Comparing Skylake-S and Skylake-X/SP Performance Clock-for-Clock

If you’ve read through the full review up to this point (and kudos), there should be three things that stick in the back of your mind about the new Skylake-SP cores: Cache, Mesh and AVX512. These are the three main features that separate the consumer grade Skylake-S core from this new core, and all three can have an impact in clock-for-clock performance. Even though the Skylake-S and the Skylake-SP are not competing in the same markets, it is still poignant to gather how much the changes affect the regular benchmark suite.

For this test, we took the Skylake-S based Core i5-6600 and the Skylake-SP based Core i9-7900X and ran them both with only 4 cores, no hyperthreading, and 3 GHz on all cores with no Turbo active. Both CPUs were run in high performance modes in the OS to restrict any time-to-idle, so it is worth noting here that we are not measuring power. This is just raw throughput.

Both of these cores support different DRAM frequencies, however: the i5-6600 lists DDR4-2133 as its maximum supported frequency, whereas the i9-7900X will run at DDR4-2400 at 2DPC. I queried a few colleagues as to what I should do here – technically the memory support is an extended element of the microarchitecture, and the caches/uncore/untile will be running at different frequencies, so how much of the system support should be chipped away for parity. The general consensus was to test with the supported frequencies, given this is how the parts ship.

For this analysis, each test was broken down in two ways: what sort of benchmark (single thread, multi-thread, mixed) and what category of benchmark (web, office, encode).

 

For the single threaded tests, results were generally positive. Kraken enjoyed the L2, and Dolphin emulation had a good gain as well. The legacy tests did not fair that great: 3DPM v1 has false sharing, which is likely taking a hit due to the increased L2 latency.

On the multithreaded tests, the big winner here was Corona. Corona is a high-performance renderer for Autodesk 3ds Max, showing that the larger L2 does a good job with its code base. The step back was in Handbrake – our testing does not implement any AVX512 code, but the L3 victim cache might be at play here over the L3 inclusive cache in SKL-S.

The mixed results are surprising: these tests vary with ST and MT parts to their computation, some being cache sensitive as well. The big outlier here is the compile test, indicating that the Skylake-SP might not be (clock for clock) a great compilation core. This is a result we can trace back to the L3 again, being a smaller non-inclusive cache. In our results database, we can see similar results when comparing a Ryzen 7 1700X, an 8-core 95W CPU with 16MB of L3 victim cache, is easily beaten by a Core i7-7700T, with 4 cores at 35W but has 8MB of inclusive L3 cache.

If we treat each of these tests with equal weighting, the overall result will offer a +0.5% gain to the new Skylake-SP core, which is with the margin of error. Nothing too much to be concerned about for most users (except perhaps people who compile all day), although again, these two cores are not in chips that directly compete. The 10-core SKL-SP chip still does the business on compiling:

Office: Chromium Compile (v56)

If all these changes (minus AVX512) offer a +0.5% gain over the standard Skylake-S core, then one question worth asking is what was the point? The answer is usually simple, and I suspect involves scaling (moving to chips with more cores), but also customer related. Intel’s big money comes from the enterprise, and no doubt some of Intel’s internal metrics (as well as customer requests) point to a sizeable chunk of enterprise compute being L2 size limited. I’ll be looking forward to Johan’s review on the enterprise side when the time comes.

Benchmarking Performance: CPU Legacy Tests Intel Skylake-X Core i9-7900X, i7-7820X and i7-7800X Conclusion
POST A COMMENT

264 Comments

View All Comments

  • FreckledTrout - Monday, June 19, 2017 - link

    Missing the 7820x on the power draw graph. Reply
  • Ian Cutress - Tuesday, June 20, 2017 - link

    The 7820X power numbers didn't look right when we tested it. I'm now on the road for two weeks, so we'll update the numbers when I get back. Reply
  • chrysrobyn - Monday, June 19, 2017 - link

    In my head I'm still doing the math on every benchmark and dividing by watts and seeing Zen looking very different. Reply
  • Old_Fogie_Late_Bloomer - Monday, June 19, 2017 - link

    I'm sure I'm wrong about this, but it makes more sense to me that the i9-7900X would be a (significantly) cut down HCC die instead of a perfect LCC. i9 vs i7, 44 vs 28 lanes, two AVX units instead of one?

    And yet the one source I've found so far says it's the smaller die. It's definitely the LCC die, then?
    Reply
  • Ian Cutress - Tuesday, June 20, 2017 - link

    HCC isn't ready, basically. LCC is. Plus, having a 10C LCC die and not posting a top SKU would be wasteful of the smallest die of the set.

    Also, delidding a 10C SKU.
    Reply
  • Old_Fogie_Late_Bloomer - Tuesday, June 20, 2017 - link

    Well, it wouldn't be a waste if Intel's yields weren't good enough to get fully functional dies. The fact that Intel is not just releasing fully functional LCC chips but announced that they would be the first ones available suggests that they have no trouble reliably producing them, which is pretty impressive (though they have had plenty of practice on this process by now).

    Thanks for the response; I thoroughly enjoyed the review and look forward to further coverage. Exciting times!
    Reply
  • Despoiler - Monday, June 19, 2017 - link

    Considering Ryzen is in the desktop category and these Intel chips are HEDT, we need to wait to see what Threadripper brings. AMD won't have the clock advantage, but for multithreaded workloads I suspect they will have more cores at a cheaper price than Intel. Reply
  • FreckledTrout - Monday, June 19, 2017 - link

    I wouldn't say AMD wont have a clock advantage once you get to the 14 and 16 core chips. They might not but you saw the power numbers and thermals, Intel very well may have to pull back the frequency as they scale up the cores more than AMD will. Reply
  • FMinus - Thursday, June 22, 2017 - link

    Actually I think it's the other way around. AMD might have clock advantage on higher core models thanks to not going with the monolithic approach. Easier to to cool those beasts but power is still an issue.

    If you imagine four 1800x on one interposer, you can see them reaching 4GHz on all of those dies, that said the power consumption would be massive, but easier cooler as the intel 16 core variant.
    Reply
  • Lolimaster - Tuesday, June 20, 2017 - link

    The 1995X will have a stock 3.6Ghz for the 16cores, same as the 7900X with just 10. Reply

Log in

Don't have an account? Sign up now