BSOD when stressing CPU out of the blue

I've been having problems with my last two motherboards so I got a new one (RMA) a month ago. On the last motherboard a mosfet burnt through when powering on. I am thinking this might have damaged my CPU.

My current setup is

i7 860 - Scythe Mugen 2
Asus P7P55D-E
2 x Corsair Vengeance 4GB
Gigabyte GTX570 SOC
Samsung 830 256GB
Intel Postville X25-M M80GB SSD
WD 2TB / WD 200GB / WD 60GB
Corsair CX 600W

Never overclocked.

I got the new motherboard a month ago. Since about two weeks I have been getting BSOD's (0x00000124), getting worse and worse. I play a lot of Battlefield 3, but I couldn't even stay in the game for longer than 5 seconds without BSOD-ing.

I read that BSOD 0x00000124 usually means decrease/increase VTT or Vcore. I tried upping the Vcore to 1.29, which I saw helping someone else with the same sort of issue. I ran Prime95 but my PC instantly shutdown, going into thermal shutdown. I tried to bring it down gradually but still would get extreme temperatures.

Eventually I settled for 1.1V Vcore and 1.15V VTT (IMC). They were both only slightly higher than my (varying) default settings and would keep the temperature around max 80-82 when stressing. I really couldn't go much higher.

I ran Prime95 and everything stays stable. Battlefield 3 works now without BSOD-ing, but the problem is it runs really bad. It seems to skip frames, even though I have a steady 60-70fps. It jitters and feels really laggy, like rubber banding, but it's not gfx card related. It really seems the CPU influenced it. My gfx card is totally fine. Here are two examples (1080p):



My question is, is my CPU slowy dying? It should be able to get a lot higher than 1.1V without getting really hot, shouldn't it? If other people can get up to 1.3V orso. Could the previous motherboard mosfet burnout have caused the CPU to die? Or has my new (2nd hand RMA) motherboard slowly died, not being able to give proper voltages?

I am really doubting if I should get a cheap i7 replacement, a cheap motherboard replacement or get Ivey Bridge to pull me over to Haswell.

I really hope for some input. Thanks!

It should be able to get a lot higher than 1.1V without getting really hot, shouldn't it?

Click to expand...

What cooler are you using? How old is the fan and when was the last time you thoroughly cleaned it?

It's a Scythe Mugen 2 Rev B, had it since I bought the CPU three years ago. It isn't dusty though at all. I cleaned the base when placing it on the new motherboard.
Even though the system gets hot, Battlefield 3 still should run normally and not have these stutters all of a sudden I would think. Couldn't it be a sign that the CPU is broken?

Is the current clock speed at the default of 2.8GHz, and do you have 2 x 4GB of RAM or 4GB total? Do you have Turbo and / or HT enabled? Also what's the model of RAM, including the current DRAM frequency, timings, and voltage?

Hi redduc900,

Yeah the current clockspeed is default, with Turbo and HT enabled. I have 2 x 4GB Corsair Vengeance (CMZ8GX3M2A1600C9 4GB PC3-12800U DDR-1600)

I am still a bit of a noob when it comes to timing settings etc, so I've made some screenshots.

I'd probably start by bumping IMC voltage to 1.25V, taking DRAM voltage off Auto and setting it to ~1.56V, setting Vcore to 1.20V w/ Turbo enabled (can be lowered later if need be), and adjusting Timing Mode to 2N w/ DRAM Frequency manually set to DDR3-1600. Also disable PCIE and CPU Spread Spectrum, and enable LLC (Load-Line Calibration). You could also try setting Ai Overclock Tuner to XMP, and set OC Tuner to either Good or Better Performance.

Cool, thanks. That's a lot I can go on from here. Got a busy evening up ahead, but I will post back results asap. I hope the Vcore bump will still maintain OK temperatures, since my last try made them go pretty far up. Or will the other settings compensate for that?

May I ask, what are your thoughts behind these settings? Will the system be better balanced and maybe make Battlefield run normal again, also without BSOD-ing?

Ok I set IMC to 1.25V, set DRAM to 1.5625V, set Vcore to 1.20V (Turbo still enabled), set Timing Mode to 2N and set the frequency to DDR-1600. Also enabled LLC and disabled CPU/PCIE Spread Spectrum.

When I booted I already noticed fluctuating high idle temps (48-56C!). On default I idle around 39-42C. I ran Prime95 and within one second the PC shutdown in thermal protection.

It would seem that the set voltages usually shouldn't cause these extreme temperature rises, or at least not make it go into thermal protection? I still really don't think it's the cooler though, since it ran fine for 3 weeks after placing the new motherboard and then started giving the BSOD 0x00000124, which isn't temperature related.

I just tested Prime95 again on default BIOS settings and I instantly got BSOD 0x00000124 again.

It really seems a slight bump in Vcore and IMC is the sweet spot for not BSOD-ing, but then I still have a system which can't play games properly and ofcourse isn't running as it should be.

