help overclocking i5-2500k

Plec · 19 Jul 2011 at 16:25

snadge said:
before
core temps (MAX): 66 - 74 - 72 - 69

after
core temps (MAX): 62 - 69 - 66 - 62

so its about 5'c cooler so far

5 degrees is a big improvement in the area of temps you were dealing with.

snadge said:
will have to use the cheap stuff that comes with the Gellid if I re-sit it again and I will spread it out myself

maybe im putting too much on..?

You 'may' be able to improve your temps with another reseating - and there's nothing wrong with the Gelid TIM it will do the job. So, if you feel you want to try again don't worry about using the pre-supplied gunk.

For future refernence you only need a small pea sized blob in the centre of the CPU and let the pressure and heat of the HS spread it out.

snadge · 19 Jul 2011 at 16:37

This is the problem.. a small pea can be any size hehe..
how do you know if its 'too much' when you take it off?

should I try 4.4Ghz on Prime95? the user above says he gets 65'c @ 1.352v - would I have to change the voltage? - i had it on 4.4Ghz last night for a bit and didnt crash (without changin voltages) how do you know when you have to change voltages?

Plec · 19 Jul 2011 at 17:42

snadge said:
This is the problem.. a small pea can be any size hehe..
how do you know if its 'too much' when you take it off?

When it's leaking over the sides of the CPU or you have lots of build up on the sides.

snadge said:
should I try 4.4Ghz on Prime95? the user above says he gets 65'c @ 1.352v - would I have to change the voltage? - i had it on 4.4Ghz last night for a bit and didnt crash (without changin voltages) how do you know when you have to change voltages?

YOu can if you want to - you can leave the voltage on auto and run cpuz while priming - and monitor the voltages you get.

Disable 'spread spectrum' and enter the memory settings manually - and make sure the memory's voltage is under 1.58V (most sets will run at 1.5V which is the ideal)

Also, take a look at this thread - link as i've discussed/advised an overclock with a similar build and clock so you'll have a lot of your questions answered in there.

ahar · 19 Jul 2011 at 17:50

snadge said:
remember i dont have any case fans... do you? do they make much diff? we are ordering two Yate Loons on..?

I would say that a rear exhaust would make a difference as the heat sink fan is blowing hot air around the case, some of which wont escape before being recirculated.

I use Yate loons for my case fans and they seem to do a good job for me.

snadge · 19 Jul 2011 at 19:16

are they silent the Late Yoons?

I have installed the small loud case fan that come with it on the back so that HSF was blowing air to it and the case fan sucked it out through back and its reduced temps by 3'c - so a big one in front and back should make a difference

when I removed the HSF earlier there was a TINY bit of paste over one of the sides but barely

Everyone else seems to have higher voltages even on Auto... and cooler temps

what has the memory got to do with it? (spread spectrum)
in the other thread he says he disabled C1E...whats that?
the memory (when I last checked) was 1.5v

I will reboot now into 4.4Ghz with everything on Auto and see what happens

snadge · 19 Jul 2011 at 19:37

...ok at 4.4Ghz
CPU-Z says Core Voltage is still 1.056
Temps on Prime95 after 5 minutes smallFFT 60'c (in Tray from Real Temp_

MAX TEMPS
62 - 69 - 67 - 63

AVG TEMPS
61 - 68 - 65 - 61

this is with a small 80mm case fan on back going 3200rpm

Plec · 19 Jul 2011 at 20:00

snadge said:
...ok at 4.4Ghz
CPU-Z says Core Voltage is still 1.056
Temps on Prime95 after 5 minutes smallFFT 60'c (in Tray from Real Temp_

MAX TEMPS
62 - 69 - 67 - 63

AVG TEMPS
61 - 68 - 65 - 61

this is with a small 80mm case fan on back going 3200rpm

Your voltatges at load are fantastic as 4.4GHz(?) - you may have a gem like my own.

Your temps are fine (as said before) it's only in comparison that they seem high - so don't worry about it (not all components run the same.)

snadge · 19 Jul 2011 at 20:06

thanks. cos its really got me why other peoples are at least 10'c cooler than mine...even in idle ive seen the gellid benchmark 19'c when mine is 34'c ...?? something doesnt seem right

are those voltages correct/safe? it stays same 1.056 voltage for 4Ghz or 4.4Ghz and 1.044 on Idle (1.6Ghz) everyone else has 1.1 to 1.3 or so ...will running at 4.4Ghz on 1.056v cause any problems?

can I ask, why disable 'spread spectrum'..? I couldnt find it in BIOS

thanks for all the help

edit: just realised we have practically same setup hehe

Plec · 19 Jul 2011 at 21:22

snadge said:
thanks. cos its really got me why other peoples are at least 10'c cooler than mine...even in idle ive seen the gellid benchmark 19'c when mine is 34'c ...?? something doesnt seem right

Even identical setups will have their own, separate, quirks- no components are the same so when you add multiple components the results can be dramatically different (relatively) - and yet everything on the surface may be identical.

snadge said:
are those voltages correct/safe? it stays same 1.056 voltage for 4Ghz or 4.4Ghz and 1.044 on Idle (1.6Ghz) everyone else has 1.1 to 1.3 or so ...will running at 4.4Ghz on 1.056v cause any problems?

