Bad chip or bad cooler :(

[railven]

I redid my fan lay out and am now trying to OC my processor manually.

System Info:
Intel i5 2500k
ASROCK Z68 Extreme 3 Gen 3 (BIOS 1.3 - for some reason I can't get it to flash to the newer stuff, eitherway newer stuff is to support IVB mostly.)
Corsair Dominator 2x4GB DDR3 1600mhz 9-9-9-24-1N @ 1.5v
Coolermaster 212 Evo+ (stock fan)
Sapphire HD 7970 w/Accelero Cooler (1125/1625 @ 30c Idle/60c load)
Corsair Cabride 500R
Corsair 750W TXv1

Fans:
2x120 front intake
1x140 panel intake
2x120 chimney exhaust
1x120 rear exhaust

Hope that covers the system info, using the AutoOC'er option on the ASROCK I've safely clocked the CPU to 4.9ghz using about 1.39v with temps getting as high as 80-90c, which of course is too hot for me.

Using the same AutoOC'er at 4.6 I would still use some where in the 1.37v with about the same temps, more in the mid 80C.

Well, after some suggestion in the cooling department I swapped some fans and decided to go for a manual OC.

Using the following options (off the top of my head):
100 BLCK
44x Multiplier
C1E - OFF/C2 - Off/C -6 OFF/ C-Package - OFF
EIST ON
Speed Spectrum Off
RAM manually set (as above)
Core Offset 0.005v
Turbo Offset 0.004v
All other voltage Auto
PLL Volt - On/Auto

I get the following after 1hr stress test:

These temps still seem unsatisfactory. I can't even get the chip stable at 45x multiplier raising the boost offset to +0.020v - that should be more than enough for 4.6, right? It will climb up to 1.32v before BSODing.

I'm starting to think I'm hitting the temp ceiling, but I don't get why if before using the auto OC'er I hit 4.9ghz stable with almost the same temps and higher volts.

 

[Arkaign]

The voltage component section of the motherboard can degrade over time. Power supplies and of course processors ran at too high a voltage can also degrade.

To make sure things are still kosher with your components, set it back exactlly the way it was settings-wise, and see if you still get the "4.9" you had before with similar or hopefully slightly improved temps with your new fan layout.

 

[railven]

The voltage component section of the motherboard can degrade over time. Power supplies and of course processors ran at too high a voltage can also degrade.

To make sure things are still kosher with your components, set it back exactlly the way it was settings-wise, and see if you still get the "4.9" you had before with similar or hopefully slightly improved temps with your new fan layout.

Thanks for the suggestion. I'm going to try that when I get home.

I'll be strapping on a second fan to see if that helps. Maybe I'm doing something wrong with my cooler.

 

[moonbogg]

bad chip or bad cooler? Sounds liek you got a chad booler.

 

[Maximilian]

Im pretty sure its just that your cooler isnt good enough. That isnt a bad cooler for a mild lower voltage OC but for what you are gunning for you need something bigger.

 

[Kenmitch]

Adding the second fan to your cooler should help with your temps I'd think.

Not sure what BSOD your getting but sometimes 0x124 can be caused by PLL voltage being too high or low for a overclock....For clarification I'm not talking PLL overvoltage.

 

[buklau]

I have a similar set up like you do 2500k, hyper 212 evo (mx-2) and asrock p67 pro3 b3 motherboard. The temperature I'm getting is like 31-38c idle across all cores and 61-65c load with prime 95 running while the evo fan spins at 1250RPM.

asrock p67 setting

dram = 1866 9-9-9-27 1t @ 1.425v samsung 30nm 4x4gb
multiplier = 45x (can't hit 46x stable with any voltage)
vcore = manual @ 1.335v (1.336v load in windows cpu-z)
pll = 1.627v
rest = auto

short power limit = 150w (even increase both to 250+ still can't hit 4.6)
long power limit = 130w

core limit = 150

This old asrock board served me well for 1 year and I am thinking about getting another 7850 to do crossfire, so I bought another board asus p8p67 pro rev3 refurbished board on geeks.com for $75. To my suprise, with this asus board I was able to run 4.5ghz with merely 1.3v vcore where as the asrock board cannot.

p8p67 pro setting in bios

LLC = extreme
VRM = 110%

multiplier = 45x

CPU vcore = manual @ 1.300v
pll = 1.6000v

rest = auto

I was even able to run at 4.6ghz with 1.35v and 4.7ghz with 1.4v! But as soon as I reach 4.6, I noticed my load temperature increased by 8c and another 8c with 4.7.

4.5 @ 1.3v = 58-62c load (p8p67 pro)
4.5 @ 1.336v = 61-65 load (p67 pro3)

4.6 @ 1.35v = 62-69c load (p8p67 pro)

4.7 @ 1.40v = 69-77c load (p8p67 pro)

Even when I set my cpu fan speed to max 1900RPM, temperature only drop by 3-4c compared to 1300 (silent mode). The case I'm using is antec 300 with stock fans 140mm top exhaust and 120mm rear exhaust both running at low and my room temperature is about 80F with ac on during these summer times.

So the first thing I'd do is to clean out your old TIM with robbing alcohol, apply a fresh layer of TIM to your CPU in vertical line according to arctic silver 5 instruction, secure down your 212 evo cooler a little bit on all 4 sides, then tighten top right screw, bottom left, top left and then finally bottom right so the TIM will spread out perfectly. Now go back and re-test with prime 95 to see if your temperature improves. Other thing I'd looking into would be to clean out any dust accumulated on your fans and heatsink and be sure to keep your room temperature below 85F if possible or turn down the overclock when your room gets too hot.