What is wrong with my system

[hunkeelin]

Been playing witcher 3 today. Got freeze and hangs that requires restart. I was thinking it might be nvidia driver or me over clocking my gpu.

Just a while ago, i am gaming witcher 3. My computer suddenly restarts itselfs. gpu isn't OC, cpu isn't OC. This is very strange I am suspecting either my mobo or psu is giving out.

My mobo/cpu/gpu/ram is all new 2011 -v REV 5960x 980 in sli.

my psu is 4 years old 1000w cooler master silent pro 80 plus

I ran speccy and here's my voltage

CPU CORE 0.928 V
MEMORY CONTROLLER 1.016 V
AVCC 3.344 V
3VCC 3.344 V
VIN4 1.016 V
VIN5 1.968 V
VIN6 0.400 V
VIN7 0.280 V
VIN8 0.928 V
VIN9 0.408 V
VIN10 1.016 V
VIN11 1.016 V
VIN12 0.440 V
VIN13 0.424 V
+5V 5.080 V
+3.3V 3.344 V
+12V 12.192 V
VIN3 0.888 V
VIN4 0.784 V
VCORE 1.856 V

This is bothering me to no end. Event log gave me the same old useless kernal power 41

Update #1: The games freezes with audios too.

Does that deduce anything? Nvidia driver? Unstable OC?

This morning I oc my card. I got msi 980 oc edition in sli i just oc to 1390mhz +87mv 122% power limit. It should be stable. But game freezes with audio.

Can that be a psu problem? From my experience it sounds like a driver/game problem.


However, I am still perplex on yesterday's incident where my computer restart by itself while gaming. Cuz that's omre incline with hardware issue than software issue.\


Update #2

I just got a bsod when gaming witcher 3 with NO OC.

 

[daveybrat]

You said you're running 2 GTX 980's in SLI. Is that correct? Your Sig shows 2 GTX 970's in SLI.

Either way, remove one of the cards and try running your games without SLI. If it works fine, it might be time to RMA or look for a new power supply.

 

[ShintaiDK]

It could sound like a PSU issue. Try run with a single GPU and see if it happens. Since the powerdraw will be obviously lower.

And try run some stresstest like Prime95+Furmark or something to see if it triggers it as well.

 

[TheELF]

Witcher 3 is really demanding and will push your temps up so check for temps first,even at stock clocks if you haven't cleaned the cards they might be raising temps too high.

 

[BonzaiDuck]

I think I'm with the others on the PSU. You're very close to the end-of-warranty if that particular CM model is a good one. I'm guessing you read Speccy with the system at idle, and I don't see any problem there, unless someone else corrects me about a reading of the other voltages excluding those immediately applicable to the PSU itself: 12V, 5V and 3.3V.

But it could also be a GPU problem, a GPU driver problem or similar.

IF you updated the drivers for the 980's, did you make sure the installer did a "clean install?"

Ultimately, if I were building a system with a $1,000 processor and two $600 graphics cards, I wouldn't insert a 4-year-old PSU into the mix. Good or bad, I would've swallowed the price of a new one. And I would go directly to the source of many brands, to buy a top-tier Seasonic.

But rich or poor, flush with cash or otherwise, go through some of the troubleshooting recommendations posted by our colleagues before you start replacing parts.

 

[dark zero]

And how about RAM problem? Maybe the problem comes from there.

 

[hunkeelin]

Update: The games freezes with audios too.

Does that deduce anything? Nvidia driver? Unstable OC?

This morning I oc my card. I got a msi 980 oc edition i just oc to 1390mhz +87mv 122% power limit. It should be stable. But game freezes with audio.

Can that be a psu problem? From my experience it sounds like a driver/game problem.


However, I am still perplex on yesterday's incident where my computer restart by itself while gaming. Cuz that's omre incline with hardware issue than software issue.

 

[ElFenix]

Why are you overclocking your card when you already have stability problems?

 

[hunkeelin]

Why are you overclocking your card when you already have stability problems?

