please someone take a few pic of the operation of the X2 before and after for us new ppl. :)
Printable View
please someone take a few pic of the operation of the X2 before and after for us new ppl. :)
All I can say is WOW.
Great job guys. :up:
I have done it so it is to late.
here is a (how to remove ihs from a amd64)
It in danish but the pic is good.
how to .
here is the tool i use.
Do you guys have any recommendations on whether a Hyper6 would be safe to use with the IHS removed? Before you flip out, I know it is heavy, but I am using it vertically (in a desktop style case), not in a tower, so there should not be any leverage issues on it, just a decent amount of weight evenly dispersed from the top.
To MMIQHW, if you see this, where did you get your bumpers from for the CPU? I would appreciate a reply or a PM, thanks man!
only if it was in english.
p.s. do you need to wear gloves or anything while doing this?
I agree with the helping in temps (it makes quite a huge difference with watercooling) and OC by removing the IHS, sure did work great for me, running 10x285 @ 1.5v stable so far, need better memory to go higher with the fsb and low multi ;)
@rsGTS
One set from an old Alpha Cooler. The other ones used ontop of my venice and my winchester are from creative labs speakers. You could buy things like this for example from tesa, the producer of the legendary tesafilm.
Greetz,
Erik
The swiftech one that screws down. You don't want to use a big heavy clip down type of sink on the exposed core. Too much pressure on the corners when mounting and remounting.Quote:
Originally Posted by rsGTS
In posts 3-6 they mention one core going higher than the other. How is this possible? Both cores on my x2 4800 have the exact same clock.
Are you using some special version of clockgen made for the x2?
I am using bios 702-2.
They just set both cores to 2,7ghz for example but test them one on one
hmm, I wasn't aware I could test the cores seperately
Yeah we only talked about a 1000 times about it. :p:
If you end up having perfectly matched core... something is wrong... you should go buy a lotto man! In general though it seems about 150-250mhz difference between cores (which could certainly have been better matched with associated cost by AMD I bet)
..and mostly the difference is caused by a bad connecting IHS
I am just going by what CPU-Z says 2905.5 for both cores. What thread talks about testing cores seperately...I have been out of the loop.
You set the affinity to specify one core or the other.
No, the issue was one core would overclock higher without crashing. They are both set to the same o/c all the time, but testing could isolate each core at the same o/c via the afinity thing and see which one was stable.
I will have to research setting affinity. I haven't seen this.
You can assign affinity of any process from task manager and prime has an option already integrated that allows you to set affinity.
I see now. Thanks.
OK, so screw down ones are better? Like the Zalman 7000cu?Quote:
Originally Posted by xgman
I got a 7000cu mounted on my XP 2500.. screws are down as far as they go (not sure if i was supposed to do that lol) but no probs with that so far.. so i think you're right...
But what i'm wondering though is.. how about the 9500?.. I'm planning to get that with a 4400+ when both are available here, but seeing the results in this thread i really want to remove the IHS as well.
Will that be dangerous considering the 9500 is taller?..
In a tower case the screws will be horizontal (on the DFI board)
i'd think it would be safer with a heavy sink if the screws were vertical.Quote:
(x) socket (x)
I guess i better find some damn good bumpers/washers... or is it really not that fragile?Quote:
(x)
socket
(x)
Not a challenge but I haven't seen seen anything that clearly proves this..... :)Quote:
Originally Posted by sai
A solid stable 200mhz gain on core0 and a 15C overall load temp drop in my case is all the proof I need. :)
Glad you got the bump, but your success hardly supports a challenge that AMD’s manufacturing process is faulty or that a poorly mounted IHS is the norm rather the exception. If this was the case we would have heard much more ‘bout this long before the X2 was born……. :)
Well, it looks like one side of the die, or one core, or whatever is getting good contact with the ihs, and one isn't, therefore one clocks higher than the other regularly. When taking it off, it seems both will do higher rate because of better contact. I don't see how this could be seen before X2 was out, as there would need to be two cores to show this happening. As for the ihs making bad contact on single core cpus, people have done this ever since 754 iirc, and have been getting better results in some cases. It looks like the IHS just sucks in general, but can definately effect the x2 as cooling every part of the die is important, and the ihs doesn't make good contact with all parts on the x2, especially one of the cores.
As for proof...I think 4 or 5 people have posted in this thread about substantial gain from this, on the avg 200mhz or so. That's enough proof for me.
Oh yeah, and to those guys who risked destroying their brand new expensive X2's to find this out: Much respect and mad props. :clap: