i5 2500k Voltage?

[mohammod]

Hey guys,
Recently I followed TTLC's video on youtube. And trying to do the first step (lowering voltage to minimum) I had issues.
He states to go to about 1.1 and then go from there lower until it wont boot.
I tried 1.1, then it got to the windows logo thing before boot then froze.
Tried 1.2, 1.3 all the same thing.
What could I be doing wrong?
Could anyone possibly suggest a good video with an Asus board?
Thanks again.



Intel i5 2500k
Hyper 212
8Gb Patriot G2
Asus 7850HD Top V2
Asus p8z68-v pro Rev 3
Force GT SSD
Mushkin 600W
NZXT Phantom 410 Full Tower

 

[seumasbeathan]

What clock speed are you trying for?

 

[mohammod]

Im running on optimized defaults to try to under volt my CPU. So stock clock speed.
Basically to my knowledge I should try to find and start at my lowest stable voltage, then work my way up while increasing my clock speed.

But I would be shooting for around 4.5, depending on what my CPU can do. If it can do more then I'd consider water cooling if necessary.

 

[Master&Puppet]

Generally you'll be looking at using up to 1.25v - 1.4v dependent on how good your cooling is.

 

[airdeano]

ok, (not mentioned) you did set your LLC (load line calibration) to extreme?
set your RAM timings and voltage to rated speeds on the RAM?
multiplier should be 35 (turbo off) at 1.080-1.1v on CPU voltage.
youve missed some steps.. check out the written guide here
youll prolly need to start at page 3.

 

[mohammod]

Master, that's not for stock clock speed is that?
And yes I have set Load Line Calibration to extreme. And is setting RAM timings and such important to under volting my CPU?
Thanks for the replys

 

[Master&Puppet]

Na that is the overclocked voltage for when you are finding x45. You can watch TTL's video for much of it. Sandy and Ivy are very similar.

 

[mohammod]

The thing is, I couldnt get those voltages on stock clock...

 

[airdeano]

Toms vid/guide worked fine for my asus MIVE-Z and MIVG-Z mainboards
i reached 35 on 1.075v loaded OCCT for 30min and held 48° (23.5°A ambient)
H100 swiftech helix @ 7v (1100rpm) stable. 1.065 fail after 6min. 2500k

 

[mohammod]

Hmm thats odd. Maybe im doing something wrong... even though there wasn't much to change haha.
I also just ran OCCT on optimized defaults and I noticed that my clock speed isnt the "default" Turbod speed?

Its running at x42 multiplier, 4326.33 - 4327.71 (from what ive seen so far) MHz Core speed, 103.03 MHz Bus Speed, 1.328 (ranges from 1.288-1.344 from what ive seen so far) Core Voltage.

Also my max temp was 60°

update/
I have since been playing with my voltages trying to get an undervolt to begin with. With Turbo boost turned off I can run at 1.05 comfortably (Never tried anything lower to be honest), with a 20min test on OCCT. However with Turbo Enabled with AUTO Turbo I had it running on 1.075 for about 10 minutes before I BIOS'ed. My Clock speed was usually around 3411 with occasional leaps to 3612. I Then tried making Manual Turbo mode and BSOD'ed or Froze while booting. I then increased Core Voltage to 1.1 and made it into Windows only to crash quickly. Getting Dump files with error codes of x101, 1e, all which point to more vcore. Is my CPU part of the unlucky Silicone lottery? Or is there still hope for me to reach 4.5GHz comfortably on my 212 cooler?

 

[airdeano]

Hmm thats odd. Maybe im doing something wrong... even though there wasn't much to change haha.
I also just ran OCCT on optimized defaults and I noticed that my clock speed isnt the "default" Turbod speed?

Its running at x42 multiplier, 4326.33 - 4327.71 (from what ive seen so far) MHz Core speed, 103.03 MHz Bus Speed, 1.328 (ranges from 1.288-1.344 from what ive seen so far) Core Voltage.