I am running several test. I just got a bsod with no oc while gaming witcher 3. Have you heard of driver/game that can cause bsod. Bsoding is more like hardware issue than software. Before that I was able to run metro LL benchmark 10 times without problem while OCINg the card. I am really confused.

 

[hunkeelin]

I don't think is ram problem because I am not getting any freeze when i'm working. I boot up vm too so the ram usage is higher for me when I am working than gaming.

 

[hunkeelin]

https://www.reddit.com/r/techsupport/comments/36ink4/witcher_3_bsod_after_about_5_minutes/

^ it seems people owning nvidia card is experiencing bsod in witcher 3 as well.

 

[BonzaiDuck]

I am running several test. (1) I just got a bsod with no oc while gaming witcher 3. . . . (2)Have you heard of driver/game that can cause bsod. Bsoding is more like hardware issue than software. (3) Before that I was able to run metro LL benchmark 10 times without problem while OCINg the card. I am really confused.

(1) Then, apply some simple logic to your testing. If you got a BSOD WITH NO OC, it would occur WITH an OC, would it not?

(2) Positively . . absolutely . . . "YES!" This is the confusion that greets less-experienced overclockers, and it has also happened to me. BSOD can also occur from a driver bug, driver conflict, a conflict in the assignment of computer resources or lack thereof. The Stop Code may "show" an indication that there's a "hardware problem," but it may not derive from either a hardware defect or even an OC setting. Such a problem can even show a BSOD for buggy software!

And (3) -- As you discovered in the reddit link and as I just said, buggy or badly-designed software can result in freezes, resets and . . . B . . S .. . O . . . Ds!

Also, ElFenix has a clear eye about this. If you had no stability problem before you installed this . . Witcher 3 game, hopefully you stress-tested the living daylights out of your system before concluding "no stability problem."

Also, I have to tell you, perhaps to ask another question. You say you have an MSI GTX 980, and you clocked it to a modest 1,390 Mhz with a voltage bump of 87mV and power limit pushed to 122%. I haven't looked keenly at the specs for the 980 card, but the default top speed for it is 1,317.

And . . so the specs to my GTX 970 (2x SLI) cards are -- on the surface -- identical: 1316 Mhz and about 7016 Mhz for the memory clock. I was able to push the core clock to 1440 Mhz and memory clock to 7,400 Mhz on both cards before raising either voltage or power limit. Getting to 1470Mhz core and 7,500 Mhz memory took me only +20mV in the voltage, and I need not increase my power limit (maximum 110%) over 100%. I've set the power limit to 105%, though.

Son! Son! Lemme give you some advice! Slow down a little bit on this business! You HAVE a GREAT graphics card there! Don't go pushing switches and sliders all at once! Leave the memory clock at stock; increase the core clock by maybe 20Mhz at a time, and stress for 15 minutes with FurMark each time, raising it another increment until it crashes. BAck it down, and start increasing the RAM clock maybe 50Mhz increments until it does the same thing. The rest of it you can discover in the OC'ing guides for these cards.

But before you do that -- figure out of your problem is either Witcher, the NVidia driver or something else before fiddling around with AfterBurner.

 

[hunkeelin]

(1) Then, apply some simple logic to your testing. If you got a BSOD WITH NO OC, it would occur WITH an OC, would it not?

(2) Positively . . absolutely . . . "YES!" This is the confusion that greets less-experienced overclockers, and it has also happened to me. BSOD can also occur from a driver bug, driver conflict, a conflict in the assignment of computer resources or lack thereof. The Stop Code may "show" an indication that there's a "hardware problem," but it may not derive from either a hardware defect or even an OC setting. Such a problem can even show a BSOD for buggy software!

And (3) -- As you discovered in the reddit link and as I just said, buggy or badly-designed software can result in freezes, resets and . . . B . . S .. . O . . . Ds!

Also, ElFenix has a clear eye about this. If you had no stability problem before you installed this . . Witcher 3 game, hopefully you stress-tested the living daylights out of your system before concluding "no stability problem."

Also, I have to tell you, perhaps to ask another question. You say you have an MSI GTX 980, and you clocked it to a modest 1,390 Mhz with a voltage bump of 87mV and power limit pushed to 122%. I haven't looked keenly at the specs for the 980 card, but the default top speed for it is 1,317.