Looking at the mosfet burnout on my previous board, maybe that really did give it a shock. Another thing what happened is that I accidently dropped the CPU from about 10CM holding it above the table when installing it on the new motherboard. Maybe this has played a part as well. Could these things combined also lead to my occurring problems, giving higher idle temps, BSOD's, not being able to up the voltages too far etc?

I must note that Battlefield 3 always seemed slightly jittery, not as smooth as could be, even on my first motherboard a little over a year ago. Same goes for FarCry3 more recently. Graphics and SSD seemed in optimal form, so I was always wondering how the unsmoothness would happen. I had enough frames per second, but it just wouldn't feel smooth. Maybe my CPU was already slowly decaying before?

[edit] I set the BIOS back to default and then set Ai Overclock to X.M.P. which left my Vcore on default, but set IMC to 1.20V, DRAM to 1.50V, freq to DDR-1600, Timing to 2N. Also set OC Tuner to Better Performance. I ran Prime95 for a short while and it seemed ok with maximum temps of 85C (still high). Then I ran IntelBurnTest on Very High but I barely made it past 2 runs and got the following error. Before with the small Vcore bump to 1.1V and IMC to 1.15V I was able to do 10 runs on Very High, but would get the same error on Maximum after a few runs. Also high temps as well (around 90C!).



Do you have any more thoughts on the whole situation or something I could try?

If you're sure the HSF is mounted correctly, temporarily disable HT and Turbo then re-run IBT.

Just to see if it passes?

Been playing BF3 for 1.5 hour now, CPU temperature just hit 90. In BF3 the CPU isn't at 100% load, but maybe the GPU also helps heat it up. I just don't get it.

[edit] PC just shutdown by thermal protection. Was 89C at the time. Now after reboot, the CPU temperature is going down really slowly, still stuck at 60C atm. GPU runs hot as well, but was able to test it properly with Furmark without any issues.

Is it actually possible that a broken CPU or motherboard can cause these extreme temperatures?

Check to make sure the heatsink is mounted correctly thoose temps are way to high!

Did you apply new TIM when you replaced the motherboard and intalled the cpu cooler?

It worked ok for three weeks though, untill the BSOD's started and I had to up voltages. Also the idle temperatures are still around 40C. I don't think they would be if the heatsink wasn't mounted correctly.

Yeah I cleaned the cpu and the cooler with 70% alcohol before applying paste when I swapped motherboards.

My guess is somehow the cooler came loose because even with a core voltage of 1.2v it should not be running that hot. Are you 100% positive the cooler is making good contact to the cpu?

I would recommend taking it off again, reapplying TIM and reseating it.

Do you know for sure that the CPU heatsink fan is spinning? That's sounding about right for a non-spinning fan.

Ok, I ran IBT again with HT and Turbo turned off. It seemed to be going in the right direction. My temperatures didn't go past 74C and averaged around 69C, almost a 20 degree difference with the previous test, a lot lower. Even still, it failed shortly after the third run



What conclusion can I take from that then?

peanutbudder said:

My guess is somehow the cooler came loose because even with a core voltage of 1.2v it should not be running that hot. Are you 100% positive the cooler is making good contact to the cpu?

I would recommend taking it off again, reapplying TIM and reseating it.

Click to expand...

I'm not even able to reach 1.2V without it going into thermal protection. Running on auto atm, I think the BIOS said it was around 1.088V.

The cooler seems properly mounted and evened out (the screws are all in place). I can't actually see the base though from looking into the case.

I will still try and take it off and re-mount it with some new paste, but I doubt it's going to change anything.

Bobnova said:

Do you know for sure that the CPU heatsink fan is spinning? That's sounding about right for a non-spinning fan.

Click to expand...

Yeah the CPU-fan is spinning, I can see it. Speedfan also says it's spinning at 1383RPM.

And I ran it again. This time with the case open. I let the system cool down, went down to 29C (my case builds up a lot of heat I guess). I ran it, temperatures were quite a bit lower than the previous test, but again, shortly after the third run it stopped again.



Didn't go above 58C at full load.

How can the error be temperature related then? There must be something else happening.

What case do you have?

An Antec Sonata II

I tried a game of Battlefield 3 with Turbo and HT disabled. I already had set all video settings to low before.

This is how it runs now:

(1080P)

It's even more unplayable than before. I don't think disabling HT and Turbo would have this effect for other people. It's actually worse than the vid shows, but it was just a quick capture. It really jerks and lags behind badly. It seems this only shows even more there is something really wrong.

Temperature stays fine since I've put it on low settings, the GPU stays a lot cooler as well, but it's unplayable.

Boot to Memtest86+ installed to a thumb drive, and check the RAM for stability at the modules current settings. If Memtest passes, bump the IMC voltage to a temp. level of 1.30-1.35V, w/ Vcore still at 1.15V and all other voltages on Auto (HT and Turbo disabled).