CPU Benchmark Performance: E-Core

In this batch of testing, we're focusing primarily on the E-cores. Intel claimed that the performance was around the level of its Skylake generation of processors (6th Gen to 10th Gen, depending which slide you read), and we had to put that to the test. In this instance, we're comparing to the flagship Skylake processor, the Core i7-6700K, which offered 4C/8T at 91 W. We also did a number of multi-threaded tests to see where the E-cores would line up.

In order to enable E-core only operation, we used affinity masks.

Single Threaded

(3-2b) Dwarf Fortress 0.44.12 World Gen 129x129, 550 Yr(3-3) Dolphin 5.0 Render Test(4-8a) CineBench R20 Single Thread

(8-1c) Geekbench 5 Single Thread

In these few tests, we can see that the E-core is almost there at 4.2 GHz Skylake. Moving down to 3.9 GHz, perhaps something like the i7-6700, would put it on par. 

Multi-Thread Tests

(1-1) Agisoft Photoscan 1.3, Complex Test(2-1) 3D Particle Movement v2.1 (non-AVX)(2-2) 3D Particle Movement v2.1 (Peak AVX)(2-5) NAMD ApoA1 Simulation(2-6) AI Benchmark 0.1.2 Total(3-1) DigiCortex 1.35 (32k Neuron, 1.8B Synapse)(4-2) Corona 1.3 Benchmark(4-3a) Crysis CPU Render at 320x200 Low(4-5) V-Ray Renderer(4-8b) CineBench R20 Multi-Thread(5-1a) Handbrake 1.3.2, 1080p30 H264 to 480p Discord(5-1b) Handbrake 1.3.2, 1080p30 H264 to 720p YouTube(5-1c) Handbrake 1.3.2, 1080p30 H264 to 4K60 HEVC(5-2c) 7-Zip 1900 Combined Score(5-3) AES Encoding(5-4) WinRAR 5.90 Test, 3477 files, 1.96 GB(7-1) Kraken 1.1 Web Test(7-2) Google Octane 2.0 Web Test(7-3) Speedometer 2.0 Web Test(8-1d) Geekbench 5 Multi-Thread

Having a full eight E-cores compared to Skylake's 4C/8T arrangement helps in a lot of scenarios that are compute limited. When we move to more memory limited environments, or with cross-talk, then the E-cores are a bit more limited due to the cache structure and the long core-to-core latencies. Even with DDR5 in tow, the E-cores can be marginal to the Skylake, for example in WinRAR which tends to benefit from cache and memory bandwidth.

CPU Tests: SPEC MT Performance - P and E-Core Scaling CPU Benchmark Performance: Windows 11 vs Windows 10
Comments Locked

474 Comments

View All Comments

  • Netmsm - Monday, November 8, 2021 - link

    @Wrs Assume we give you this, just drop AMD related things.
    Let's focus on performance per watt of new Intel Architecture. I asked some questions about this which is a pertinent topic to this article but you dodged. Just talk about 241 watt being sucked by 12900k :))
  • ajollylife - Sunday, November 7, 2021 - link

    To be fair Intel's def not perfect. The i225v lan bug comes to mind
  • DominionSeraph - Saturday, November 6, 2021 - link

    When Intel doesn't have this issue it is AMD.
    Intel demands validation and spends hundreds of millions on engineers to help motherboard manufacturers fix their issues before release because they realize that the motherboard is an INSEPERABLE part of the equation. You can fanboi all you want over the CPU, but if there are no good motherboards you have a terrible system.
    AMD just dumps their chip on the market, leaving the end-user with a minefield of motherboard issues among every manufacturer.
    When Intel has the problem solved but AMD doesn't, it's an AMD issue.
  • Qasar - Saturday, November 6, 2021 - link

    DominionSeraph have you tried this on a different x570 board ? it COULD be just that board, as was already stated. i have been running a 5900x now for about a week, and no issues at all with disk access, its the same as the 3900x i replaced, and im using an asus x570 e-gaming board.
    maybe take a step back, realize that its NOT an amd issue, but a BOARD issue. to be fair, intel has had its own issues over the years as well, while you see them as perfect, sorry to break it to you, but intel also isnt perfect
  • Netmsm - Sunday, November 7, 2021 - link

    Oh I got it, the fact that you're encountering unusual issues, despite myriads of people who haven't complained of the same problem you have, means that AMD got many incompatibility issues.
    That's a brilliant reasoning which brings us to a pure conclusion: AMD is bad. We're done here.
  • Dug - Monday, November 15, 2021 - link

    @DominionSeraph
    "Note Intel doesn't allow "dog sh1t motherboards" to happen, especially at the $300+ price point. That makes it an AMD issue."

    You obviously haven't looked at or experienced the issues that plague many motherboards. Maybe just read some reviews or visit manufactures forums to get an idea of how bad it is before making such a naïve statement.
    It has more to do with the manufacturer of the board than the CPU.
    That is like saying nvidia doesn't make crap video cards. Well you need to look at all the failures by each manufacturer, not nvidia itself.
  • madseven7 - Saturday, November 6, 2021 - link

    You put that chip in an ASROCK motherboard. That's the reason for the issues.
  • xhris4747 - Tuesday, November 9, 2021 - link

    Lol 😅
  • FLORIDAMAN85 - Friday, November 12, 2021 - link

    I, too, have made the error of purchasing an ASRock board.
  • mode_13h - Saturday, November 13, 2021 - link

    So far, my ASRock Rack server board is still going strong.

Log in

Don't have an account? Sign up now