Causes of random restarts except PSU failure

[Ilmiont]

Hi guys,

 

Looking for a bit of help/diagnosis on something although I'm pretty certain I know what is wrong...

 

Basically, I built my computer back in May using an AMD A10-7850K... since then until last month I was using the A10's integrated graphics.

Then I bought an Asus Radeon R9 270X 4Gb edition... and since then, occasionally, the computer freezes up completely so that I have to press the reset button or the monitor simply goes off and does not return until rebooted.

The computer is powered by a 500W Corsair CX500M PSU and I believe this may be the problem... insufficient wattage but are there any other possible causes?

In terms of hardware drawing power, I have 3 SATAs hooked up to a HDD, SSD, and optical drive; 2 case fans; stock AMD CPU cooler, AsRock FM2A88X Extreme4+ mobo.

 

Any suggestions appreciated but I'm thinking I need a new PSU soonish.

 

Ilmiont

 

[Mindovermaster Moderator]

The PSU is not considered low with that setup. A bit lacking, yes, but not crucial.

 

If it freezes up, but still runs, it is not the PSU. Can you verify by taking the card out and try with your onboard, see if it freezes again? There may be something wrong with yourgfx card.

 

How are your temps?

[MikeFu84]

Rough calculations, not 100% accurate but give you an idea.

 

Each hard drive will take about 10W of power, the SSD about 3W. Motherboard could use up to 80W and the CPU up to 125W. Each stick of RAM uses about 3W each as well as each fan. Optical drive uses about 30W. High end GPUs can use between 250-330W roughly.

 

Based on your build:

3 x HDD @ 10W - 30W

SSD - 3W

DVD - 30W

Motherboard - 80W

CPU - 100W

2 x RAM @ 3W - 6W

2 x Fan @ 3W - 6W

GPU @ ~225W

-----

Total ~480W

 

Personal opinion, only having a 20W buffer, you may want to upgrade that PSU.

[Luc2k]

Rough calculations, not 100% accurate but give you an idea.

 

Each hard drive will take about 10W of power, the SSD about 3W. Motherboard could use up to 80W and the CPU up to 125W. Each stick of RAM uses about 3W each as well as each fan. Optical drive uses about 30W. High end GPUs can use between 250-330W roughly.

 

Based on your build:

3 x HDD @ 10W - 30W

SSD - 3W

DVD - 30W

Motherboard - 80W

CPU - 100W

2 x RAM @ 3W - 6W

2 x Fan @ 3W - 6W

GPU @ ~225W

-----

Total ~480W

 

Personal opinion, only having a 20W buffer, you may want to upgrade that PSU.

270X is nowhere near 225. Your figures are somewhat irelevant even if they were true since it would require all components at full load all the time. Also OP didn't mention doing anything intensive to cause the freezes. If there's anything wrong with the PSU, insuficient power ain't it.

 

OP, how are your temperatures looking? Also try and change the GPU drivers to different versions and see if it reacts the same.

[spaceelf]

I'd recommend checking with a different video card or a different PSU.  I had similar problems until I switched out my PSU.

 

(I went down to a 500w Antec EarthWatts and have a 660 in here, so wattage is unlikely the problem.)

 

Also update BIOS and chipset drivers is always a good idea - 

http://support.amd.com/en-us/kb-articles/Pages/desktop-apu-catalyst-software.aspx

http://asrock.com/mb/AMD/FM2A88X%20Extreme4+/?cat=Download&os=BIOS

[Sandor]

Check event log...it'll likely give you a good idea.

[spaceelf]

Check event log...it'll likely give you a good idea.

Never said anything useful for me.

 

I'm inclined to believe ASrocks voltage regulator on these boards is twitchy (Extreme6+ here.)

[Ilmiont]

OK responding to a few posts in one...

 

It occurs generally whilst streaming video or playing games. I often occasionally get "AMD Display Driver stopped responding and successfully recovered" warnings... I am on the latest graphics drivers and the card idles at ~25C.

Actually, the card does take 225W I think to whoever doubted that.

 

I took the card out and put it back in this evening so I will see it makes any difference, I also checked the PCI power connectors and reseated the RAM.

 

I have only the card for 6 weeks and the rest of the build since May.

 

Ilmiont

[eddman]

GPU @ ~225W

Actually, the card does take 225W I think to whoever doubted that.

 

No, it doesn't.

