Deliding and lapping my i5 3570K (3770K update) - lowering my temps

[Folterknecht 2,196]

before:

thermal "paste" under the IHS :

after:

... -30°C
 

[Sipher351 241]

3 hours ago, Folterknecht said:

<snip>

Nice! How high did you get it? Broke 5k I hope.

[Folterknecht 2,196]

6 hours ago, Sipher351 said:

Nice! How high did you get it? Broke 5k I hope.

I just completed the operation yesterday and didn't have the time to test out the chip. Before I delidded it, I was running it @1.26V and 4.7GHz  and it seemed stable, though hot as you can see from the first screenshot. That 's why I didn't run extensive stability tests, but the system made a stable impression.

 

Yesterday after delidding and producing the "after" screenshot, I tested 5GHz for around 45min with Prime small FFTs and 1.35V. Just punshed in the numbers and it didn't crash, bluescreen or produced stalled workers in Prime95. But I can't say if it's stable or what voltage it really needs for certain speeds yet.

 

Edit:

5GHz @1.35V resulted in 70°C.

[Folterknecht 2,196]

On 3.1.2018 at 4:30 AM, Sipher351 said:

Nice! How high did you get it? Broke 5k I hope.

Yep - hit's 5GHz stable with 1.35V.

---------------------------------------------

 

 

[Fulcrous 1,979]

Mfw you hit 5.0 but I can't with kaby lake even after a delid...

Fucking silicone lottery.

[Folterknecht 2,196]

5 hours ago, Fulcrous said:

Mfw you hit 5.0 but I can't with kaby lake even after a delid...

Fucking silicone lottery.

golden sample ... the first since my old Q9550 E0

[Fulcrous 1,979]

So apparently my previous BIOS was screwing up my overclock headroom. It caused random instabilities anything above 4.7GHz.
I ended up updating to latest BIOS that alleged to fix those issues. Could boot at 5.0GHz @ 1.37. Many series of p95 BSODs later and I crashed at 1.4v 15minutes in.

Currently stress testing 5.0GHz @ 1.41v
#fingerscrossed

Update: Considering stable for now until I do a long stream session.

[Folterknecht 2,196]

1.4+V with that chip and you 'll probably see long term degradation. Also check for WHEA errors in event viewer from time to time - as long as these occur the system isn't really 100% stable.