1. TechSpot is dedicated to computer enthusiasts and power users. Ask a question and give support. Join the community here.
    TechSpot is dedicated to computer enthusiasts and power users.
    Ask a question and give support.
    Join the community here, it only takes a minute.
    Dismiss Notice

Intel Core i7-9700K tips up on Geekbench

By LemmingOverlrd ยท 14 replies
Sep 5, 2018
Post New Reply
  1. It seems Intel's upcoming Core i7-9700K CPU, this season's refresh of its Coffee Lake architecture, has tipped up on Geekbench revealing some interesting performance numbers.

    First of all, let's recap the Core i7-9700K: this unreleased CPU is based on Intel's latest optimization of the Coffee Lake architecture. '9th' generation Core architecture, and recent leaks indicate it is a 3.6GHz/4.9GHz CPU with eight physical cores and no hyperthreading. It is rated as a 95W CPU. More recently, there's been an ongoing debate on whether the 9700K (and its big brother, the Core i9 9900K) has its IHS soldered on, rather than pasted with TIM.

    According to the CPU scores on Geekbench, the processor reached 6297 points in single-core performance and 30152 points in multi-core performance. This is barely a 4% improvement over the 8700K, but represents a comfortable (~20%) lead over the Ryzen 2700X in a similar configuration, greatly due to its high single-core turbo boost of 4.9GHz.

    Looking at multi-core performance we see the 9700K performance lead quickly erodes. Its 30152 score doesn't quite reach the Ryzen 2700X's 30552 (that's less than a 1% difference). However, if you take into consideration it does not sport hyperthreading, the 9700K doesn't fare too poorly. Intel's eight physical cores are doing the work of 16 logical cores on the Ryzen 2700X.

    Here, it seems, is where Intel's 14nm++ process comes into play. Intel is pushing the envelope on this node and if the leaked specs on the Core i7 9700K are to be believed, then all this performance on a 95W design is quite impressive.

    IPC continues to be Intel's main lead over AMD, and with the Coffee Lake R optimization, Intel seems to have kept AMD at bay, for now. Of course, other factors -- apart from actual benchmarking -- will come into play when the processor finally releases, and knowing Intel we expect this CPU will carry a hefty price tag, as Intel is traditionally uncompromising on its margins.

    Permalink to story.

  2. Gmn17

    Gmn17 TS Rookie Posts: 17

    Awesome news
  3. Puiu

    Puiu TS Evangelist Posts: 3,434   +1,889

    The improvements in Intel's 14nm++ process seem to be fairly good. They'll need every bit of extra power efficiency and performance to keep up with future 7nm competition, at least until their own 10nm process is finished.
    Reehahs likes this.
  4. hood6558

    hood6558 TS Evangelist Posts: 353   +110

    Surprising that even with twice as many threads, the Ryzen only beats it by 1% in multi-threaded tests. Better luck next time AMD. Maybe Ryzen 2 will be the miracle your fans keep hoping for. But by then, Intel's 10nm will be out, and setting new records for IPC and clock speed.
  5. Puiu

    Puiu TS Evangelist Posts: 3,434   +1,889

    The difference in clocks speeds is really big and it's just one test. There are benchmarks and applications that can or cannot take advantage of the extra threads. We'll just have to wait and see the final numbers to make a proper comparison. I also want to see the price point at which the 8700k will launch.
    Theinsanegamer and Evernessince like this.
  6. Evernessince

    Evernessince TS Evangelist Posts: 3,993   +3,479

    Take it with a grain of salt. The 8700K is pretty close to the 2700X on geekbench. There's a reason it's not a go to test for multi-core.

    This has only been the 5th amazing performance rumor this month. people need to calm down and wait for real benchmarks.
    meric likes this.
  7. erickmendes

    erickmendes TS Evangelist Posts: 572   +247

    Cost x Benefit is king to me... so I still would choose AMD, even for gaming.
    i3kingwizardop and avioza like this.
  8. Theinsanegamer

    Theinsanegamer TS Evangelist Posts: 1,547   +1,767

    Competition is a miraculous thing.
  9. Lew Zealand

    Lew Zealand TS Guru Posts: 663   +553

    Not sure what to think here but Geekbench is a low level series of tests which for my uses are more valuable assessing relative performance within a similar series of processors. Ie: comparing all Skylake or Ryzen 1. And it's useful for CPU and memory clocking to quickly confirm your clocking changes are actually having an effect. But as a real world performance indicator? Only in a very general sense.
  10. pcnthuziast

    pcnthuziast TS Evangelist Posts: 584   +193

    Intel shady marketing at work. The 9700 and 9900 are identical physically if I'm not mistaken, but have different firmware. 1 physical chip, 2 skus. They're playing their position well based on what AMD has out though.
  11. Lew Zealand

    Lew Zealand TS Guru Posts: 663   +553

    Isn't that how chip binning works in general? I don't think all desktop Intel i5s are merely i7s with "defective" HT. I assume they're less power efficient/stable at higher frequencies so they're binned lower and have HT turned off.

    It's all artificial product segmentation in one way or another, so I wonder if people's underwear bunching would be quite so pronounced if Intel had just called the exact same product an i5-9700.
  12. i3kingwizardop

    i3kingwizardop TS Booster Posts: 41   +17

    We need a new unit of measurement. Instead of Instructions Per Clock, we need Instructions Per Dollar, aka bang for the buck. Looking forward to AMD's answer in the 2800X.
  13. enemys

    enemys TS Maniac Posts: 180   +175

    "IPC continues to be Intel's main lead over AMD"
    That's not true, since IPC in most workloads is virtually the same, as TechSpot itself has measured: https://www.techspot.com/article/1616-4ghz-ryzen-2nd-gen-vs-core-8th-gen/page5.html
    Intel's main advantage is clock speed - both with and without OC they can reach about 20-25% higher clocks, which mostly explains the performance parity between the processors.
    Lew Zealand likes this.
  14. Lew Zealand

    Lew Zealand TS Guru Posts: 663   +553

    True, 3% IPC difference is a lot less than 20% clock speed difference. Funny how you can know the details but miss the big picture.
  15. prodrom

    prodrom TS Rookie

    All this controversy is because of geekbench. It is not better and most of the time it's worse than legacy benches. It brings down real world performance to about 20 percent of the theoretical performance of Intel and amd cpus. This is simply ridiculous.
    We are in a weird period during which cpu manufacturers try to hide in plain sight. If they were decent , they would say :"look , theoretical perf is x because ipc is y , frequency is z bla bla bla..." Most cpus reach about 75 percent of their theoretical peaks with some significant variations, but I have never seen one that can only perform at 15 to 20 percent. This is retarded.
    For what it's worth(Even though companies try to hide) , amd cpus must have a theoretical 16 dp per cycle and 32sp per cycle. Intel's core iI7s which do not have AVX - 512 have the same performance. So as others have pointed out Intel's lead is mainly due to higher frequencies, barring xeons and higher desktop implementations like 7980 xe which do have avx-512.

Add your comment to this article

You need to be a member to leave a comment. Join thousands of tech enthusiasts and participate.
TechSpot Account You may also...