Same VID as my L644 chip

1344 FFT is like an acceleratedsystem stability study, so that you can test stability in a shorter time with Prime95.
Say if you can passed the 1344 with 90%+ available RAM for 15 mins, then your system should be stable for 12 hours of 4k FFT.

This post has been edited by owikh84: Feb 15 2017, 06:45 AM

Nice chip

Here's mine...


Somehow 5.2GHz @ 1.424v freezes in Cinebench R15 so I think better stop here.
Got an offer for a 7700K binned at 5.2GHz @ 1.3v stable R15, $450 USD
Hmm, I think I will just wait for Ryzen first

Exactly, bcos of the identical core performance between KBL and Skylake.
But to get that 6700K overclocked to more than 4.8GHz is quite a challenging task LOL.

Draw of luck.
For record, only 56% of 7700K has ability to do 5.0GHz+.
6700K would be less and rare.



Source: https://siliconlottery.com/

I am using a custom watercooling solution as follows:

Looks like this, except that the graphics card is now air-cooled


This post has been edited by owikh84: Feb 22 2017, 09:17 PM

Nice OC, in future please also include a screenshot of the CPU-Z's motherboard tab in order to get listed in the Post #.
Sounds like you will need a better cooling. Here's my 6600K @ 4.5GHz @ 1.370v cooled by custom WC of course :




It is normal to have those vCore reading in Windows different from the value set through the UEFI BIOS.
It is different from one board vendor to another.
vCore shown in OS-based software is real time and correct one, but the most accurate is of course using a calibrated digital multi-meter (DMM).
The differences is also contributed by the level of CPU voltage Load-line Calibration (LLC) you're using.
Different LLC level will give you higher vdroop (lower vCore at load) or vdrop (higher vCore when loading). For instance,

My 6600K OC 4.5GHz BIOS set 1.35v, LLC 7, but inside Windows CPU-Z reads 1.360v (Idle) and 1.376v (Load).

This post has been edited by owikh84: Mar 2 2017, 07:22 AM

Very nice bro, welcome to the Club

This post has been edited by owikh84: Mar 2 2017, 10:49 PM

Every brand has different LLC behaviour, idle and load voltages.
I suggest you to try the LLC levels one by one.

Use Realtemp or Coretemp to monitor core temps, not CPU package temp like certain software.

Here's my 6700K running at 4.5GHz @ 1.312v, cooled by a custom watercooling kit comprising of dual 360 radiators:

CPU temp during gaming was hovering around 60-70c depending on the game titles.

Temp fluctuates as the CPU load level goes up and down. Look at this example:

What's your motherboard, RAM and cooling?
Can't wait to see your OC results

Benching time...



7700K
ROG Maximus IX APEX
G.Skill TridentZ RGB DDR4-4266MHz 2x8GB







Stock memory frequency - DDR4-4266MHz CL19-19-19-39-2T @ 1.40v



Board and RAM are strong overclockers, managed to push the memory to DDR4-4300MHz without touching the stock voltage of 1.40v


This post has been edited by owikh84: Apr 8 2017, 07:39 AM

Lol, so you're the one who ordered the RAM from IdealTech?

OCbility is huge on this Samsung B-die memory kit. The most important thing is you need to play a lot with VCCSA and VCCIO voltages.

But first of all, have you managed to get the RAM run at its specified frequency as of XMP @ DDR4-4266 C19?

This post has been edited by owikh84: Apr 8 2017, 10:40 AM

Kaby Lake doesn't like 1T and requires voltage bump, addition training, and skill that's why I don't bother using it. 2T for now.
Buying a 4266 kit but running at 4133 IMO kind of waste of money, you should just take the 4133 instead of this, just saying.

Today decided to make a nameplate for ROG Maximus IX APEX.



Keep pushing it!

Pretty sure you can do that manually bro
How much is your 7700K's load voltage @ 5.1?
What's your chip's VID bro? (Read through UEFI with everything default and power saving features disabled)

Here's what I found after few days playing with the 4266 kit

1) with XMP profile loaded at 4266 19-19-19-39 @ 1.40v, the RAM gives a lot of 55 code at reboot. This could be due to the PCB with the LED causing unncessary interferences or the mobo (BIOS) is not fine-tuned to achieve such a high mem frequency
2) the RAM runs fine with perfect reboot at 4133 19-19-19-39 @ 1.35v
3 G.Skill RGB software is still in beta stage, so it has compatibility/conflict issues with ASUS Aura software in which both software can't be used at the same time. In my case, I just use ASUS's software to control both mobo and RAM

Well I don't buy this kit just to see it will be downclocked
Keep pushing

This post has been edited by owikh84: Apr 9 2017, 08:02 PM

First Pentium G4560 here!
I used to own the same Vengeance LED 3200C16 kit but never really had the time to push it until I decided to let it go.

The label indicated that it's using Samsung ICs, but no idea what's the specific type of die it used.

Realbench 85C max here at 5.0GHz under custom watercooling:


Well, OC stability is still a debatable topic since ages ago because of different experiences by different overclockers.
You can choose to skip stability testing, gaming without instability issue and call it stable, but some other people might think achieving 100% stability is a mandatory because they need it for their work. I'd say it's up to you to decide what you need.

RealBench uses AVX instructions thus generating a lot of heat. If you're not using AVX-based programs then there's no use of it. Gaming won't take any benefit of it. Like acther said, you can use the AVX offset to "control" the heat. For instance, 5.0GHz OC with AVX offset of 2 will reduce the frequency to 4.8GHz hence reducing the heat. For gaming you still get 5.0GHz with no speed reduction.