "MultiCore Enhancement" in BIOS, then run a stress test to see if your rig is stable? – guest-vm Jun 14 '16 at 0:43
  • Yes, speedstep; I never check if the frequency and voltage was going down with higher temperatures; that's why your comment abut throttle was confusing to me. The board is a Z97i-Plus; I didn't touch the bios since when I did assemble the rig; it may be possible that an update was installed (or is needed); I will check that – rataplan Jun 14 '16 at 6:14
0

Thanks Guest for checking with me; it seems that the problem is twofold.

1) the CPU is defective; I did check with Intel and we did run together a whole set of diagnostics, and it seems that the temp values are off by a bit. Not considering the CPU broken, but they will investigate further, so I am waiting to see what is going on.

2) Asus motherboards have some sort of nasty OC; which is set for some reasons by default. I did check in the bios and there was a setup for turbo; which would increase operational speed of the CPU (in the limits of what is considered safe), at cost of slightly lower energy efficiency. I did run with turbo on, for years; but now, due to the CPU giving troubles, that is not working anymore; so I did set to basic setup the motherboard; and at least now I can use the computer without crashing every time. I did not notice loss in performances so far, but I did not have an extended session with a game that would cause the reboot; so I have to give a try this weekend.

Just in case someone may experience the same issue as me; I am leaving these info here. The Intel tech was knowledgeable, so it was not that bad.

  • 1
    Most Z-chipset mobos have MCE on by default, not just Asus. – guest-vm Jun 15 '16 at 6:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.