Also my max temp was 60°

update/
I have since been playing with my voltages trying to get an undervolt to begin with. With Turbo boost turned off I can run at 1.05 comfortably (Never tried anything lower to be honest), with a 20min test on OCCT. However with Turbo Enabled with AUTO Turbo I had it running on 1.075 for about 10 minutes before I BIOS'ed. My Clock speed was usually around 3411 with occasional leaps to 3612. I Then tried making Manual Turbo mode and BSOD'ed or Froze while booting. I then increased Core Voltage to 1.1 and made it into Windows only to crash quickly. Getting Dump files with error codes of x101, 1e, all which point to more vcore. Is my CPU part of the unlucky Silicone lottery? Or is there still hope for me to reach 4.5GHz comfortably on my 212 cooler?

bus speed 103.3? huh.. sounds like you have the AI Overclock Tuner engaged.
thats not optimized defaults (F5 in BIOS) and you really cant change the lockouts
unless you save the profile and tweak and save (not worth it)

i like the cm 212+/EVO, but 4.5 is gunna be alotta warm. i used one at 4.6 on an
open bench with 29° ambient and loaded at 1.270v wavering at 78° (stock blademaster fan)
and in a case with 25° ambient and loaded 1.270v eating 82°.
there is hope.. but for an easier feeling, id look for a used D14 or H100/i (if case
compatible).

 

[mohammod]

I think that was the case airdeano, Since I updated the thread I believe Ive fixed that problem. haha.
But I cant even achieve 3.3 with 3.7 Turbo boost on 1.1 volts. Is that normal? Should I keep increasing Core Voltage until I have a stable system with that setup? Even if that means almost reaching stock voltage? haha. On TTLC's video he showed that he had stock with turbo boost running on 1.05 =(
Im just curious as to what it can do because for the time being it seems to be running pretty cool, at least in my opinion. Running at 3.3GHz WITHOUT Turbo on 1.05 volts boost i was running on average at about 40°. This voltage fluctuated on CPU-Z showing 1.056 - 1.064 while Core speed fluctuated from 3310.74-3311.18. The Bus being 100.34.
I feel as there is no point in me purchasing a H100/i if my CPU cant even handle the Overclock anyways at a stable Voltage. If this is a 5.0GHz + CPU then Id be more inclined to purchasing a H100/i. My case has more than enough space; NZXT Phantom 410 Full Tower.

 

[airdeano]

I think that was the case airdeano, Since I updated the thread I believe Ive fixed that problem. haha.
But I cant even achieve 3.3 with 3.7 Turbo boost on 1.1 volts. Is that normal? Should I keep increasing Core Voltage until I have a stable system with that setup? Even if that means almost reaching stock voltage? haha. On TTLC's video he showed that he had stock with turbo boost running on 1.05 =(
Im just curious as to what it can do because for the time being it seems to be running pretty cool, at least in my opinion. Running at 3.3GHz WITHOUT Turbo on 1.05 volts boost i was running on average at about 40°. This voltage fluctuated on CPU-Z showing 1.056 - 1.064 while Core speed fluctuated from 3310.74-3311.18. The Bus being 100.34.
I feel as there is no point in me purchasing a H100/i if my CPU cant even handle the Overclock anyways at a stable Voltage. If this is a 5.0GHz + CPU then Id be more inclined to purchasing a H100/i. My case has more than enough space; NZXT Phantom 410 Full Tower.

1.10v is highly comfortable. yeah, some can go lower with some tweaks, but
its not a contest, but consistancy is what your are looking for.
the .008v VDROOP (LLC) is a normal fluctuation.
now you are ready to take the next step and tune for 4.0GHz with a slight
bump in voltage (to maybe 1.125-1.135v) and retest.. each time you are
sucessful, record that combo and move up to the next, till temps are not
safe to hold (80°).

