Follow Slashdot blog updates by subscribing to our blog RSS feed

 


Forgot your password?
Close
typodupeerror
×
Intel Stats Upgrades Hardware

Skylake Breaks 7GHz In Intel Overclocking World Record (hothardware.com) 85

Posted by timothy from the of-course-the-handlers-will-get-the-credit dept.
MojoKid writes: Intel's latest generation of processors built on the Skylake architecture are efficient as well as seriously fast. The flagship, Core i7-6700K, is an interesting chip as it's clocked at a base 4GHz, and can peak at 4.2GHz with Turbo Boost. Of course, as fast as the 6700K is, overclocking can always help take things to the next level, or at least temporarily explore future potential. In Chi-Kui Lam's case, he did just that, and managed to break a world record for Intel processors along the way. Equipped with an ASRock motherboard, G.SKILL memory, and a beefy 1.3KW Antec power supply — not to mention liquid nitrogen — Lam managed to break through the 7GHz barrier to settle in at 7025.66MHz. A CPU-Z screenshot shows us that all cores but one were disabled — something traditionally done to improve the chances of reaching such high clock speeds.
This discussion has been archived. No new comments can be posted.

Skylake Breaks 7GHz In Intel Overclocking World Record More

Skylake Breaks 7GHz In Intel Overclocking World Record

Comments Filter:

  • About 4 times less performance than without OCing (Score:5, Insightful)

    by Nikademus ( 631739 ) * <renaud.allard@it> on Tuesday February 09, 2016 @10:09AM (#51469227) Homepage

    This is a very fun experiment as it lets you play with liquid nitrogen.
    However, the CPU has now only 1 core instead of 8 and only about 1.6 times the clock frequency. This means a huge decrease in performance...

    • "4 times less performance" than what?
      I HATE that wording on tech and science sites.
      1/4 the performance is the correct wording, unless comparing to a difference.

      • Re: (Score:1)

        by Anonymous Coward

        4 times less... times is multiplication, less is subtraction.
        So without overclocking, its performance is a base of "100%".
        After overclocking, its performance is 100% - (4 * 100%), making it -300% now.

    • Re: (Score:2)

      by Junta ( 36770 )

      It's actually only 4 cores, so the super fast overclocking isn't as bad as you say, it's almost half as fast as pre-overclock.

    • Re: (Score:2)

      by Kjella ( 173770 )

      However, the CPU has now only 1 core instead of 8 and only about 1.6 times the clock frequency. This means a huge decrease in performance...

      Amdahl's law says that depends on what you're doing. Also it has 4 cores/8 threads but yeah. This is obviously just for doing it. As someone who started with a 0.985 MHz C64 and got a 1.2GHz Athlon not so long into the new millennium I'm quite underwhelmed though, despite the IPC improvements.

    • I'm curious also if that one core was even running at full tilt. Processors these days sleep by default. So is this test one of "boot up and run some cpu measuring utility" or did it at least run at 100% utilization for a while?
    • You're making the ridiculously optimistic assumption that users are running software that can actually make use of more than one core. (It's 2016; why do all mainstream languages still generate single-threaded code by default?)
    • Circuits using liquefied gases for cooling passed 20 GHz over 40 years ago, so this fun experiment is the equivalent of strapping a JATO onto a bicycle.
  • The skylake broke the record for OCing a skylake CPU, big whoop.

    • Re: (Score:3, Informative)

      by chris200x9 ( 2591231 )
      Forgot link http://www.engadget.com/2007/0... [engadget.com] so TFA is wrong, it doesn't even beat records for and intel processor it literally is just a new skylake record.

      • The Intel 10GHz prediction (Score:4, Interesting)

        by Gr8Apes ( 679165 ) on Tuesday February 09, 2016 @10:58AM (#51469695)
        In 2000, Intel predicted they'd have 10GHz chips by 2011. Here it is, 2016, and they've been going backwards since around 2006 with a peak roughly 3.8GHz production chip.
        • Comment removed based on user account deletion

          • Re: (Score:1)

            by Gr8Apes ( 679165 )
            Intel had to be beaten in the x86 arena for that to matter. There were all sorts of better designs around at that time. It kind of goes back to Intel not really being all that good of a chip nor innovator. They were taking lessons from their pal in announcing vapor to keep people from jumping ship. Sadly, it worked again.

  • Barrier? (Score:3)

    by AmiMoJo ( 196126 ) on Tuesday February 09, 2016 @10:23AM (#51469345) Homepage Journal

    What is this 7GHz barrier? Is it like the sound barrier or something for CPUs?

