The Intel Skylake-X Review: Core i9 7900X, i7 7820X and i7 7800X Tested
by Ian Cutress on June 19, 2017 9:01 AM ESTComparing 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:
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.
Facebook
Twitter
RSS
264 Comments
View All Comments
Gothmoth - Monday, June 19, 2017 - link
i don´t care about powerdraw that much if i can COOL the CPU and keep the cooling quiet.but in this case the powerdraw is high and the heat is crazy.
and all because of intel insisting to save a few dollar on a 1000 dollar CPU and use TIM?
WTF....
Ej24 - Monday, June 19, 2017 - link
I wish amd would have released Threadripper closer to ryzen. That way amd wouldn't make comparisons of ryzen to Intel x99/x299. They kind of shot themselves in the foot. AM4 is only directly comparable to lga115x as a platform. R3, 5 and 7 are only really intended to compete with i3, 5, and 7 consumer parts. Amd simply doubled the core count per dollar at the consumer line. It's merely coincidental at this point that ryzen core count lines up with Intel HEDT. The platforms are not comparablein use case or intent. All these comparisons will be null when Threadripper/x399 is released as that is AMD's answer to x299.Ej24 - Monday, June 19, 2017 - link
how is the 7740x, 112w tdp only drawing 80w at full load? I understand that tdp isn't power draw but thermal dissipation. However the two values are usually quite close. In my experience, max turbo power consumption surpasses the tdp rating in watts.For example, my 88w tdp 4790k consumes 130w at max 4 core turbo. My 4790S a 65w tdp consumes 80w at max 4 core turbo. My 4790t, 45w tdp, consumes 55w at max 4 core turbo. So how is it the 7740x consumed 80W at max utilization??
AnandTechReader2017 - Tuesday, June 20, 2017 - link
Agreed as on http://www.anandtech.com/show/10337/the-intel-broa... the all-core load for the i7 6950X the all-core load is 135W yet on this graph it's 110W. Something is wrong with those load numbers.Ian Cutress - Tuesday, June 20, 2017 - link
It's consumer silicon running a single notch up the voltage/frequency curve. Probably binned a bit better too. 112W is just a guide to make sure you put a stonking great big cooler on it. But given the efficiency we saw with Kaby Lake-S processors to begin with, it's not that ludicrous.Flying Aardvark - Monday, June 19, 2017 - link
This is an interesting time (finally), again in CPUs. To answer the question you posed, "Ultimately a user can decide the following". I decided to go mini-ITX this time. Chose Ryzen for this, and initially the 1800X. Had to downgrade to the 1700 due to heat/temps, but overall I don't think anything competes against AMD at all in the Node202 today.That's one area where Intel is MIA. Coffeelake will be 6C/12T, 7700K is 4C/8T. R7-1700 is 65W and 8C/16T. Works great. I paired mine with a 1TB 960 Pro and Geforce 1060 Founders Edition.
If I moved to anything else, it would be all the way to 16C/32T Threadripper. I'm really unimpressed by this new Intel lineup, power consumption and heat are simply out of control. Dead on arrival.
Gothmoth - Monday, June 19, 2017 - link
what mobo and ram did you use? is your ryzen build really stable?i need full load stability 24/7.
Flying Aardvark - Monday, June 19, 2017 - link
What, you don't need just 60% stability? Yes it's stable.I did have one bluescreen and it was the Nvidia driver. I think it's unlikely most people would run into whatever caused it, because I use a triple monitor setup and lots of programs / input switching, and it crashed upon a DisplayPort redetection.
I bought the Geforce 1060 because it was the most efficient and well-built blower fan cooled GPU I could find. But buying again, I'd go for the best Radeon 480/580 that I could find.
I never had a bluescreen for decade running Intel CPUs and AMD GPUs so I dislike changing to AMD CPUs and Nvidia GPUs.. but I think it's safest to run a Radeon. Just less likely to have an issue IMO.
Other than that, no problems at all. Rock solid stable. I used the Biostar board and G.Skill "Ryzen" RAM kit.
Gothmoth - Tuesday, June 20, 2017 - link
it´s something different if as system is stable for 2-3 hours under load or 24/7 under load.. capiche? :-)Gothmoth - Tuesday, June 20, 2017 - link
btw... thanks for your answer.i use a triple monitor setup and use many programs at once... what sense would a 8-10 core make otherwise. :-)