Has anyone delidded Haswell yet?

So...people de-lidding, but no one pushing their chips to see if they can get a better overclock. WTF is wrong with this picture? lol
 
HydroBudz said:
So...people de-lidding, but no one pushing their chips to see if they can get a better overclock. WTF is wrong with this picture? lol
Click to expand...

Judging from what Kyle was saying in his review, you have to get a very very good part just to push it to 4.8ghz.

But yeah if you are de-liding then I expect nothing short of dry ice cooling and massive overclocks, are those days gone with Haswell :confused::(
 
From what I've seen on other forums people are getting a 10-20c drop from de-lidding. Temps were mostly the factor when pushing these chips. So, 4.8 shouldn't be an issue with reduced temps.
 
I've had a peek at a few ivy delidding threads... why are people putting the IHS back on instead of adding a shim and doing it old school style?
 
I've had a peek at a few ivy delidding threads... why are people putting the IHS back on instead of adding a shim and doing it old school style? Is the IHS really needed to increase the surface area?
 
Dephcon said:
I've had a peek at a few ivy delidding threads... why are people putting the IHS back on instead of adding a shim and doing it old school style? Is the IHS really needed to increase the surface area?
Click to expand...

I think theyre putting the IHS back on to spread the pressure and prevent their cores from being crushed moreso than as a way to increase the surface area.
 
I'm pretty hesitant to ever try this, especially with Haswell's die having the integrated VR - more chance for damage even if you're careful, and i'm generally not too excited about throwing away a 350$ chip. Still, there's nothing to indicate that Haswell is any different than IVB in this respect, delidding can shave 10c from peak temperatures.
 
If you guys find any overclocking attempts on delidded chips drop a link please. I want to know if delidding increases maximum overclocks. My prediction is that it won't most of the time, but I wanna see the data.

Thanks for the links.
 
There's a thread at OCN with users that have delidded, and you can expect load temps to be 10-12C lower if you're lucky. Keep in mind that de-lidding Haswell is slightly more risky than it was with IVB due to the integrated VR - certainly not something i'd want to do, but that's just me.

Seeing as overclocks with Haswell are temperature limited, I see no reason why it wouldn't help overclocks - but that isn't the main issue. The main issue with Haswell is that there is so much variance in terms of what voltage a particular clockspeed requires, more so than with prior CPUs. One CPU may require 1.125V for 4.6GHz while another requires 1.35V - the quality of your chip is everything with Haswell. But, back to your question - there's a thread over @ OCN with exactly the information you're seeking.
 
Going to be delidding a 4770k in the near future and over clocking it with a Gigabyte Z87X-OC.

I had alot of fun delidding 3570k and 3770k. Haswell is a greater challenge and I hope it works out.
 
XacTactX said:
If you guys find any overclocking attempts on delidded chips drop a link please. I want to know if delidding increases maximum overclocks. My prediction is that it won't most of the time, but I wanna see the data.

Thanks for the links.
Click to expand...

That would indeed be interesting. A reduction of 10C - 20C is great, but the question is whether it allows a significantly higher overclock (not just 100 - 200 MHz) with an average CPU.
 
i delidded my 4770K using the vice + hammer method from overclock.net. I was really surprised how easy it was. A couple light hits (nothing more than the weight of the hammer from about 6 inches) and the adhesive broke loose enough to lift the chip off the ihs. Personally i don't trust myself enough to do the razer method, no way im steady or patient enough for that.

I put ceramique on it temporarily and saw temps drop about 15 deg. CLU is on the way and should be here today or tomorrow.

Before delid:
4.2ghz @ 1.25v - around 92Deg in Prime

After delid:
4.6ghz @1.32v - around 85 Deg in Prime

Im still playing with the overclock and will see how much farther i can go with the CLU. My chip seems pretty finicky though, i can get up to 4.5 with 1.26v but anything over 4.5 seems to require a massive jump in volts so im not sure at the moment what i will keep it at.
 
obliv said:
i delidded my 4770K using the vice + hammer method from overclock.net. I was really surprised how easy it was. A couple light hits (nothing more than the weight of the hammer from about 6 inches) and the adhesive broke loose enough to lift the chip off the ihs. Personally i don't trust myself enough to do the razer method, no way im steady or patient enough for that.

I put ceramique on it temporarily and saw temps drop about 15 deg. CLU is on the way and should be here today or tomorrow.

Before delid:
4.2ghz @ 1.25v - around 92Deg in Prime

After delid:
4.6ghz @1.32v - around 85 Deg in Prime

Im still playing with the overclock and will see how much farther i can go with the CLU. My chip seems pretty finicky though, i can get up to 4.5 with 1.26v but anything over 4.5 seems to require a massive jump in volts so im not sure at the moment what i will keep it at.
Click to expand...

Thanks for sharing your results with us..I have been hesitant to switch from my 3770K since I have a great chip that does 4.8Ghz @ 1.272V loaded 24/7 with Prime95 and everyday since with [email protected] am going to push for 5Ghz+ as soon as NZXT starts shipping the H630 cases so I can use my 2nd 360mm radiator..
 
Ran Prime @ 4.7 Ghz / 1.4 Vcore. Temps were 96-100. And not delided.

I should do it though. Probably could hit 5 Ghz. And I have Liquid Pro on hand too.
 
obliv said:
Im still playing with the overclock and will see how much farther i can go with the CLU. My chip seems pretty finicky though, i can get up to 4.5 with 1.26v but anything over 4.5 seems to require a massive jump in volts so im not sure at the moment what i will keep it at.
Click to expand...

Very common with Haswell. My CPU has one of those "voltage walls" at 4.3 GHz. At 4.3 it only needs 1.2V, at 4.5 or even 4.4 it requires a lot more - I only went as high as 1.25V and it was not quite stable. Not worth it to go higher for a ~5% increase IMO. I mean I could probably achieve 1.5V at close to 1.3V or more, but temps would be close to 100C in Aida64 and power draw would increase significantly.

This is why I'm wondering whether de-lidding will help much if your CPU has this frequency/voltage wall.
 
You must log in or register to reply here.
Back
Top