[jebusv20]

I installed my XSPC RX360 water cooling kit yesterday without a hitch.
My temperatures at between 50 and 60C on the die, in intel burn test @ 1.4V.
I am getting BSOD error: MACHINE_CHECK_ERROR (or something worded similar)

My current motherboard is the
ASRock Z68 Extreme3 Gen3
and I am using the exact same settings as
http://www.overclock.net/t/1190228/i5-2500k-asrock-z68-extreme-3-overclock-help-please

I have noticed that the on-die GPU is still showing up on my computer, do I need to disable it for higher overclocks?
Also I was running a screen off the on-die GPU previously, I am not sure if that was the problem,
will run some more stress tests with it unplugged.

Please help me out if you have any ideas as to why my OC is so low. I know a lot people can get 4.5V on Air with almost stock volts.

 

[jebusv20]

has been stable so far with the screen not plugged in to the motherboard @4.4Ghz.
Not sure if the Ghz drop or the screen made the difference, still really need some help with this tho.

ps: bump

 

[RockThePylon]

Check RAM timings, speed, voltage, etc...

You're right, 1.4v for 4.5ghz is high. Unless you got the worst chip evar.

 

[jebusv20]

mmkay... tried again @ 4.5Ghz and it froze up in a burn test after a few runs, it seams to be stable sometimes and can do 8-9 runs
without BSOD, other times it freezes on the first. I am running my ram at the XMP profile 1, but I can try and loosen the timings
or the frequency and run it again.

Currently at 4.4ghz It feels stable,oh and... it wouldn't even post at 5ghz.

 

[jebusv20]

just tried 4.5Ghz with memory @ 1333, and it BSOD'd on the 3rd run of lynpack. Should I try a higher voltage? I have a feeling it would be pointless, and obviously useless when I am looking for a 24/7 OC

 

[MacNcheese]

Start again.

Set memory to the rated times when your overclocking CPU. Especially an unstable CPU. 1.4 is very high for 4.5, so stop and start again. With a weird BSOD message like that, I suggest going back to default settings and running stress tests then. You know your CPU is stable and your RAM is too, so if you get BSOD, its something else. If not, get to overclocking one notch at a time.

 

[daman246]

Is the memory it happened lots of times with me it. sometimes it run 5-10run of intel burn but sometimes if BSOD in the first run

 

[jebusv20]

thanks for replying, I ran IBT for about 30 runs last night on all stock, as I didnt plan on overclocking at night (when its cold).
and there were no problems, IBT just finished now at 4.4ghz 1.4V with ram back at XMP, and all the setttings exactly as they are in the link on the OP (with higher vcore)

 

[Silver_WRX02]

Try running stress test on stock setting first to see if system is stable.

 

[jebusv20]

mmkay, I am using the very high settings on IBT is that the problem? If it is the memory what should I do try and fix it, I have already tried a run at 1333mhz.

 

[MacNcheese]

Quote:

Originally Posted by jebusv20

thanks for replying, I ran IBT for about 30 runs last night on all stock, as I didnt plan on overclocking at night (when its cold).
and there were no problems, IBT just finished now at 4.4ghz 1.4V with ram back at XMP, and all the setttings exactly as they are in the link on the OP (with higher vcore)

That happen to me too, but my case was a little bit "better" per say. I needed a lot more voltage per 100mhz to go from 4.4 to 4.5 that it just made sense for me to stay at 4.4 rather than 4.5. Maybe 4.3 is your sweet spot? have you tried that stable to see what volts are required? you can probably make a chart for 4.2/4.3/4.4 and then pick the best volts. The performance difference is so marginal at that point that it is worth it to save your chip

 

[MacNcheese]

Quote:

Originally Posted by jebusv20

mmkay, I am using the very high settings on IBT is that the problem? If it is the memory what should I do try and fix it, I have already tried a run at 1333mhz.

Highest setting on IBT are the ones you use to seal the deal on the overlock... solve your memory problem by running it at stated speeds.

edit: soz dubl pst

 

[jebusv20]

just looked at coretemp and it is reporting my load temps as:
core #0: 64
core #1: 74
core #2: 74
core #3: 70

these seam aweful high for a water cooled chip albeit at 1.4V
my HDD and SSD (air is straight from the front) are reporting temp of 30C
I have a really high ambient, but maybe I didnt seat the cooler right to be getting 75s?

 

[jebusv20]

other than vcore and using the pll overvolt setting, are there any other volts that should be changed?

 

[jebusv20]

hey guys I am back from work, I am still not able to run IBT on very high above 4.4ghz.
I tried everything above to no avail, before work I used the preset overclocking tool to try 4.6 (see if it was THAT easy)
but again, BSOD while I was at work (no idea how long It lasted I had it set to run for about 2 hours)

I looked again at the EUFI and found that PLL Ovolt is disabled with the overclocking presets so I enabled it
and I am 90 runs deep at 4.6ghz in IBT running a standard test.

I am afraid to run very high and disapoint myself, in the case that It is unstable and I have some form of multiplier wall at 45
can I get some advice as to bumping to BCLK? If I bump it from 100 to 110 at 44, I can get a very reasonable OC (4840), not sure
what else I would have to fix tho.

 

[CapitanPelusa]

I have same board and same chip and i ran into the Machine Check Error and stability problems just like you. For me the fix was to try out differenct Load Line Calibration settings until i was able to be stable.

Start at 5 and test, then move to 4,3 etc until you are stable.

I remembered doing some other things ill try to replicate the error again and repost with what i did to get rid of it but LLC settings was a big part of it.

I didnt have to mess with RAM, i kept XMP profile.

Ill replicate error when i get back from work hopefully.

 

[Darkslayer7]

Post all your voltages ( Vcore , VTT , vccio ... ). Something ain't right here .
What king of ram are you using ? 1.5V or 1.65 , or even higher ?
How many ram sticks ...

 

[jebusv20]

thank you for trying to help

unfortunately I've been running with LLC at 1 or 2 all day.

 

[jebusv20]

memory is
Corsair Vengeance CMZ8GX3M2A1600C9 8GB (2x4GB) DDR3

so 2 sticks at 1.5V, I'll post other voltages, as well as octal code for BSOD error next time it dies.

mmkay, so it just BSOD'd after I posted this in the middle of its 3rd IBT run on very high.
BSOD code was:

MACHINE_CHECK_EXCEPTION
0x9c

Volts are all set to Auto with Vcore set to offset 0.05V
DRAM = 1.53
PCH = 1.06
CPU PLL = 1.83
VTT = 1.05
VCCSA = .925

the offset shows 1.44V in CPU Z with LLC at 1

 

[jebusv20]

Found this on another post:

"0x9C = Increase QPI/VTT, but increasing vcore has helped in some instances". Try adding a little bit of vcore. You may have a bad cpu so if vcore stabilizes it at stock then I would suggest RMA if you still have warranty."

VTT is at 1.05V now, what is 'increasing it' mean, is that 1.1V or 1.4?

edit: just bumped it to 1.15V as it is only a few steps above what I was getting before. Will do some research first.

PS:

this cpu is stable under THESE conditions

50+ runs of IBT on Vhigh
4.4Ghz @ stock volts
4.4Ghz @ 1.3V fixed

200+ runs of IBT on standard
4.6Ghz @ 1.05v offset