Things I’ve learned, published for the public benefit
Hope This Helps header image

The right way to stress-test an overclocked PC

Consider the following situation: You’ve overclocked your CPU, set the core voltage (Vcore) to some reasonable number, and then it’s time for some stress testing to make sure your rig is stable. You follow all the standard recommendations that you can find on overclocking forums: you run Prime95 for 12-24 hours, do a few hours of FurMark, and a few hours of IntelBurnTest for good measure. All the tests complete without a hitch. Congratulations! Your system is now considered stable.

But then you run Crysis or Battlefield 3 and you get random lockups and reboots. What’s going on?

The problem is that you stress-tested the CPU and GPU separately. That doesn’t guarantee that your system will be stable when both the CPU and the GPU are under load. Interestingly, loading the GPU can make your CPU unstable!

In other words, to be stable under combined CPU+GPU loads, your CPU may need a higher Vcore than it does for isolated CPU loads. That’s why the Vcore you arrived at using Prime95 or IBT may be too low to guarantee CPU stability when the GPU is under stress.

Here’s the story of how I realized it:

I had overclocked an i5 3570K to 4.3 GHz at 1.25 Vcore. The system had successfully completed the following stress tests:

  1. Prime95 Blend (all cores) for 14 hours
  2. Prime95 Small FFT (all cores) for 8.5 hours
  3. FurMark for 3 hours
  4. IntelBurnTest on Standard for 2 hours
  5. IntelBurnTest on High for 1 hour

Most overclockers would agree that these results look rock-solid. However, when I ran IntelBurnTest and FurMark simultaneously, I was shocked to see FurMark fail almost immediately. I repeated the experiment several times and the time to failure was always between 5 and 30 minutes. I started coming up with hypotheses:

Hypothesis #1: The PSU can’t supply enough power for the combined CPU+GPU load. The Corsair HX520W supplies 480 watts on the 12V line; an overclocked 3570K + HD7850 shouldn’t draw more than 300 W, but there’s also the motherboard, and the PSU is kind of old – who knows, maybe the capacitors have aged?

Refutation: After replacing the PSU with the Be Quiet! Straight Power E9 580 W (564 W on the 12 V line), the symptoms were exactly the same. It wasn’t a PSU problem!

Hypothesis #2: The increased heat production is causing the GPU, CPU, or video card VRMs to overheat.

Refutation: (1) GPU and CPU core temperatures were carefully monitored during stress testing and did not exceed 90°C. (2) During isolated CPU/GPU stress testing, CPU/GPU core temps were the same, yet there were no crashes. (3) With an open case and all fans set to maximum, the CPU+GPU test failed just as quickly. It was not a heat issue.

That left me with only one hypothesis.

Hypothesis #3: Loading the GPU is somehow causing the CPU to lose enough power to become unstable. In other words, the CPU Vcore is set high enough for isolated CPU load, but not high enough for combined CPU+GPU load. This was a novel hypothesis; I hadn’t seen it discussed in any of the overclocking resources that I had read.

Confirmation: In the words of Colonel Hans Landa, that’s a bingo! Upping the Vcore from 1.25 V to 1.275 V improved the stability by a large margin. Instead of failing after 5-30 minutes (as verified in multiple tests), IBT+FurMark failed after 108 minutes. (Notice that if the crashes had been due to overheating, the test would have failed more quickly than before, as increasing the CPU Vcore made the CPU hotter.)

Still, the system was not fully stable. In order to make it IBT+Furmark stable for several hours, I would have probably had to increase Vcore to around 1.3 V. Incidentally, this is close to the Vcore that the CPU chooses automatically for a 4.3 GHz clock speed, if you select the so-called “offset mode” instead of forcing a fixed Vcore. The lesson here would be that Intel knows what they’re doing when choosing these automatic settings.

The right way to ensure your overclock is fully stable

Run FurMark and IntelBurnTest at the same time for a few hours – at least as long as you expect to operate with combined CPU+GPU load. For example, if your typical gaming session is 3 hours, you should run IBT+FurMark for at least 3 hours. Of course more is better.