 

[mohammod]

I actually went ahead and did that today airdeano.
I had a few hiccups but I think it went okay in the end. 1.165v at 4.0 it worked for 15 mins. Then I turned off OCCT and it Crashed about a minute AFTER OCCT was off.. So I turned it up to 1.175v same thing happened. 15 minutes worked fine max temps reaching 48°. Went to 1.185v and it BSOD'ed about 30minutes without even touching OCCT (Was talking to a friend on skype and played a game of LoL)
I then went to 1.2v and its been a few hours playing LoL with no crash's. Have yet to OCCT test it for time though.
Thanks again airdeano! =)

 

[airdeano]

have fun with overclocking and enjoy..

i keep a log on what worked and what didnt (and why) so if someone has a solution
you can go back and recreate the overclock and make corrections/adjustments.

 

[mohammod]

Haha fair enough.
Im curious if you have a general guideline for voltage for certain clocks? Like you said 1.125 - 1.135 for 4.0. What would be a guideline for 4.5GHz? Because if I went to 4.5 there should I test 1.25,1.275,1.3,1.325,1.35? Since they would just be way too low anyways hah.
Thanks again. Appreciate the help.

 

[airdeano]

ahhh grasshopper...
i could tell you, but then id have to kill you... LOLOL j/k
the exercise in overclocking and tom releasing his information was that "not all
printed or verbal is the gospel.." and short cutting isnt fair to you.

someone telling you the numbers pretty much does nothing for you. and if you
had to recreate them, you couldnt because the procedure wasnt learned.

and that statement "no point in me testing...." tom would flame you (id edit it asap)
getting to 4.5 would be as easy as getting to 4.4 from 4.3 and so on...
skipping the steps in between doesnt allow you the knowledge on how you got there.
and id loose my moderator badge (i like my mod badge) if i told you, too...
if you were to go to 4.2 from your 4.0 like toms steps.. it'll be clear on to get to
4.3/4.4/4.5 plus your boards power phase can handle alot.

 

[mohammod]

LOL. Yeah Yeah. You don't even know meeee. You don't know what I'm capable of!
But I know that. I think I'm just being too eager to achieve 4.5 LOL.
I tried doing that once but it resulted in being unstable for me (Who would have guessed) haha. But I agree. I am learning by going through the steps thus far! Increase vcore, increase vcore, increase vcore. Wow it works. Crash. Increase vcore. So much fun in learning the steps! =P
How long should I test my current setup (1.21v at 4.0GHz) to test for a 24/7 stability? Also should I aim for not passing 60° on OCCT if I want it to be 24/7?
But done! (Now edit that so there is no evidence!)
But I guess so. I Like it too. You should give me something nice that will make me just as cool as you. =)
But hmm. Fair enough. I currently have Power Phase on Extreme and Internal PLL Overvoltage disabled. Should I enable the PPL Overvoltage?
Thanks for all your help, I really do appreciate it. =)

 

[airdeano]

go 30minutes.. then go down .005v in voltage.. retest.. till it crashes.
within 30min the temps stabilze and enough torture to the processor to get the
big bugs out.
PLL only if you use off-set method i believe.. mines sill enabled. some disable and
temps come down slightly.. thatd be something to look into with master&puppet's
guide to look at..

 

[mohammod]

Sounds good. Im in the process of doing that at the moment now.
But hmm. kk. Ill look into it.
Quick question. I was running 4.1 at 1.210v and it passed OCCT Stable for 30minutes. With max temperatures of 45,48,48,47. About 20 minutes AFTER the test (was browsing youtube and such) I crashed with an error of xD1. I found an "OVerclockers BSOD code list" And it said = add QPI/VTT voltage? Any tips on this?
Thanks!

~Edit.
What temperatures should I be looking to stay under if I want to maintain this as a 24/7 overclock? Somewhere around the ~60 range?
Currently at 4.3GHz at 1.260 v, im sitting at max 53,56,56,54 after a 30min OCCT.
Thanks again!

~Edit x2
What could be the reason for crashes occuring well after the OCCT test is done? For example with my 4.3 I did the 30min test and all was fine. About 1.5 hours after I finished (And played some LoL With some friends) I Froze with x124 being the problem. Fix to this is to add more vcore. Why would it fail after ~2 hours? Is this normal?
Thanks again..... haha