 

http://www.techpowerup.com/reviews/EVGA/GTX_970_SC_ACX_Cooler/23.html

 

130W peak

 

173W max while stress testing with Furmark a.k.a unrealistic

 

Nowhere near 225W. That's just the theoretical possible max power delivery: 75W PCI-E slot + (2 * 75W 6-pin) = 225W

 

Sure, your card is overclocked, but a 5% or 12% higher frequency won't increase the power consumption that much; not even close.

 

EDIT:

 

http://www.techpowerup.com/reviews/ASUS/R9_270X_Direct_Cu_II_TOP/24.html

 

137W peak

 

178W max (stress testing - unrealistic)

[Sandor]

OK responding to a few posts in one...

 

It occurs generally whilst streaming video or playing games. I often occasionally get "AMD Display Driver stopped responding and successfully recovered" warnings... I am on the latest graphics drivers and the card idles at ~25C.

Actually, the card does take 225W I think to whoever doubted that.

 

I took the card out and put it back in this evening so I will see it makes any difference, I also checked the PCI power connectors and reseated the RAM.

 

I have only the card for 6 weeks and the rest of the build since May.

 

Ilmiont

I had a GTX760 that started doing that with nvidia drivers...stopping responding and then recovering. It finally melted itself about a week later. Never OC'ed or dirty...somehow it still overheated and went in for RMA.

[Luc2k]

It occurs generally whilst streaming video or playing games. I often occasionally get "AMD Display Driver stopped responding and successfully recovered" warnings... I am on the latest graphics drivers and the card idles at ~25C.

Actually, the card does take 225W I think to whoever doubted that.

What exactly is the latest in your case? 14.4 WHQL, 14.7RC3, 14.8 APU WHQL (which also supports desktop GPUs in the ini), the R9 285 driver unofficially modified to work with all GPUs or the 14.x RC1 OpenCL 2.0 driver?

 

I gave you a link to a review that measures ONLY GPU power consumption with professional hardware so I'm not sure where you're getting the 225W figure from. 130W is the peak value in Metro: Last Light, which is a very intensive game and 173W is the maximum consumption, and that value is hit by stress testing the card.

 

We can try and help you out, but that means you have to actually read what we have to say and link.

 

Going by "AMD Display Driver stopped responding and successfully recovered", I would recommend a clean driver reinstall. Use DDU in safe mode and remove the old ones, then install the "latest driver" of your choice. The modified R9 285 driver is pretty good from reviews, but 14.7 RC3 is a safe choice. I haven't read anything on the OpenCL one since I've just found out about it.

[spaceelf]

stock AMD CPU cooler

Speaking of which I don't know that I'd qualify that thing as a 'cooler.'  Heh.  More like a joke.

 

My CPU temps went down to the 40s when I slapped a decent one in here.

[farmeunit]

OK responding to a few posts in one...

 

It occurs generally whilst streaming video or playing games. I often occasionally get "AMD Display Driver stopped responding and successfully recovered" warnings... I am on the latest graphics drivers and the card idles at ~25C.

Actually, the card does take 225W I think to whoever doubted that.

 

I took the card out and put it back in this evening so I will see it makes any difference, I also checked the PCI power connectors and reseated the RAM.

 

I have only the card for 6 weeks and the rest of the build since May.

 

Ilmiont

Try using an older driver as well, first.

 

Even if you don't think it's the PSU, it's worth testing IF you can get your hands on a spare from a friend or something.  It may have plenty of wattage, but not be stable.  I've been buying Corsair lately myself after using Antec for just about my whole life building machines.  Have had good luck with them so far, but ALL parts will have some bad instances from time to time.

 

I had a computer once that showed the display driver crashing every day, several times a day and swapped the PSU out, and never had an issue again.  That would an old P4 with onboard Intel, so a little different, but just an example.

[MikeFu84]

270X is nowhere near 225. Your figures are somewhat irelevant even if they were true since it would require all components at full load all the time. Also OP didn't mention doing anything intensive to cause the freezes. If there's anything wrong with the PSU, insuficient power ain't it.

 

OP, how are your temperatures looking? Also try and change the GPU drivers to different versions and see if it reacts the same.

 Irrelevant at this point now that he's brought up the driver issue but Asus claims up to 225W (http://www.asus.com/Graphics_Cards/R9270XDC24GD5/specifications/).

[farmeunit]

 Irrelevant at this point now that he's brought up the driver issue but Asus claims up to 225W (http://www.asus.com/Graphics_Cards/R9270XDC24GD5/specifications/).

eddman explained how this works up above a few posts.