    • It's like deja vu. The calculations travel back in time, so they have to do it all over again. It's like deja vu. The calculations travel back in time.

    • Re:Barrier? (Score:4, Funny)

      by Joe_Dragon ( 2206452 ) on Tuesday February 09, 2016 @10:42AM (#51469497)

      Need to get it to 88Ghz

    • Re: (Score:2)

      by tibit ( 1762298 )

      That's the clock multiplier limit on that chip. You can't seem to be able to set it any higher...

    • A practical limit for silicon-based CPUs. I've been told that military uses a different semiconductor material to run CPUs at 100GHz at a much higher temperatures.

      • A practical limit for silicon-based CPUs. I've been told that military uses a different semiconductor material to run CPUs at 100GHz at a much higher temperatures.

        I'm not sure they are CPUs in the same sense. You can easily find simpler circuits that operate at such frequencies, e.g. microwave amplifiers, but a modern CPU involves much more than the raw switching speed of transistors. Keeping the core in sync with itself will be harder with a wavelength of 3 mm (This would be for 100 GHz in vacuum, in a solid it would be even less).

      • https://www.comsol.com/blogs/h... [comsol.com]

        Essentially the answer is transistor noise, then transistor count, and heat. (Not the Speed of Light [reddit.com]).

        Because power increases with (clock speed)^3, but executing speed only increases by (clock speed)^1, people go for parallel processors.

    • It's the same as the 90 mph wind barrier for buses. [wikipedia.org]

    • We do seem to have been stuck at a max of 4G processor speed for a number of years. Has Moore's law been declared dead?

      • Re: (Score:3)

        by fnj ( 64210 )

        What the hell does Moore's law have to do with GHz? All it is, is an observation on the growth over time of the number of transistors possible to build into a dense IC. As far as I can tell, it is still operative.

      • We do seem to have been stuck at a max of 4G processor speed for a number of years. Has Moore's law been declared dead?

        No, and neither is Moore... He's 87. I don't think he will live long enough for his law to be declared dead..

  • Not the overclocking record (Score:5, Informative)

    by hajile ( 2457040 ) on Tuesday February 09, 2016 @10:32AM (#51469423)

    I believe the official Guinness record is 8.429GHz on an AMD pre-release bulldozer in 2011. Another record was set at 8.723GHz on an AMD FX-8370 in 2013, but I don't recall it being "official".

    http://www.tomshardware.com/news/bulldozer-amd-overclock-guinness-record,13431.html

    http://www.theregister.co.uk/2014/09/02/amd_fx_series/

    • It does say 'Intel Overclocking World Record'. Which is just silly. One can assume that Intel will always hold the Intel Overclocking World Record.

  • Yeah that's feasible for everyday use with all but one core disabled. Not. This is a plateau of unusable.

    I'm curious to see how it performs with all cores enabled at that multiplier/voltage?

  • Apparently AMD still holds the record though (Score:4, Informative)

    by The-Ixian ( 168184 ) on Tuesday February 09, 2016 @10:49AM (#51469575)

    As impressive as 7GHz is, it's not the highest frequency we've ever seen. Over at HWBOT, we can see that the overall world record belongs to The Stilt, who pushed an AMD FX-8370 to a ridiculous 8,722.78MHz

    • I was about to quote the exact same thing. What world record? What 7GHz barrier? AMD CRUSHED that a long time ago.

  • Overclocked processor pops bag of popcorn in next room.

  • They probably would have hit 10GHz if they got 7GHz on some garbage Asrock board. If you're not familiar, it's a budget spin-off of ASUS and I've had nothing but problems with defective Asrock motherboards at my computer repair shop.
    • Asrock's products are crap at the lower end, but at least from what I've _heard_ their high end stuff is the same as everything else.

    • Considering the problems the genuine Asus boards have given me over the years, I can only imagine how terrible Asrock must be. I'd take my changes with ECS first.

  • It be like: strip a car down to its bare chassis, removing all extraneous weight, seats, anything that's there for comfort or convenience, then keep cutting out more metal - panels, half the floor, roof. Bolt on a huge turbo and nitrous and watch it do a ¼ mile in 8 seconds. Great, but it ain't going to get you to the shops and back.
  • Run it at something like 4-5GHz and save on the power consumption. Hope that the multi-core helps here

  • I remember back when even 4Ghz required liquid cooling. This was when people were doing it to Pentium IV chips. It's surreal knowing that 4Ghz is now the "stock speed" while 7Ghz is the new threshold to reach.

Slashdot Top Deals

Every nonzero finite dimensional inner product space has an orthonormal basis. It makes sense, when you don't think about it.

Close