Extreme Overclocking of Skylake (7.02566 GHz)

Subject: Processors | February 6, 2016 - 09:00 PM |
Tagged: Skylake, overclocking, asrock, Intel, gskill

I recently came across a post at PC Gamer that looked at the extreme overclocking leaderboard of the Skylake-based Intel Core i7-6700K. Obviously, these competitions will probably never end as long as higher numbers are possible on parts that are interesting for one reason or another. Skylake is the new chip on the liquid nitrogen block. It cannot reach frequencies as high as its predecessors, but teams still compete to get as high as possible on that specific SKU.

View Full Size

The current world record for a single-threaded Intel Core i7-6700K is 7.02566 GHz, which is achieved with a voltage of 4.032V. For comparison, the i7-6700K is typically around 1.3V at load. This record was apparently set about a month ago, on January 11th.

This is obviously a huge increase, about three-fold more voltage for the extra 3 GHz. For comparison, the current world record over all known CPUs is the AMD FX-8370 with a clock of 8.72278 GHz. Many Pentium 4-era processors make up the top 15 places too, as those parts were designed for high clock rates with relatively low IPC.

The rest of the system used G.SKILL Ripjaws 4 DDR4 RAM, an ASRock Z170M OC Formula motherboard, and an Antec 1300W power supply. It used an NVIDIA GeForce GT 630 GPU, which offloaded graphics from the integrated chip, but otherwise interfered as little as possible. They also used Windows XP, because why not I guess? I assume that it does the least amount of work to boot, allowing a quicker verification, but that is only a guess.

Source: HWBot

Video News

February 6, 2016 | 10:14 PM - Posted by Anonymous (not verified)

7Ghz wow that is insane. my i7-6700k is running at 4.85Ggz with a air cooler.

February 6, 2016 | 10:44 PM - Posted by Scott Michaud

Yeah. It keeps getting harder and harder to increase frequency after some inflection point.

February 6, 2016 | 11:16 PM - Posted by Esa (not verified)

This is pretty funny though, the clockspeed race has pretty much died off http://hwbot.org/benchmark/cpu_frequency/

February 6, 2016 | 11:45 PM - Posted by Scott Michaud

Yeah, I don't know how much further that graph will go while we still push radio frequency EM through silicon.

February 7, 2016 | 02:22 AM - Posted by Master Chen (not verified)

>Only one core enabled and no HT
That's pretty lame of an "OverClocking", actually. Kiddie stuff.

February 7, 2016 | 05:43 PM - Posted by arbiter

well that is pretty much how all overclocks are done is only 1 core enabled rest turned off. Yea i agree its just an epenis kinda thing where as 2 cores at 4ghz is better but it is what it is.

February 7, 2016 | 07:24 PM - Posted by Scott Michaud

There's categories for all different types of configurations. Extreme overclocking is usually "one thread as high frequency as you can, and get a valid OS boot" though. Often doesn't even include benchmarks as far as I know.

February 8, 2016 | 05:23 AM - Posted by Master Chen (not verified)

Not really. There are truly hardcore actual OverClockers out there who manage to get exceptionally high results with all cores enabled and sometimes even while HT alongside them. That is THE true OverClocking, in my opinion, since, essentially, getting a multi-core chip and them deliberately gimping it down to one core and one thread just to boast on your "high OC result" is nothing more than just a lame cheating (at least in my eyes it is).

February 8, 2016 | 01:23 PM - Posted by Anonymous (not verified)

I disagree. Think of it more like challenging the architecture itself, pushing it to its limits. It's basically the same exact thing as Intel's own turbo mode, but like "turbo turbo".

February 7, 2016 | 01:41 PM - Posted by Anonymous (not verified)

CPU clock speeds are not going to go up, and in fact they are going to be going down with more cores added instead, so I guess that with more of the game code accelerated on the GPU anyways with DX12, and Vulkan, there will not be as much of a need to worry about CPU performance!

To think that by the time AMD gets its Zen Based SKUs to market there will be less of a need for as much CPU side improvements owing to more things gaming being able to to be done on AMD's ACE units via Vulkan/DX12, so AMD may not be at such a disadvantage if it only gets Zen's IPC performance around the Ivybridge-Broadwell levels of IPC!