[Luc2k]

 Irrelevant at this point now that he's brought up the driver issue but Asus claims up to 225W (http://www.asus.com/Graphics_Cards/R9270XDC24GD5/specifications/).

If he was going with your advice, he would have bought a new PSU without even mentioning the driver issue so I don't see how clearing up misinformation is irrelevant. Manufacturers can claim whatever they want to cover their asses, but we should work with real figures, not make-believe, and that includes the rest of your power consumption figures.

 

A little update on my previous post: ignore the OpenCL driver.

[Ilmiont]

What exactly is the latest in your case? 14.4 WHQL, 14.7RC3, 14.8 APU WHQL (which also supports desktop GPUs in the ini), the R9 285 driver unofficially modified to work with all GPUs or the 14.x RC1 OpenCL 2.0 driver?

 

I gave you a link to a review that measures ONLY GPU power consumption with professional hardware so I'm not sure where you're getting the 225W figure from. 130W is the peak value in Metro: Last Light, which is a very intensive game and 173W is the maximum consumption, and that value is hit by stress testing the card.

 

We can try and help you out, but that means you have to actually read what we have to say and link.

 

Going by "AMD Display Driver stopped responding and successfully recovered", I would recommend a clean driver reinstall. Use DDU in safe mode and remove the old ones, then install the "latest driver" of your choice. The modified R9 285 driver is pretty good from reviews, but 14.7 RC3 is a safe choice. I haven't read anything on the OpenCL one since I've just found out about it.

I'm using 14.4 right now, the latest release edition but will reinstall later then.

 Irrelevant at this point now that he's brought up the driver issue but Asus claims up to 225W (http://www.asus.com/Graphics_Cards/R9270XDC24GD5/specifications/).

Yep, that is where I sourced that from... didn't realize it uses so much less IRL.

 

 

I will be upgrading the PSU as soon as I can anyway but the issue here is finance... I'm only 15 and although a CX750M is only ?70, I still need to find that having just bought a Logitech G27 and the R9 270X itself :D

[philcruicks]

I don't think anyone has posted this yet, so a good PSU calculator is

http://www.extreme.outervision.com/psucalculatorlite.jsp

 

From your description of events its not likely to be the PSU though, I've had 2 upgrades in the past (both new GPU's) one going on as I type, where the PSU was no longer powerful enough, currently got an 4 year old 600W PSU, put new card in and unlike your description the PSU not providing enough power in my case (both now and a few years back) just causes the system to re-boot, no drivers crashes, no freezes, only happens when playing intensive games like BF4, working one second, then powers down, and beep back to BIOS. lack of power isn't going to cause driver crashes.

 

I've just upgraded to an 850W and waiting to the new case to be delivered (too big for my current one).

 

If you still think its the PSU then before ordering a new one try downclocking the card, its what I've done while I'm waiting for the new PSU, MSI Afterburner software lets me drop the power usage of the card, 10% was enough for me and now I can play games without issues, just obviously not running at full power.

[Ilmiont]

OK this is an interesting twist then... looking around the net shows that "AMD Display Driver stopped responding and has successfully recovered" is far from an uncommon issue on the R9 270X... worryingly though, it seems as though it gets worse with time and there are no specific remedies detailed other than reinstall drivers which I'm doing right now. Maybe it will be fixed in a later Catalyst then, I'm on 14.4 right now with graphics drivers 14.10.1006-140417a-171099C.

[+Zlip792 MVC]

"AMD Display Driver stopped responding and has successfully recovered"

 

This error is nothing related to Power Supply issue, its mroe Timeout Detection Recovery, pretty common on nvidia as well. It used to happen with me on GTX 460 with 800 Watt Xigmatek PSU, so I don't think it would be related to PSU.

 

My recommendations:

- Clean latest GPU driver install, might be it could fix it. AMD is pretty famous to leave leftovers from driver un-installation so you can use this utility: http://www.guru3d.com/files-details/display-driver-uninstaller-download.html

Latest GPU driver - http://www.guru3d.com/files-details/amd-catalyst-14-x-rc1-(14-410-september-2)-download.html

- For hacky purpose, you can manipulate Windows TDR ability : https://forums.geforce.com/default/topic/550192/geforce-drivers/wagnard-tools-ddu-gmp-tdr-manupulator-updated-09-26-14-/post/3846695/#3846695