One tricky part to combined CPU+GPU stress testing is that FurMark needs some free CPU cycles in order to stress the GPU properly. If you fully load the CPU with IBT, Furmark will be bottlenecked and the GPU load will be much less than 100% – on my system it was around 75%.

The solution is to find the right settings for IntelBurnTest that will result in stressing the CPU while leaving just enough free CPU cycles to allow FurMark to get GPU load above 95%. In my case, the highest combined load was produced with IBT set to “High” and 3 threads.

Another piece of advice is to stick to “offset mode”, i.e. let the CPU set the Vcore automatically depending on the clock frequency. This may result in rather high voltages and high power draw (for example, a 3570K @ 4.3 GHz and full load has a Vcore of 1.32 V, which is higher than usually reported by overclockers at this clock speed), but at least you’re using a Vcore that has been blessed (and presumably extensively tested) by Intel.

My personal choice was to scale back my overclock to 4.1 GHz and choose offset mode. This translates to a Vcore of only 1.192 volts under load and, as far as I can tell, complete stability (IBT+Furmark 5 hours with no error) – though of course you can never be 100% sure…

FAQ

Isn’t it overkill to stress-test the CPU and GPU at the same time? That depends. If you’re absolutely sure you’re never going to reach full or near-full CPU+GPU load, then I guess you can stick to standard, isolated stress testing. However, bear in mind that there are games which can place a very high load on the CPU and the GPU at the same time.

Are you saying that most overclocked systems out there are not truly stable? It would seem so. Most overclockers determine a “stable” CPU Vcore based on isolated CPU testing with something like Prime95 or IntelBurnTest. My experience shows that you need a much higher Vcore to ensure stability under combined CPU+GPU load. I suspect most OC’d systems would fail the IBT+Furmark test outlined here rather quickly; they might also crash when running certain games. Of course, my findings would be more trustworthy if someone replicated them – please post your experiences in the comments!

Why would loading the GPU make the CPU unstable? I’m no hardware engineer, but here’s a wild guess: Loading the GPU drains some of the power away from the CPU (or causes small voltage dips from time to time). Therefore, when the GPU is loaded, the CPU will need a higher Vcore to remain stable.

What was your exact setup?

  • Motherboard: Asus P8Z77-V Pro
  • CPU: Intel i5 3570K
  • CPU cooling: Noctua NH-U12P SE2, 120mm Enermax T.B. Silence fan
  • Video card: AMD Radeon HD7850 (MSI Twin Frozr III, 2 GB)
  • GPU cooling: Accelero S1 Plus, 120mm Enermax T.B. Silence fan
  • 16 GB RAM (two 8 GB sticks, Kingston HyperX)
  • SSD: Crucial M4 256 GB, HDD: Western Digital Red 2 TB (WD20EFRX)
  • Case: Fractal Design Define R4
  • Case cooling: 1 rear 140mm Fractal Design Silent Series R2 fan
  • PSUs tested: Corsair HX520W / Be Quiet! Straight Power E9 580W

26 Comments so far

  • bluewoodtree

    If I have a “basic” set-up I would dare to overclock my PC. From my experience overclocking is just temporarily useful, I think on the long run it will decrease the lifetime of your system where there is no need. I would only overclock my machine in order to bridgeover a foreseeable upgrade.

  • Kurare

    Having same trouble, read and liked!

  • LocoGringo

    thanks for the different perspective this was a ton more helpful than scrolling forums all day

  • DeOmZ

    Nice tip we have here..I’m now trying this test for almost an hour now and still no errors.. Does it mean that it’s stable?

  • harry

    Nice article dude! on what you had written, i just passed every stressed test.. after a day playing high intensive graphic game, like battlefield 3 my system crashed.. got blue screen! omg.. stabiity test is not 100% reliable.

  • Velmurugan

    Great Man Thanks For sharing and Can u advice something for my Problem below:
    OC my C2Duo to 3.4GHz but not stable.

