Overclock suddenly fell apart - 4930k - ASrock

C

clayton006

[H]ard|Gawd
Joined
Jan 4, 2005
Messages
1,089
So I'm referring to the main system in my sig (4930k, ASrock extreme 11). I booted up my system after a few weeks of sitting and it is very unstable.

Ran memtest overnight, did 2.5 passes without issue (overclock on, but no men O/C as I don't run one). Updated the UEFI, put everything back to stock / auto (auto voltage gets set at 1.3v and with LLC at 5 it goes down to 1.22v under full load) and I can't get this thing stable at stock.

OCCT fails on a core within 10 minutes, prime95 ran a bit longer and locked up.
Temps are a bit warmer now that my Titans are in the loop, but they are acceptable (all under 60c), I probably need to throw another pump in there and larger 2nd rad (I have two now).

One thing I did do was lock in the core multi to 39. Previously this board on auto would only turbo to 3.6 which was odd. I haven't tried auto with the new UEFI update.
This CPU was doing 4.6 @ 1.3v with no issues last time I used it.

I kinda hate this ASrock motherboard to be fair, I'm thinking that this is probably the culprit. I believe should go with the Asus Rampage IV Black Edition and get a solid performer in this area.

Any thoughts?
 
you know many times you pay through the teeth for fancy buttons on the top end boards, most of the mid end do quite well example, my M5A99X EVO gets extremely close results to the sabertooth and ROG based 990FX boards and was a good $50-$120 cheaper at the time, but hey I didn't get fancy buttons :(

anyways, A just cause it is liquid cooled does not automatically mean it will be more stable, if anything that means folks usually clock them that much higher with more voltage which will make them fall apart sooner.

Liquid was just the cpu, or full cpu and board?

Ample case aifflow?

It could be that you need to bump voltage one notch or reduce speed one notch from where you had it, or you got unlucky and your chip finally said screw you
 
Well I'm working on it stock right now. The GPU's and CPU is under water. The VRMs are being cooled by fans as well as the sink over the PLX / LSI chips at the bottom.

Running at stock and still having issues tells me either CPU or Mobo has had it. I'm hoping this board has had it because I can't stand it.

Based on the fact that I haven't pushed the voltage past 1.37v for 24/7 operation I doubt I fried the CPU.
 
clayton006 said:
Well I'm working on it stock right now. The GPU's and CPU is under water. The VRMs are being cooled by fans as well as the sink over the PLX / LSI chips at the bottom.

Running at stock and still having issues tells me either CPU or Mobo has had it. I'm hoping this board has had it because I can't stand it.

Based on the fact that I haven't pushed the voltage past 1.37v for 24/7 operation I doubt I fried the CPU.
Click to expand...

Que? I'd say you fried the CPU with that voltage.
 
The 1.37 was for a few benchmark runs. 24/7 voltage is around 1.33v. I didn't think that was an issue on this chip?
 
It seems to be passed around freely that staying below 1.4v for 24/7 on water (or as long as temps are okay) is the rule of thumb here. If there is a chart somewhere else that says otherwise I'd like to see it.
 
1.35v is max intel seems to recommend air or water means nothing really electrons can and still do damage. 1.4v is max that users recommend.

Thing is its an electronic circuit, just cause the average chip does not fry with x voltage does not mean another will not. I doubt this is the issue though.

Another thought, the QPI or whatever it is called that controls the IMC how much voltage was it giving, I have read on improper use of XMP that it sets to high a voltage and this can kill the chip or at least make it quite unstable overtime as from first gen I series till now they are all tuned slightly different and as a result they do not all have the same settings, again just a thought.

24/7 under liquid with active cooling at 1.33 is fine as long as temps were good, 1.37v I do not know as Intel chips by and large are more prone to electronic migration then AMD chips which just kind of die outright lol.

Extreme speeds high voltage can and does kill chips, I do not know the average for 6 core Intel chips in regards to expected speeds/voltages etc but I have read many a thread on said chips burning out seems moreso on liquid then air.

Hopefully its just the motherboard, but that is the prime guess for me, maybe that QPI thingy has to high a voltage set and if its anything like AMD the IMC does not at all like a high voltage on it with a low core voltage this is asking for trouble.
 
This has little bearing but the XMP setting for my Crucials have two settings. 1 setting being 1600 cas-9 at 1.35v (vDIMM) and another with 1.5v. I use the 1.35v profile.

If the 4930k died a great death at a few runs at 1.37 then I must have had a very weak chip, or a horrible mobo. I do have a replacement plan on the chip... but I'd hate to use it if I don't have to. I guess at this point I need to start swapping out components.
 
the memory controller and the cpu are tied at the hip for all modern processors, so it does have bearing. low voltage one core but high voltage/speed on the other can make a wicked hotspot and progress electro-migration and such effects, IMC is a portion of the actual die, to hurt this WILL hurt the cpu even if the actual core is not "hurt"

Few runs at 1.37 if it is above Intel spec as in they tested above 1.35 and the vast majority of cases it dies long before their recommended life span of say 5 years at nominal speed/voltage. Chips are extremely well engineered this is why they give the recommendations that they do for temps, voltage, clock speeds etc, which is also why they do give a decent overhead cause they generally test the crap out of them, also why we can or are able to overclock them, cause they do test quite a bit as there does have to be +/- to some degree.

More then likely it is the motherboard I would imagine as generally speaking if say rating was 1.35v max at x clock and you exceed this voltage to say 1.4v you might lose x% of intended lifespan in 90% of cases or some such thing hard to say as from what I recall based on the makers guidance (AMD, Intel, Nvidia and such) try to get a ~10 years at their ratings(for the actual chip not the vregs and such obviously) which is quite in the ballpark if cooling is maintained and stable voltages etc, so for most folks I think ~7 years if all goes well which seems to be about right, plenty of folks I know have their old systems still fully functional for this long.
 
You must log in or register to reply here.
Back
Top