Where AMD could take the lead is maybe just making its APUs interposer based for its Laptops/PCs(Some versions) SKUs and having the ability to place a more powerful GPU die on the interposer package along with HBM, and more directly connecting the GPU die to the Zen cores with some wide/thousands of traces directly from the Zen cores to the GPU, with lower memory/communication fabric clocks providing higher effective bandwidth than any PCI based connection would allow! At least for laptop APU SKUs based on an interposer with HBM the laptop OEMs would not be able to gimp the memory channels like they did with the Carrizo based laptop SKUs.

An Interposer based Laptop APU SKU would allow AMD to control the CPU to GPU, and the HBM Memory side of the SOC equation and put AMD in control of that part of user's experience, all while making it less expensive for the for the laptop OEMs on the PCB design side if AMD could at least have the ability to have a top end laptop gaming APU SKU support 32GB of HBM Memory, while having 8/16 GB of HBM options for its lower cost interposer based APUs.

Interposer based APUs for laptops will also be able to be much more power efficient owing to the fact that many thousands of wide HBM memory traces, as well as the same for CPU to GPU/other wide communication fabrics will allow for the clocks to be lowered for the Memory, and CPU to GPU connection fabrics, and saving on the power usage as well as the space and PCB complexity savings.

February 8, 2016 | 10:11 PM - Posted by Anonymous (not verified)

This is completely off topic.

February 7, 2016 | 05:06 PM - Posted by Danny Lim (not verified)

It's impossible to reach 7GHz with Air Cooling solution.

February 8, 2016 | 10:17 PM - Posted by Anonymous (not verified)

Unless you would consider liquid air (-317.83 degrees fahrenheit) to be "air cooling".

February 7, 2016 | 06:36 PM - Posted by Anonymous (not verified)

Article commends AMD for overclocking then slams them about IPC in the next breath... Dat bias is showing again

February 7, 2016 | 07:26 PM - Posted by Scott Michaud

Actually, it shows that you're biased enough to not even see "Pentium 4" before the IPC statement. :p ((Edit: Whoops! Accidentally said "your biased" when I changed the sentence from one containing "your bias" to another with "you're biased". :x))

February 8, 2016 | 10:50 AM - Posted by [CoFR]Prodeous (not verified)

To make it easier to see, just place that as a separate paragraph, since when you "speed read" through it, it does feel it relates to FX rather then P4.

Funny part is no mater how badly P4 performed (even worse then AMD's FX file), Intels market manipulation and near monopoly managed to do a lot of damage to AMD.

Still looking at this "Achievement" with Skylake. Seeing how it is just one core without Hyperthreading, makes this "achievement" look like a cheat. FX managed to get it running as the full CPU.

Would have been interesting to see performance results of the 7Ghz 1 core Skylate vs 8Ghz 8 thread FX. :)

February 8, 2016 | 03:21 PM - Posted by Scott Michaud

I should actually look up what the record is for Skylake in a category that requires a benchmark.

February 8, 2016 | 12:19 PM - Posted by Dusty

Many competition overclockers use a very stripped down version of XP so that there is the shortest amount of time, and processes between power up and screenshotting the validation on CPU-Z. They will also boot from a USB drive to bypass the SATA storage subsystem.

February 9, 2016 | 10:30 AM - Posted by UnfriendlyBanana (not verified)

1 core, 1 thread...
Now for the 1 million dollar question.
Will it play Crysis???

April 1, 2016 | 11:31 AM - Posted by phobosq (not verified)

Forgive me the archeology, but I'm surprised that no one noticed CPU-Z has voltage detection bug and it reports vcore at its double value. So for this run "only" 2.016v was needed - still a lot, but far more feasible on 1c/1t with LN2 cooling. This issue is observed on this model as mobo used was in engineering (beta) stage, not yet released to production.
The part of argument about FX and Pentium selected for low IPC makes also little sense to me, they are in top 20 of raw clock rate for three reasons: good scaling with cold, good scaling with voltage and lack of so-called cold bug. Low IPC is a coincidence here, Intel increases IPC on their CPUs since Core 2 Duo, and max clock frequencies on LN2 remain around 7 GHz with each generation.

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Lines and paragraphs break automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd> <blockquote><p><br>
  • Web page addresses and e-mail addresses turn into links automatically.

More information about formatting options

This question is for testing whether you are a human visitor and to prevent automated spam submissions.