• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

***Official Ivybridge Overclock Thread***

eyetrip said:
Can't help myself - 5GHz, 1.336V.

Decided to swap to Aida64 for my stability testing - these guys were told by Asus and Intel it was good ofr Ivy testing, so am going to give it a go.

Only 12 minutes in, temps of hottest core hit 91.

Not used it before though, so not sure yet how long I need to run it for...
Click to expand...

if it passes Aida and fails Prime95 .... does that still mean its stable ?

sure its another test to use - but I'd go with whatever gets processor hottest... does Aida even use AVX ?

I find Prime95 27.7 temps not much hotter than Handbrake - which also uses AVX

in fact I find Handbrake tougher stably test than Prime95
 
Buckster said:
if it passes Aida and fails Prime95 .... does that still mean its stable ?
Click to expand...

Hmmm, I dunno.

Maybe will give Prime a go after, and see how it compares.

I have an issue with things like Prime and IBT the same way I have an issue with Furmark.

I will install rosetta as well, as I found that to be better than Prime and IBT for finding instability.
 
:) as I said I find a nightly version of Handbrake is one of best tests for stability

certainly most likely to introduce WHEA errors anyway :) and my 3770k gets toasty whilst encoding ....!!
 
Passed an one hour and 15 min of Aida, fired up IBT, and after a few seconds of high, system hung :)

Also, figured I was just being greedy going for 5GHz. 4.9 is quite manageable, with temps maxing out in low 90's on hottest core in IBT very high, and takes around 1.264V under heavy load to do so.

Already far better than 99% of people with a 3770K will see, so will leave frying my CPU til further down the line, when CPU getting older and I feel OK removing the IHS :)
 
Well it seems I've hit a wall at anything over 4.8GHz with my chip. To get to 4.8 it needs 1.376v (CPU-z) to remain stable during stress testing. The temps are high 80's on cores 1, 3 and 4 with core 2 at 93. There is usually an 8-10 degree difference between core 2 and the other cores. I'm thinking of removing the IHS and replacing the TIM with some Liquid Pro to try and bring those temps down and hopefully even them out too.
 
Gibbo_GTI said:
Well it seems I've hit a wall at anything over 4.8GHz with my chip. To get to 4.8 it needs 1.376v (CPU-z) to remain stable during stress testing. The temps are high 80's on cores 1, 3 and 4 with core 2 at 93. There is usually an 8-10 degree difference between core 2 and the other cores. I'm thinking of removing the IHS and replacing the TIM with some Liquid Pro to try and bring those temps down and hopefully even them out too.
Click to expand...

Giving serious thought to removing my IHS as well.

Just need to decide if it's really worth it.

I know my chip will do 5GHz - it's just the temps I don't like so much.

With IHS removed, even slight possibility of hitting 5.1....

But if I much it up, not only will have thrown £250 down the toilet, but I may not get another chip as good....
 
I've just done a Prime95 run with Real Temp and CPUID HW Monitor running. There is a 10 degree difference between the two programs with HW Monitor being lower. I've noticed a few posts where people are using HW Monitor to measure temps so which one should I use?
 
eyetrip said:
Giving serious thought to removing my IHS as well.

Just need to decide if it's really worth it.

I know my chip will do 5GHz - it's just the temps I don't like so much.

With IHS removed, even slight possibility of hitting 5.1....

But if I much it up, not only will have thrown £250 down the toilet, but I may not get another chip as good....
Click to expand...

Heres a tip, if you put your CPU in the oven at about 80C for 10 mins the silicon glue should go pretty soft.after that I would suggest a thin softish plastic spatula kind of thing to just lever the IHS off. It's what I plan to do with my 3770k and seems the safest bet of not killing the cpu imo :p
 
Ollie1132 said:
Heres a tip, if you put your CPU in the oven at about 80C for 10 mins the silicon glue should go pretty soft.after that I would suggest a thin softish plastic spatula kind of thing to just lever the IHS off. It's what I plan to do with my 3770k and seems the safest bet of not killing the cpu imo :p
Click to expand...

Or just IBT it for 10 mins, shut down, pull out (burn fingers :D), grab sharp utility knife and pray!