    [PROBLEM]Comp restarts after blend tests(Prime95) after 1.5 hours

    MObo–>Asus P5G41-MLX
    Processor—>C2Duo E7300 stock(2.66GHz)
    FSB Freq—>343(CPU not Posting More than That)(FSB WALL)
    DRAM Freq—>1029 OR 1372 OR Auto (Only 3 options)

    NB voltage—>[auto] —>CANNOT FIG out THE AUTO VOLTAGE(EVEN IN Hwinfo32,Cpuz monitor Etc.)

    Memory Voltage—–>1.86 volts
    VTT voltage—–>[auto]—>cannot fig out

    Vcore Voltage—>+50mV OR+100mV OR +150mV—.Using +150

    RAM: Coarsair 4GB DDR3 1600 (PC3 12800)

    CPU fan—>Using Stock Fan

    Normal Temp–>57 Deg.celcius
    During Load Temps—>88 deg.celcius

  • itsalways

    88 degrees? That is danger zone for temps. check your cooling. anything over 70 is bad. your pc is probably shutting down to protect the cpu.

  • Ben

    Cough cough, AIDA64 Extreme Edition does this for you..

  • Protec

    I’m not the most PC-tech sawwy guy..but i have managed to overclock my i7 950 @ 4,2ghz. I think the best way to test for stability is using your computer. Do gaming, watch movies, what ever. If it handles your everyday tasks without crashing = system is stable. I always run the IBT and Prime95 then i started a game: Instant Crash. I mean instant. Put some more voltage = stable. So the IBT and Prime95 tests alone does not mean jack.

  • Anton

    Hi, I like your thinking and while I agree with your assumption, I do not agree completely.

    While cpu stress and gpu stress tests suceed individually, but fail together, this means it is a failure of stable voltage supply. Not all psu’s are made equal and some allow for voltage drop and noise on the rails which will certainly cause the effect you are experiencing, sometimes only a little can cause a big issue. Yes an increase will solve this issue but I feel you are only hiding the real problem.
    Note that the psus you used are of high standard but are not immune to this.

  • Sven Bent

    running IBT with reduced threads is a silly way to give up cpu resources to furmark.

    Use proper cpu priority instead. Run IBT on idle cpu priority should make it use near 100% of all cores but still give furmark all is needed cpu resources to run at full speed.

    furmark alone = 38fps
    furmark + pime95 at 3 threads = 38 fps 75% cpu utilisations
    furmark + prime95 at 4 threads = 38 fps 96-98% cpu utilisations

    Prime95 runs in idle as default as you can see no loss on furmark form using all threads in idle priory but a lot better cpu utilization

    Furnmark alone 38fps
    furmark + IBT Vhigh at 4 threads normal = 21fps 98-99% cpu utilisations
    Furmakr + IBT Vhigh at 3 threads normal = 38fps 74-75% cpu utilisations
    Furmark + IBT Vhigh at 4 threads idle = 38fps 97-98% cpu utilisations

    above cpu utilization is just prime95/IBT not total cpu utilization.
    So yeah dont cut down threads just learn to use proper cpu prioritization

  • jashnioshis

    i will check this… been using OCCT for combined stresstests and PSU checking too… i still think this should be a PSU or motherbosard issue, or both at the same time… probably some vdroop happenening that should be adressed with LLC if the motherboard and psu are up for it.

  • samabaikbone

    I too find the same thing. Xeon L5408 @ 3.2 ghz pass prime95 but Crysis Warhead crashed in just 1 minute play. Upped Vcore from 1,33 to 1,36 volt solved the crashed. Strangely in that game, cpu usage hovering between 50-60%.

  • Randy

    Nice article! I’ve always wondered about this. I pass Prime, then launch a game, CRASH! Why?? Although a better PS may minimize V drop, a better solution is to up the V core voltage. After all we are talking about using our existing equipment here… (If I had the money to buy a high-end PS, I might as well buy a faster processor, thus eliminating the V-drop issue due to more stable Mhz running at less power.)… It amazes me that one cannot find this concept in any of the forums! As much as those self-proclaimed experts clamour on about this and that, they miss this essential hardware stress issue! Thank you for helping me to see the FULL PICTURE, reality.