And . . so the specs to my GTX 970 (2x SLI) cards are -- on the surface -- identical: 1316 Mhz and about 7016 Mhz for the memory clock. I was able to push the core clock to 1440 Mhz and memory clock to 7,400 Mhz on both cards before raising either voltage or power limit. Getting to 1470Mhz core and 7,500 Mhz memory took me only +20mV in the voltage, and I need not increase my power limit (maximum 110%) over 100%. I've set the power limit to 105%, though.

Son! Son! Lemme give you some advice! Slow down a little bit on this business! You HAVE a GREAT graphics card there! Don't go pushing switches and sliders all at once! Leave the memory clock at stock; increase the core clock by maybe 20Mhz at a time, and stress for 15 minutes with FurMark each time, raising it another increment until it crashes. BAck it down, and start increasing the RAM clock maybe 50Mhz increments until it does the same thing. The rest of it you can discover in the OC'ing guides for these cards.

But before you do that -- figure out of your problem is either Witcher, the NVidia driver or something else before fiddling around with AfterBurner.

Furmark is not stressing my system hard enough. My gpu temp is only up to 70. When I game witcher i get to 76~78. when OC it will be around 80~82. When I OC and ran furmark it doesn't pass the 80c mark.

 

[BonzaiDuck]

Furmark is not stressing my system hard enough. My gpu temp is only up to 70. When I game witcher i get to 76~78. when OC it will be around 80~82. When I OC and ran furmark it doesn't pass the 80c mark.

Did you tweak the AA and other settings in FurMark to put a bigger burn on your card? Try 4x AA. Or 8x AA.

Heaven Benchmark is another tool, and it's pretty limp for generating temperatures.

Howsoever! Whichever stress-test you use, make sure you close down any other monitoring software (like ASUS Suite, ASUS Tweak, MSI Afterburner, HWMonitor, etc.) if voltage, temperature and other monitoring is presented in the stress-test software. Otherwise, you may generate a "false-positive" for instability . . . and . . . cause your own BSOD. [Back to that, again . . ]

Also, another point. As we'd discussed in the forums for CPU overclock stressing, you may indeed want to see how hot you can make the CPU (or graphics card) for assessing your cooling strategy and assessing worst-condition stability.

But it isn't necessary to determine stability for ordinary real-world situations. The advice I'd seen for final balls-to-the-wall testing of a stable overclock for graphics suggested using Heaven for 24 hours. For CPUs, OCCT: CPU is a good test for 4 hours. So is Intel Extreme Tuning Utility. But they don't push the temperatures as high as LinX or Intel-Burn-Test.

One more . . . I'm betting that one reason your temperatures are so high -- whether in Witcher or anything else -- derives from pushing your voltage slider all the way to 87 mV. For my 970 cards, that's the absolute limit. You shouldn't need to use any additional voltage to go considerably beyond the core clock you'd cited.

I can't speak with total authority about your GTX 980, because I don't have one. But my 2x GTX 970s should be the next best thing for comparison, and they're MSI cards, too. I expect that the 980 card will run hotter than mine. But it can't be by that much.

Thanks for calling my attention to Metro LL. I'll want to try it out.

 

[BonzaiDuck]

Here's something I found for just a few minutes Googling:

http://www.gamersnexus.net/guides/1680-overclocking-gtx-980-maxwell-tutorial/Page-2

It seems to suggest that you can reach ~1,450 Mhz without unlocking the voltage slider or increasing the voltage.

But you really need to get this other instability problem squared away. Do a couple marathon stress/stability verification runs with some good stress/bench tools. Stress-test your CPU. Let the system run for long periods at idle speeds to see if there is any idle instability.

Get all that sorted out, and you should actually get a much better clock for that card without even touching the voltage. That also means you'll get lower temperatures than those you cited.

Also, do you check your Windows Event Logs for errors and warnings? Some are benign; others need a little research so you can eliminate them. It's worth a lot to have a clean-running rock-solid system before you go ahead and tweak the graphics card.