i5-2500k Overclocking question

Associate
Joined
16 Jul 2012
Posts
280
(new to desktop cpu overclocking)

So I've got an i5-2500k in an MSI Z77-G43 with a H75 Corsair AIO water cooler.

Changed the cpu multiplier to 40x in the bios with the everything else the same as I want it to clock down when not doing anything special.

Running prime95 gives solid results after 15 minutes of running 100% load.

Question is why task manager (set to performance logical cores) shows a cpu speed of 4.8ish Ghz when in load and 4.7ish when no load. Surely it should say 4GHz under load and 2/3 when not?

CPU-Z reports 4Ghz limit and Real temps show 50C when under load so seems alright?

See this pic for example:

Capture_zps6b367f83.png


edit: just to be clear, my aim is to run about 4.2GHz so when I'm gaming/intensive work it runs at this speed but when I'm in Windows (browsing, word documents etc), it runs at the normal stock speeds etc). Any tips to achieve this would be great :D
 
Ignore what task manager says for cpu speed its wrong.

Says mines at 7.77GHz lol

Prime needs longer than 15mins to be sure its stable...2hrs minimum in my books ..4 better or 8 perfect :)

OCCT was good for triping my system up after only 20 mins when prime was fine for 4 hrs +

Temps look fine but voltages look high for 4.0GHz....have you left vcore on auto?
Looks like its at 1.3v now but going upto 1.37v under load.
Use the offset for the vcore and loadline calibration to control the vdroop under load.

Nothing wrong with 1.37v but that voltage should get you up to more like 4.5GHz even on a poor clocking chip like mine.

If you leave the power saving features on it will clock down at low load.

Mine drops to 1.6GHz but voltage stays at 1.37v
Then under load it ups the vcore to 1.38v via LLC (non uefi bios i have)

100% stable and temps max out at 70c and mid 50s in games.

Hope that helps :)
 
Well if its maxing out at 50c with 1.37v then why not try for 4.5GHz? :)

Dont leave it on auto as board will overvolt it for sure...they all do

General rule is under 1.4v and 80c for 24/7 clocks/usage but if temps are so low then beyond 1.4v should be okish...just depends on what you are willing to risk :)

Thats coming from someone who killed a 2600K cpu....could bench at 5.55GHz and silly volts but looked like it finished the chip off before its time.

Intel RMA saved my day though :)
 
Yes :)

If stable then drop it if not then raise it.

As i said occt is good at tripping up and unstable overclock pretty quick but prime95 is the "proper" test.
 
You will soon find the "wall"...the point where cpu needs big increase in vcore for another 100MHz...i have a 3570K that will do 4.3GHz on stock but then needs a lot extra to get to 4.6GHz.

All depends on what you want and how hard you are willing to push it.

Most would say 4.5GHz is more than enough though.

Have fun :)
 
No problem at all mate :)

If it wasn't for others helping me out over the years i still wouldn't have a clue what i was doing...and still don't compared to the knowledge of others on here :)
 
No problem at all mate :)

If it wasn't for others helping me out over the years i still wouldn't have a clue what i was doing...and still don't compared to the knowledge of others on here :)

I decided to screw it and do it now as its fun haha. This is after an hour at stock voltage and 42x

Capture2_zps54c893ab.png


Seems good temps again? :)

max of 55C
 
Cant complain with that :)

I notice the 12v rail of psu reads very low...dont worry about that will be a false reading by software...if it was truly that low pc wouldnt boot/run at all :)

Noticed also TMPIN1 has max reading of 87c? dont know what that sensor is for though...in previous screeny it showed a min of -89 so prob a false reading.

speedfan can be a good program also for various temps plus you can see from graphs which ones vary properly as load applied.

http://www.almico.com/speedfan.php
 
Back
Top Bottom