  • nintendomaniac64

    For proper Prime95 stability testing, you really should run for 24 hours with a priority setting of “10”.

    Also the settings you choose can make a big difference. The following thread contains two of the commonly recommended settings:
    http://www.overclock.net/t/838244/prime95-a-quick-dirty-guide-to-the-custom-settings

  • Sven Bent

    AIDA64 does not stress cpu even close to what linpack or prime95 does. it can run for hours where prime95 and linpack shows errors after 5 mins. tested on multiple intel and amd systems with provoked errors by lowering voltage

  • nintendomaniac64

    If you’re really paranoid, one big thing is that, if your CPU has an integrated GPU (like your 3570K does), then for stability testing you really should remove any discrete graphics cards and stress-test both the iGP and the CPU at the same time.

    For the iGP I personally use the ‘Graphics stress test’ in the “Intel Extreme Tuning Utility”, while for the CPU I use the the “x264 v2 stress test” from Overclock.net set to normal priority and twice the amount of threads than my CPU supports (so for a 4790k you’d set 16 threads).

  • nintendomaniac64

    Got an update to my method – while running the Intel graphics stress test and the x264 stability test, open a modern web browser and have it run the title screen demo loop of something like Sonic Spinball (Master System version) on Archive.org’s in-browser MESS emulator.

    Doing something like this is key for rock-solid stability because I ran the Intelg graphics stress test and the x264 stability test all night without problem, but not even 30 minutes into running those two PLUS Sonic Spinball, sure enough x264 would crash (which is what happens when you’re unstable).

  • Rawr

    Why did you get a separate cooler for the 7850? It’s a low power consumption graphics card. In fact, I’m surprised you got such a weak graphics card (the 7850 is for budget gamers) consideing all of your other components are far more powerful and more expensive.

  • GuL916

    Hi, that’s a very interesting article.

    “Hypothesis #3: Loading the GPU is somehow causing the CPU to lose enough power to become unstable.”

    Just a quick note to tell you that in my case, on a low cost motherboard, that was memory that became unstable, probably due to the same explanation. Increasing memory voltage a little bit solved the problem.

    Thanks for the idea !

  • Kyle L

    I totally agree with this article. I recently purchased a core i7 6700K and a new ASUS Z-170A motherboard. I was reading these unreasonable values for overclocking people were getting. 1.28-1.3 Volts for 4.6Ghz. Meanwhile my prime95 was failing with Blend test until I got to 1.35 volts (which is what is set automatically for 4.6GHZ in the BIOS. Interestingly enough, 1.28 volts worked for me at 4.6GHZ doing burn test, and ROG benchmarks. Prime95 was the only one that complained and most people on their OC guides didn’t even use prime95 saying that it created an “unrealistic workload”. But to me, if it’s not stable under ALL programs, it’s not truly stable. It seems that the “perfect storm” situations, such as stressing RAM/CPU/GPU is where you really need to make sure you are stable. Great article!

  • Ross s

    I Just want to state that depending on the architecture it can be the extra load on the CPU data wise from the PCIE bus and not the extra power draw being “stolen” I mean this happens with RAM overclocking as well where you have to up the voltage of the memory controller but thats a lot easier to detect.

    In modern chips where the “northbridge” is integrated you can sometimes achieve stability by increasing that voltage or the PLL between that and the other subsystems. Instead of just the cpu voltage.

  • sulko

    Hi,

    different story here,

    I have R52600 @ 4.1GHz 1.325V with GTX 1060 6GB (+200/+500Mhz – Core/Mem),

    thing is when I load Furmark I have 75% GPU load, but when running Furmark together with Prime95 (heavy) GPU load drops to 20-35% – anyway its stable, not crashing.

    Power consumption of this build is some 330W max, PSU is CORSAIR CX600 (aged, but there should be plenty of reserve).

    any ideas?

    Thanks

Leave a Reply to LocoGringo