Skip to main content

AMD's EPYC Milan Breaks Cinebench Record, Here's a 10nm Ice Lake Xeon Comparison

(Image credit: AMD)

AMD's EPYC Milan processors launched last month with 120 new world records to their credit in various applications, like HPC, Cloud, and enterprise workloads. But variants of these chips will eventually come to the market as Threadripper models for high end desktop PCs, and AMD's server records don't tell us too much about what we could expect from the PC chips. However, the company recently broke the Cinebench world record with its Milan chips, giving us an idea of what to expect in rendering work. Just for fun, we also ran a few tests on Intel's new flagship 40-core Ice Lake Xeon chips to see how they stack up against not only AMD's new record it set with the server chips, but also a single AMD Threadripper processor.

During the latest episode of AMD's The Bring Up YouTube video series, the company took two of its $7,980 EPYC Milan 7763 chips for a spin in Cinbench R23, a rendering application that AMD commonly uses for its desktop PC marketing (largely because it responds exceedingly well to AMD's Zen architectures).

As a quick reminder, AMD's flagship 7763 server chips come armed with the 64 Zen 3 cores and 128 threads apiece and have a 2.45 GHz base and 3.5 GHz boost frequency. All told, we're looking at a Cinebench run with 128 cores and 256 threads, which you can see in the tweet below:

See more

The dual 7763's scored 113,631 points, while the previous world record weighed in at 105,570 (as per HWBot rankings). AMD says it used a reference server design with conventional air cooling for the test run, so there were no special accommodations or overclocking. The system peaked at 85C and 403W during the test run. Here's AMD's official HWBot world record submission

1K Unit Price / RCPCores / ThreadsBase / Boost - All Core (GHz)L3 Cache (MB)TDP (W)
AMD EPYC Milan 7763$7,89064 / 1282.45 / 3.5256280
Intel Xeon Platinum 8380$8,09940 / 802.3 / 3.2 - 3.060270

That isn't much info to work with, but it's enough for us to set up our own test. We ran a few tests with a dual Xeon 8380 Ice Lake Xeon server we used for our recent review. Much like AMD's test system, this is a standard development design with air cooling (more details in the review). The Xeon system houses two $8,099 10nm Ice Lake Xeons with 40 cores 80 threads apiece that operate at a 2.3 GHz base and 3.2 GHz boost frequency. Yes, AMD's Milan outweighs the Xeon system, but the Ice Lake 8380 is Intel's highest-end part, and both chips come with comparable pricing.

We're looking at the EPCY Milan server with 128 cores and 256 threads against the Intel Ice Lake system with 80 cores and 160 threads. Our quick tests here are not 100% like-for-like so take these with a grain of salt, though we did our best to match AMD's test conditions. Here are our test results, with a few extras from the HWBot benchmark database mixed in:

Cinebench Benchmarks
ScoreCoolingChip Price
2x AMD EPYC Milan 7763113,631Air$15,780
1x Threadripper 3990X (Splave)105,170Liquid Nitrogen (LN2)$3,990
2x EPYC 7H1292,357Air?
2x Intel Xeon Platinum 838074,630Air$17,000
1x Threadripper 3990X (stock)64,354 All-In-One (AIO) Liquid Cooling$3,990

As you can see, in Cinebench R23, the dual EPYC Milan 7763's are 52% faster than the dual Ice Lake Xeon 8380's. AMD lists a 403W peak power consumption during its tests, but we assume those measurements are for the processors only (and perhaps only a single processor). In contrast, our power measurement at the wall for the Xeon 8380 server weighed in at 1154W, but that includes a beastly 512GB of memory, other platform additives, and VRM losses, etc., meaning it's just a rough idea of power consumption that isn't comparable to the EPYC system.

Naturally, Cinebench R23 results have absolutely no bearing on the purchasing decision for a data center customer, but it is an interesting comparison. Notably, a single Threadripper 3990X, when pressed to its fullest with liquid nitrogen by our resident overclocking guru Splave, still beats the two Xeon Platinum 8380's, though the 8380's pull off the win against an air-cooled 3990X at stock settings (measured in our labs).

Finally, we decided to see how two Ice Lake Xeon 8380's compare against a broader set of processors. Intel suffered quite a bit of embarrassment back at AMD's launch of the 64-core Threadripper 3900X for high-end desktop PCs, as this $3,990 processor (yes, just one) beat two of Intel's previous-gen 8280 Xeons in a range of threaded workloads. Intel's Xeon's weighed in at $20,000 total and represented the company's fastest server processors. Ouch.

