My CPU or RAM damaged?

[Dave3000]

I have an i7-4820k. Last night I was running Intel Burn Test with the High setting and on the 8th run my system froze. CPU core temp was 62C. I had my CPU set to turbo boost all 4 cores at 3.9GHz. Intel's only officially supports single core loads for 3.9GHz turbo for this CPU, so I was sort of running my CPU outside of specifications, but that's how my motherboard sets my CPU if I enable the XMP setting for my RAM.

One reason why I think my CPU may be damaged is because I set my board to Asus Optimized mode in the BIOS and it automatically overclocked my CPU to 4.5GHz with 1.456v vcore and I ran it that way for 2 minutes until I realized it was overclocking and overvolting my CPU when I checked CPU-Z. I didn't that that the Asus Optimized mode would overclock and overvolt my CPU. I shut down my PC and got a message saying "overclocking failure and press F1 to enter setup." I changed the settings back to XMP which does sync all cores to 3.9GHz then I ran Intel Burn Test to check if that overclock and overvolting that my motherboard did damaged my CPU.

Weird thing is that I overclocked my CPU to turbo 4.2GHz at the same vcore as when it was turbo 3.9GHz and it succesfully passed Intel Burn Test on High. So that makes me wonder if it really is the CPU that froze the system because it froze at 3.9GHz but not at 4.2GHz with the same vcore. I think that if it would freeze at 3.9GHz that one time in Intel Burn Test, it would freeze sooner and more often at 4.2GHz at the same vcore, assuming my CPU is the issue? Also this morning I ran Intel Burn Test with the stock turbo specifications of my CPU (on my motherboard, disable XMP, set AI overclock to Auto, and clocks on Per Core to Auto, and set my RAM manually). I ran Intel Burn Test at Maximum mode and it passed. Stock turbo specs for my CPU on 4-core loads is 3.7GHz.

Also I was running memtest 86+ 4.2 overnight and 3 hours 42 minutes into it my system froze, timer stopped and had to do a hard reset. Turbo is not active during memtest so that means I was running at the base clock speed of 3.7GHz. So this makes me think that something could be up with my RAM. I'm running it at XMP settings which is running as advertised.

Also I've been getting errors in Coretemp if I let it run for more than 10 minutes sometimes and Call of Juarez The Cartel sometimes kicked me back to the desktop, saying that it stopped responding but the Call of Juarez The Cartel problem happened before my motherboard automatically overclocked and overvolted my CPU with the Asus Optimized setting

I'm thinking it's either my CPU or RAM that the problem. It's surely not an overheating CPU problem as the crash happened at 62C core temp or at least it never got hotter than that during testing.

 

[PliotronX]

First try a few loops of Memtest on a thumbdrive and if it passes the RAM is good to go. Note that sometimes a memory controller or CPU error can cause an error in MT so it is not always the RAM. Run it at stock to eliminate other components. I think the motherboard would damage a VR component before the CPU would be damaged.

 

[Dave3000]

First try a few loops of Memtest on a thumbdrive and if it passes the RAM is good to go. Note that sometimes a memory controller or CPU error can cause an error in MT so it is not always the RAM. Run it at stock to eliminate other components. I think the motherboard would damage a VR component before the CPU would be damaged.

I did run memtest at stock CPU settings and stock memory settings. You need to be in Windows to have turbo boost working and memtest is not a Windows application and so my CPU was at the stock base clock speed when running memtest as I mentioned earlier.

 

[Dave3000]

I also had other RAM turn bad on me on this motherboard a couple of years ago and it would error out in Memtest within a minute. I hope my motherboard is not damaging RAM.

 

[PliotronX]

Ah sorry about that. Memtest can reveal strange things. When MT freezes it's usually the CPU in my experience but the motherboard can do this, mostly by fault of not supplying enough juice to the CPU or an incorrect bus related timing (strap level in previous motherboard chipsets). It is very rare for a motherboard to eat RAM but I have heard of it.

 

[Dave3000]

I had my CPU for 3 weeks at my system was stable other than the Coretemp error and Call of Juarez problem but I think that the Call of Juarez problem was a bug in the game because it happened mostly when loading up the game with the exact same error message but then later one I didn't have any problems loading the game but at one time when loading a level again with the error message mentioning the same file. And Coretemp would do the same error sometimes and stop responding while it was running. This was before that auto overclocking/overvolting incident yesterday. But the Intel Burn test system freeze and memtest happened after the auto overclocking/overvolting incident.

I'm thinking that maybe my CPU degraded enough to lose it's ability to run the turbo at 3.9GHz on all core loads or maybe it was never stable at that setting to begin with and I discovered that yesterday through Intel Burn Test? But then again memtest froze my system at stock settings with turbo boost disabled.