As mentioned before - your volatges are very impressive. Lots of forum members have to push 1.25V and upwards to get 4.4GHz stable. If what you are reporting is accurate then you have a gem of a chip for silent running (with modrate clock). It may be that you have a chip that could hit 5GHz - but the two don't neccessarily go hand in hand.

snadge said:
can I ask, why disable 'spread spectrum'..? I couldnt find it in BIOS

If you can find it, disable it (what MB do you have) as it's a function that is made defunct when clocking (well we choose to ignore it, thruth be told).

snadge said:
edit: just realised we have practically same setup hehe

What is your setup?

snadge · 19 Jul 2011 at 23:36

its my dads setup for his FSX..i have it here until Sunday when we go up to his.

2500k @ 4Ghz (until we get case fans then 4.4)
Gigabyte P67A-D3-B3
Corsair 4Gb XMS3 1.5V
MSI GTX 460 (336 cores / Fermi) @ 725Mhz
XFX Pro 550watt with EasyRail 80+
60Gb Vertex 2E SSD
500Gb Seagate HDD (SATA3)
Win7 x64

im probably gunna try re-sitting the HSF again at some point, this time I will use the Gellid paste and see what happens - if its still the same then thats that!

how is it some people have to change voltages and some dont? is it just down to the Motherboard and PSU combo?

ive looked all over BIOS for 'spread spectrum' ...cant find it

Plec · 19 Jul 2011 at 23:54

snadge said:
how is it some people have to change voltages and some dont? is it just down to the Motherboard and PSU combo?

When you try for larger clocks you usually have to fix the voltage manually for stability.

You're auto voltage reading at 4.4GHz is very impressive - and if you can confirm it as stable then you may as well leave it at auto so that you get the idle voltage benefits ~0.98V (if 4.4GHz is your target clock?).

snadge said:
im probably gunna try re-sitting the HSF again at some point, this time I will use the Gellid paste and see what happens - if its still the same then thats that!

As mentioned before your temps are fine - but it wont hurt reseating again.

ahar · 20 Jul 2011 at 07:44

Annoyingly one of my Yate loons isn't that quiet (but the rest are), but the rest of them are - I think there's a problem with that one and at £3 each I'll just replace it.

snadge · 20 Jul 2011 at 07:46

ahar said:
Annoyingly one of my Yate loons isn't that quiet (but the rest are), but the rest of them are - I think there's a problem with that one and at £3 each I'll just replace it.

how many do you have?

can you hear them?

ahar · 20 Jul 2011 at 07:55

I've got 4 of them and one coolermaster fan that came with the case. It's the intake fan on the side that makes a humming sound. The rest seem fine.

snadge · 20 Jul 2011 at 08:01

just im struggling to find some quiet ones that my dad can afford, his PC is silent, its on floor right next to me and i cant hear a thing... we'd like to keep it that way.. and he doesnt live locally so its not like we can get something else if not happy as im installing them then i will be leaving his...

I managed to reduce temps on the Gellid by 4-5'c by re-seating it, im gunna try again as im still hitting 69'c (on one core - averages 58'c in the tray icon - realtemp) at 4.4Ghz (1.056v) on Prime95 ... and that was with a noisey 2800rpm 80mm case fan on the back.

could it be all your fans how you get 55max on 4Ghz? and which core is that? RealTemp displays the lowest core temp in tray icon..is this your reading?

ahar · 20 Jul 2011 at 10:42

I would say that the temps you're getting now at 4.4Ghz are pretty comparable with what I get given you have one 80mm fan. Although my temps are a little lower at 4.5ghz and higher volts, I have more fans running and I'm using IC Diamond paste which seems to be pretty good.

snadge · 20 Jul 2011 at 12:18

ahar said:
I would say that the temps you're getting now at 4.4Ghz are pretty comparable with what I get given you have one 80mm fan. Although my temps are a little lower at 4.5ghz and higher volts, I have more fans running and I'm using IC Diamond paste which seems to be pretty good.

thats what im hoping for (case fans helping)

when you say max of 55'c is that the highest temp on the hottest core? or just what RT reports in tray? (lowest temp)

ahar · 20 Jul 2011 at 14:39

I use core temp which reports the highest core temp in the system tray, and it was 65c when running at 4.5Ghz. Not sure what the lowest was, but I tend to have a variance of up to 7c across all four when at full pelt

snadge · 20 Jul 2011 at 15:26

what do i change in BIOS to adjust core voltage?

DVID
QPI/FTT
SYSTEM AGENT
PCH CORE
CPU PLL

i want to try 1.3v at 4.4Ghz as someone has told me that my current voltages are impossible and mis-reported, HWmonitor reports same voltage as CPU-Z

joeyjojo · 20 Jul 2011 at 15:30

Plec said:
As mentioned before - your volatges are very impressive. Lots of forum members have to push 1.25V and upwards to get 4.4GHz stable. If what you are reporting is accurate then you have a gem of a chip for silent running (with modrate clock). It may be that you have a chip that could hit 5GHz - but the two don't neccessarily go hand in hand.

It's probably not stable.