In fact, those benchmark results were so amazing that we included an entire page of testing in our Threadripper 3990X review comparing two of Intel's fire-breathing behemoths to AMD's single workstation chip, which you can see here. As a bit of a redux, we decided to revisit the standings with a quick run of Cinebench R20 with the new Intel 10nm Xeons. Notably, this test is with an older version of the benchmark than we used above, but that's so we can match our historical data in the chart below:

Xeon vs Threadripper

(Image credit: Tom's Hardware)

Unfortunately, we don't have a dual-socket EPYC Milan 7763 system to add to our historical test results here, but we get a good enough sense of Ice Lake's relative positioning with this chart. The two Intel Ice Lake 8380's, which weigh in at $17,000, beat the single $3,990 Threadripper 3900X at stock settings. That's at least better than the dual 8280's that lost so convincingly in the past.

However, a quick toggle of the PBO switch, which is an automated overclocking feature from AMD that works with standard cooling solutions (no liquid nitrogen required), allows a single Threadripper 3990X to regain the lead over Intel's newest 10nm flagships in this test. Intel's latest chips also can't beat AMD's previous-gen EPYC Rome 7742's, which are 64-core chips.

Of course, this single benchmark has almost no bearing on the enterprise market that the Ice Lake chips are destined for, and the latest Xeon's do make solid steps forward in a broader range of tests that do matter, which you can see in our Ice Lake 8380 review. 

  • thGe17
    So where's the news? The dual-Epyc has +60 % more cores and has a +52 % higher score. This shouldn't be a surprise.
    The only thing to mention is at best the perf/watt rating, which is still better for Epyc because of TSMCs mature N7.
    R23 should require more time due to a higher workload, so I guess turbo's wouldn't play a role here:
    CB R23 MT
    Score/CoreScore/Core/GHzScore/WattDual Epyc 7763888362406Dual Xeon 8380933406276Obviously Intel is doing fine, the problem is still the 10nm node. It's getting better, especially if compared to 14nm+++, but the node still cannot fully compete with TSMCs N7 in this iteration.
    Reply
  • watzupken
    thGe17 said:
    So where's the news? The dual-Epyc has +60 % more cores and has a +52 % higher score. This shouldn't be a surprise.
    The only thing to mention is at best the perf/watt rating, which is still better for Epyc because of TSMCs mature N7.
    R23 should require more time due to a higher workload, so I guess turbo's wouldn't play a role here:
    CB R23 MTScore/CoreScore/Core/GHzScore/WattDual Epyc 7763888362406Dual Xeon 8380933406276Obviously Intel is doing fine, the problem is still the 10nm node. It's getting better, especially if compared to 14nm+++, but the node still cannot fully compete with TSMCs N7 in this iteration.
    I actually don't think any iterations of Intel's 10nm will be competitive in terms of core count until at least 7nm. While SuperFin seems to have improved things with higher clockspeed (also higher power consumption), Intel's current 10nm seems to be struggling with density in my observation because they are still pretty much capped out at 4 cores on their ultra low power CPUs, and around 8 to 10 cores on their desktop CPUs.
    Reply
  • amootpoint
    Who cares about cinebench ? Its a totally useless benchmark. No one makes their purchasing decision on this benchmark.
    Reply
  • thGe17
    @amootpoint: Nobody said that, expect for yourself. ;) CB is just one way to put heavy load on all CPU cores, nothing more, nothing less.
    And Ian Cutress' SPEC2017 tests show silimar results for Score/Core, Score/Core/GHz and Score/Watt. The Ice Lake Cores seem to be faster than Zen3 but here also the N7 shows better efficiency.