 

[PliotronX]

Your suspicion of IBT revealing a prior existing instability seems more plausible to me, I really don't think the CPU degraded even with 1.45v running through it for a couple of minutes. Degradation usually occurs over weeks or months, outright damage in seconds. So in essence it is not totally stable even at stock as it is? Have you tried running just one memory stick, maybe in different slots?

 

[tweakboy]

CPU is not damaged. Unless you were running it @ 105c on load.

These cpus can take 100c and still beat properly.


Mobo seems to be culprit

CPU is 105c all this time.

 

[Dave3000]

I'm running my CPU at stock settings and I can't get my system to crash on me now but then I could not reproduce the crash when running even at 4.2GHz with the same vcore as 3.9GHz. But then again Core temp stopped responding and Call of Juarez stopped responding when I was running my CPU at stock settings.

 

[aigomorla]

so wait... no working when not overclocked... working when overclocked?

 

[Dave3000]

I think that my chip just can't handle being turbo'd to 3.9GHz on all core loads, at least with the vcore on Auto. 3.9GHz is the official max turbo for my CPU. Intel Burn Test ran fine at stock CPU settings on the Max test setting.

 

[Dave3000]

I ran another Intel Burn Test again in Very High mode at CPU stock turbo settings and it passed the test. I'm still confused why memtest froze my system at stock clocks. One thing I can think of is that I had my Xbox 360 controller connected to my PC while running memtest and I disconnected it while running memtest because of the flashing light on my controller. I did the same thing before with my Xbox 360 controller while running memtest but it was on a different hardware and memtest crashed eventually, not responding to keyboard inputs.

 

[Dave3000]

Well I just ran Intel Burn Test in Very High mode with my CPU's turbo boost overclocked to 4.3GHz on all core loads at a manual 1.165v and it successfully completed the test. That's with the same vcore that Auto vcore gave me when I had all core loads at 3.9GHz turbo, which my system froze late in the test. If it was because of my CPU that I got the system freeze in the test at 3.9GHz wouldn't 4.3GHz at the same vcore give me problems in Intel Burn Test very quickly? I'm starting to think that my CPU is fine and it must be something else.

 

[Dave3000]

I just ran Memtest 86+ 5.01 overnight for 6.5 hours and it didn't freeze or error out on me. I also tested each stick separately in different DIMM slots and did 2 passes for each stick and Memtest didn't freeze or error out on me. I also just ran Intel Burn Test on Very High Setting for 20 runs and it passed. This is all at totally stock settings. I did remove my DIMMS and reinserted them before running these tests again, so maybe it was a stick that wasn't securely in the slot. One of the sticks did feel like it was slightly easier to take out than the others when I was removing the sticks.

 

[Dave3000]

Okay, today I replaced my motherboard with the same one and I was running Intel Burn Test. I am using Windows 8.1. First of all I each time I double clicked on the Intel Burn Test exe file it said that an error as been encountered and something mentioning missing MS Net 3.5 in the details so I downloaded MS Net 3.5 and finally could run Intel Burn Test. Then about 2 minutes into the test, I get a message on the screen saying that my PC as encountered a problem then my screen when blank and I had to do a hard reset.

All this was done with my CPU at stock settings. The reason why I thought it could be my old motherboard was because when I took my CPU out and looked inside the socket, there was what appeared to look like a tiny piece of crumb or dirt at the edge right next to a pin. It did not look like a bent pin. I could not remove whatever that thing was because of the pins in the way and it's such a small area and it's not loose and it's flat. I'm not going to risk using this board any longer. Could my CPU be damaged, if it is, from that foreign piece material in my socket while I was using it or could it be from that high overclock with the high voltage that my board chose to do when I chose a performance setting?

Was this just a Windows 8.1 issue with Intel Burn test?

 

[Dave3000]

About 1 hour ago I clean the area where the tiny crumby looking thing was in my CPU socket with Arctic clean and a old soft bristle toothbrush I don't use anymore and now that dirty or stain is gone. However, I noticed that the plastic where the crumby thing was is now chipped or a deep scrape, upon inspection with a magnifine glass and it's still noticeable if I look carefully without a magnifine glass. Maybe because of me picking that area with a pin and a toothpick to try to clean that stuff out of the socket. This time that area where it looks chipped is darker than the rest of the plastic. Pins don't look bent however. I don't know if I should go back to my old board and return the new one.

 

[Dave3000]

Also the reason why I'm hesitant to think that it's the CPU is because a few years ago I had problems with strange errors and crashing and it ended up being memory that turned bad but I exchanged the CPU and motherboard anyway, but it was the memory all along. Many years ago I thought I had a CPU problem because of crashing and it was a motherboard problem all along and I wasted 6 weeks waiting for a replacement CPU from the manufacturer when their was nothing wrong with the CPU.