Anyway, how big is the correlation between memory speeds and cpu core voltage when ocing to high levels?

Got rather annoyed this morning when I discovered that cpu frequency/voltage was no longer working at previously stable tested settings.

Then it occurred to me on my way into work, that I have been upping the memory speed/timings

Now, I am aware that faster memory will allow higher throughput to the cpu, but didn't really think it would be the the extent I have epxerienced.

When I initially started clocking this chip, I was getting 4.9 stable, with around 1.25V - but I had the Samsung Green on stock, which is a very relaxed 1600 11-11-11-28-2t

Then I needed around 1.26/1.27V. But I think this might be when I started upping the mem speeds to around 1866 with 9-9-9-28-2t.

When I figured out how to get memory past 1866 on the sabertooth, I set it at 2133 11-11-11-2t, but needed to up vcore to around 1.28V.

Then this morning I started failing IBT at 1.28V, and later remembered that I had changed timings to 10-10-10-2t.

I know I have done the whole thign a bit backwards - I was only just telling someone the other day to OC memory where you want it first, as it faster memory could require higher vcore when aiming for high clocks at lowest possible voltage.

So tonight I am going to put the memory back to stock, and see if I can get back to the 1.25V for 4.9GHz. If I can, it means at least that my CPU has not degraded as I was beginning to suspect.

If this is the case, then I just need to decide if I want to sacrifice a bit of cpu speed for memory speed, or vice versa.
 
Last edited:
For me removing the IHS and applying liquid metal reduced my temps from 90c to 75c on average. That is with a 4.4Ghz OC on a 3570k.
I found it hard to believe some claims in how successful it sound be in reducing the temps but I did find that the above temp differences did apply in my case. Each chip will be different of course, as will the volts and OC used. Prior to doing that raising the OC to 4.5 meant increasing my volts and thus the temps went even higher that 90c.
Caveats do apply but I'm kinda happy with what I achieved.
 
Just done the IHS removal on my 3770K

Method used was to bake in Oven at 90c for 15 minutes
Then using a thin modeling knife worked my way around lifting the IHS.
Then cleaned off all the black glue on both the CPU and IHS.
Applied liquid metal to the CPU... stuck it in the motherboard and added the IHS.
Heatsink applied using IC24 and job done.

I'm using a custom water loop, just doing the CPU with 2 140mm fans on a 240 rad.

At 4.8GHz with CPUz reporting 1.312v
Running Prime 95 on all cores with HT enabled.
Was a peak temp of low 80's C before
Now one core has just hit 68c after 1 hour... The best core has only hit 59c

So I think i'm seeing something like a 10c - 15c drop in temps.

Suddenly i'm really happy with the 3770K :)
 
SimonMaltby said:
Just done the IHS removal on my 3770K

Method used was to bake in Oven at 90c for 15 minutes
Then using a thin modeling knife worked my way around lifting the IHS.
Then cleaned off all the black glue on both the CPU and IHS.
Applied liquid metal to the CPU... stuck it in the motherboard and added the IHS.
Heatsink applied using IC24 and job done.

I'm using a custom water loop, just doing the CPU with 2 140mm fans on a 240 rad.

At 4.8GHz with CPUz reporting 1.312v
Running Prime 95 on all cores with HT enabled.
Was a peak temp of low 80's C before
Now one core has just hit 68c after 1 hour... The best core has only hit 59c

So I think i'm seeing something like a 10c - 15c drop in temps.

Suddenly i'm really happy with the 3770K :)
Click to expand...

Hmmm interesting, so you have to reapply the IHS then so the heatsink would contact the CPU ?

Any way of just directly attaching the heatsink/water block ?
 
Combat squirrel said:
Hmmm interesting, so you have to reapply the IHS then so the heatsink would contact the CPU ?

Any way of just directly attaching the heatsink/water block ?
Click to expand...

i'm sure you could. It may not tighten enough without modification, but I decided to keep using the IHS as I did not want to risk crushing the core by tightening the heatsink unevenly.

I still have nightmares following chipping the edge of an AMD Xp1800 core when installing a freon cooler.
 
You must log in or register to reply here.
Back
Top Bottom