    @watzupken: Should be interesting to see if Intel can improve Enhanced SuperFin (10nm+++) as much as for SuperFin, but it also may be possible, that the node will never be able to fully catch up with TSMCs N7(P), but in the end, this may be a minor problem, because Meteor Lake will be already manufactured on 7nm (or maybe (partially?) on TSMCs 5nm), therefore 10nm will only be used for a short period.
    Reply
  • waltc3
    This should at last put paid to the notion that "Intel's 10nm = TSMC's 7nm" but good...;) Clearly that is not the case. If Intel could make a competitive CPU it would, but it can't, so it doesn't. Try as they might, Intel these days mostly makes AMD look really good. Intel performance per watt is poor beside these Epycs. It's going to be an especially difficult road for Intel from now on as AMD is a constantly moving target.
    Reply
  • tripleX
    thGe17 said:
    So where's the news? The dual-Epyc has +60 % more cores and has a +52 % higher score. This shouldn't be a surprise.
    The only thing to mention is at best the perf/watt rating, which is still better for Epyc because of TSMCs mature N7.
    R23 should require more time due to a higher workload, so I guess turbo's wouldn't play a role here:
    CB R23 MTScore/CoreScore/Core/GHzScore/WattDual Epyc 7763888362406Dual Xeon 8380933406276Obviously Intel is doing fine, the problem is still the 10nm node. It's getting better, especially if compared to 14nm+++, but the node still cannot fully compete with TSMCs N7 in this iteration.

    Score/Price?
    Reply
  • TCA_ChinChin
    I don't think anyone will argue Milan's great power efficiency and core count, but unfortunately for AMD, there are some customers that simply don't care about that metric as much anymore. If you've already bought into any of Intel's proprietary or specialized workloads like avx-512, optane, or some machine learning extension, it doesn't matter how efficient your architecture or good it is at general computation, since Intel's not even trying to compete on that level anymore. They've shifted to marketing their fancy extensions and exclusive technologies creating an entire market segment that is completely dependent on Intel products, regardless of how good each new Intel release is. It's similar to how so much of machine learning/AI and compute is dependent on Cuda, and thus must buy Nvidia accelerators and GPUs.
    Reply
  • mac_angel
    Crysis Benchmarks?
    Reply
  • thGe17
    waltc3 said:
    Intel performance per watt is poor beside these Epycs.
    And? AMD was already there in the last decade. ;) So if they have managed to come back, why shouldn't others be allowd to do so?

    tripleX said:
    Score/Price?
    Yes, maybe, but this competition will AMD easily loose against ARM-servers, e.g. an Altra Q80-33 and the M128-30 is going to follow soon. (Obviously they use the same market mechanics like AMD does to gain market shares ;))
    And additionally, you do not know prices for OEMs, because you can be sure that these prices differ from basic stock prices. And the differences already slowly dissipate, if you look at the more relevant general/basic CPUs with lower core counts.
    And to make it even worse for you, price sometimes is only a secondary factor in the datacenter (to be precise, Lisa Su even omitted the "sometimes" in her statement).
    And in many cases, it doesn't matter at all, because AMD simply does not have the capacity. The topic is a bit more complex ...


    @TCA_ChinChin: "I don't think anyone will argue ..." correct.
    "... but unfortunately for AMD, there are some customers that simply don't care ..." nonsense, now the customers are to blame? These are no gaming/hobby products where you simply look at the FPS counter and decide which one is the most suitable CPU. Requirements in datacenters are much more complex (in and outside the cabinet).
    And yes, AMD currently has nothing competitive to offer with regards to AI and that's the reason why AMD is going to enhance the vector units of Zen4. (Btw, they currently also have nothing competitive in GPGPUs with regards to AI; here they will also need more time and hardware iterations or have to compete with a lower price).
    And yes, CUDA is heavily used for AI and other tasks, but not because its useless and Nvidia simply has a good marketing. It is a good and powerful framework and collection of tools and AMD simply has failed to provide resources years ago to put something similar for their own architecture into place. Of course a problem of resources, but certainly nothing to blabe Nvidia for.

    @mac_angel: The CryEngine heavily relies on its primary thread. This is nothing you want to throw onto a slow-clocked, 32+ cores server CPU. And in this case even 200 GB/s memory bandwith won't make a difference. ;)
    Reply
  • Fdchdfa
    thGe17 said:
    So where's the news? The dual-Epyc has +60 % more cores and has a +52 % higher score. This shouldn't be a surprise.
    The only thing to mention is at best the perf/watt rating, which is still better for Epyc because of TSMCs mature N7.
    R23 should require more time due to a higher workload, so I guess turbo's wouldn't play a role here:
    CB R23 MTScore/CoreScore/Core/GHzScore/WattDual Epyc 7763888362406Dual Xeon 8380933406276Obviously Intel is doing fine, the problem is still the 10nm node. It's getting better, especially if compared to 14nm+++, but the node still cannot fully compete with TSMCs N7 in this iteration.
    In case you blind, AMD's Epyc 7763 cost $7890, meanwhile Intel's pathetic 40 core 8380 cost $8099.
    Reply