• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

2500k 4.5 oc - first bluescreen

B

ButterB

n00b
Joined
Jan 26, 2012
Messages
3
okay... first bluescreen. This is my first attempt at overclocking. I did as much research I could to see how to oc the 2500k and it seemed easy enough.

I have run multiple intel burn tests different days, highest was low 70's, all okay. I have run 4 prime 95 tests each 8+ hours and one at 16, with no issues, again on multiple days. Its been okay for couple of weeks playing BF3, Skyrim and others with no crashes. Tonight watching a video on steam and it locked up. Restarted, ran intel burn test and then bluescreen on pass 7. grrr thought I was good.:mad:

I have used these forums for researching and never had a need to post before. Here for my first question. So running the numbers by you all to see if there is anything I should look at changing first.

2500k @4.5
Gigabyte z68xp-ud3p
hyper 212+
Sapphire 6870
8GB Vengeance ram at 1333, no oc

bios f5
multiplier x45 @100
multi-steps load line - lvl 5
vcore 1.3
qpi/vtt 1.15v
cpu pll 1.815
everything else in bios is auto/default. Those above are the settings I have adjusted.

Link To Google for Screenshots

I hope I posted enough info to get some feedback. If I could post anything else to help give anyone more info.. let me know I will try to find out. Thanks.
 
Try upping the Vcore first. 1.3 is a little on the low side for 4.5. You can also try turning on Internal PLL Overvoltage.
 
I would agree with forceman as well. I have a very similar MB as well as a 2500K. My chip, and what I have seen from a lot of other people, will go up to a certain speed with no voltage change. But at a certain point, I have to raise the voltage a bit. Mine will do 4.3 on all stock, and then to get 4.7 it takes like 1.375 with the LLC at 7, which makes the effective voltage under load 1.41. That might not be the exact voltage, but I know its close.

Don't be afraid to go to 1.4, that is plenty safe as long as you have good cooling, and honestly I see a lot of people going beyond that a bit even. Like I said as long as the temps are ok, it shouldn't be an issue. I just have a H50, and at 4.7 I will hit about 56 C under stress test. I see you have air, but I think that is a pretty decent cooler, so you might be able to do it.

Just start slowly raising your voltage and doing a stability test. You can do just like 30 minute or something to just get a preliminary stability test. Then what I usually do if it passes a short stability test at a speed I'm happy with/aiming for is run a full stability test. Prime 95, Linx, Burn Test, etc. Whatever tools you use.

Just have to use a little trial and error to find the best combination of speed/temps/voltage that you are comfortable with and that is completely stable. Good luck, and report back to us with your results. :)
 
Another vote for vcore increase. Try 1.36. That works well for me. Also a slight increase in on the ram might help as well.
 
Thanks for the replies. I haven't had much of a chance to sit down since and play with it. I upped vcore to 1.36 and was able to back it down to 1.33. It seems to run stable but the temps are higher than I want. I downloaded real temp and core temp programs (I think that are the correct names-I'm not at home right now) because I have heard they provide better temp readings. The IBT is hitting 82 C and prime95 after 2-3 hours is hitting 72-75 C. Unfortunately, this computer room during the summer will rise another 10 C or so... I am going to play around and see what I can get with my temps being in the 60-70 C range. I may have to revisit my installation of my 212 cooler, perhaps I didn't get the AS5 applied as good as I could.
 
Does this board have offset CPU voltage increase capability? If so I would definitely recommend going that route. It's by far the easier method to OC a CPU, as well as maintain all CPU power saving/idle features. My 2500k @ 4.4Ghz but idles at 1600Mhz.
 
Those temps aren't terrible. A little higher than I would expect at that voltage, but considering gaming will probably be 10C below your Prime95 temps, I think they are fine for long-term use.
 
P95 is best run in 24HR cycles for error checking on long term stability on voltage/temperature and cooling system capacity.

I suggest more Voltage on the CPu Core & running Memtest X86 for RAM testing.
 
Since it primes for so long but crashes inexplicably when watching a video, is speedstep/eist/c1e on? Disable those and see if it remains stable.

For my chip @ 4.5V, it would run at 1.3V easily for hours on load but on idle, the CPU would drop itself to 1.6GHz and 0.07V or something ridiculous and bluescreen when watching Youtube.

To counter this, I used negative offset voltage control for 4.5GHz and manually locked 1.4V for 5.0GHz.
 
So I haven't looked at offset voltage to much. Does this work in conjunction to LLC or in place of it? I am looking it up but just trying to understand how it works. Yes this board has it listed as DVID.
 
ButterB said:
So I haven't looked at offset voltage to much. Does this work in conjunction to LLC or in place of it? I am looking it up but just trying to understand how it works. Yes this board has it listed as DVID.
Click to expand...

No, offset just says: "whatever the voltage is, I'm going to use that voltage +/- [your setting]"

IE: Real voltage is 1.38V. Add negative offset of 0.08V. Effective voltage used is 1.38V - 0.08V = 1.3V. This allows power saving functions to remain on because when the CPU switches to lower state, it might drop to 1.1V. Add the negative offset and it it becomes 1.02V.
 
ButterB said:
So I haven't looked at offset voltage to much. Does this work in conjunction to LLC or in place of it? I am looking it up but just trying to understand how it works. Yes this board has it listed as DVID.
Click to expand...

One thing to consider though, if you have high settings of LLC it will actually add voltage instead of just compensating for Vdroop. This can lead to problems where your offset is correct for the high load situation where LLC is boosting voltage, but not enough for lesser loads where LLC hasn't fully kicked in (like single threaded loads). The result is that the voltage isn't high enough, because the offset isn't correct, and leads to crashing under single threaded loads. The way to counter that is to use a lower setting for LLC, and a higher offset.

For example (all numbers approximate):
4.5 Ghz, high-load, 1.4V needed, 1.25 Vid, +.10 offset, +.05 LLC - fine because total Vcore=1.4V
4.5 Ghz, light load, 1.4V needed, 1.25 Vid, +.10 offset, +.00 LLC - crash because Vcore=1.35V
 
You must log in or register to reply here.
Back
Top