Intel 3rd Gen Xeon Scalable (Ice Lake SP) Review: Generationally Big, Competitively Small
by Andrei Frumusanu on April 6, 2021 11:00 AM EST- Posted in
- Servers
- CPUs
- Intel
- Xeon
- Enterprise
- Xeon Scalable
- Ice Lake-SP
SPEC - Single-Threaded Performance
Single-thread performance of server CPUs usually isn’t the most important metric for most scale-out workloads, but there are use-cases such as EDA tools which are pretty much single-thread performance bound.
Power envelopes here usually don’t matter, and what is actually the performance factor that comes at play here is simply the boost clocks of the CPUs as well as the IPC improvement, and memory latency of the cores.
The one hiccup for the Xeon 8380 this generation is the fact that although there’s plenty of IPC gains to be had compared to previous microarchitectures, the new SKU is only boosting up to 3.4GHz, whereas the 8280 was able to boost up to 4GHz, which is a 15% deficit.
Even with the clock frequency disadvantage, thanks to the IPC gains, much improved memory bandwidth, as well as the much larger L3 cache, the new Ice Lake part to most of the time beat the Cascade Lake part, with only a couple of compute-bound core workloads where it falls behind.
The floating-point figures are more favourable to the ICX architecture due to the stronger memory performance.
Overall, the new Xeon 8380 at least manages to post slight single-threaded performance increases this generation, with larger gains in memory-bound workloads. The 8380 is essentially on par with AMD’s 7763, and loses out to the higher frequency optimised parts.
Intel has a few SKUs which offers slightly higher ST boost clocks of up to 3.7GHz – 300Mhz / 8.8% higher than the 8380, however that part is only 8-core and features only 18MB of cache. Other SKUS offer 3.5-3.6GHz boosts, but again less cache. So while the ST figures here could improve a bit on those parts, it’s unlikely to be significant.
169 Comments
View All Comments
Oxford Guy - Tuesday, April 6, 2021 - link
Reading the conclusion I’m confused by how it’s possible for the product to be a success and for it to be both slower and more expensive.‘But Intel has 10nm in a place where it is economically viable’
Is that the full-fat 10nm or a simplified/weaker version? I can’t remember but vaguely recall something about Intel having had to back away from some of the tech improvements it had said would be in its 10nm.
Yojimbo - Tuesday, April 6, 2021 - link
Because there is more than the benchmarks that are in this review to making decisions when buying servers. Intel's entire ecosystem is an advantage much bigger than AMD's lead in benchmarks, as is Intel's ability to deliver high volume. The product will be a success because it will sell a lot of hardware. It will, however, allow a certain amount of market share to be lost to AMD, but less thanwpuld be lost without it. It will also cut into profit margins compared to if the Intel chips were even with the AMD ones in the benchmarks, or if Intel's 10 nm was as cost effective as they'd like it to be (but TSMC's 7 nm is not as cost effect as Intel would like they're processes to be, either).RanFodar - Tuesday, April 6, 2021 - link
This.Oxford Guy - Wednesday, April 7, 2021 - link
So, the argument here is that the article should have been all that instead of focusing on benchmarks.Yojimbo - Friday, April 9, 2021 - link
I never made any argument or made any suggestions for the article, I only tried to clear up your confusion: "Reading the conclusion I’m confused by how it’s possible for the product to be a success and for it to be both slower and more expensive." Perhaps the author should have been more explicit as to why he made his conclusion. To me, the publishing of server processor benchmarks on a hardware enthusiast site like this is mostly for entertainment purposes, although it might influence some retail investors. They are just trying to pit the processor itself against its competitors. "How does Intel's server chip stack up against AMD's server chip?" It's like watching the ball game at the bar.mode_13h - Saturday, April 10, 2021 - link
> To me, the publishing of server processor benchmarks on a hardware enthusiast site like this is mostly for entertainment purposes, although it might influence some retail investors.You might be surprised. I'm a software developer at a hardware company and we use benchmarks on sites like this to give us a sense of the hardware landscape. Of course, we do our own, internal testing, with our own software, before making any final decisions.
I'd guess that you'll find systems admins of SMEs that still use on-prem server hardware are probably also looking at reviews like these.
Oxford Guy - Sunday, April 11, 2021 - link
It's impossible to post a rebuttal (i.e. 'clear up your confusion') without making one or more arguments.I rebutted your rebuttal.
You argued against the benchmarks being seen as important. I pointed out that that means the article shouldn't have been pages of benchmarks. You had nothing.
trivik12 - Tuesday, April 6, 2021 - link
I wish there were test done with 2nd gen Optane memory. isn't that one of the selling point of Intel Xeon that is not there in Epyc or Arm Servers. Also please do benchmarks with P5800X optane SSD as that is supposedly fastest SSD around.Frank_M - Thursday, April 8, 2021 - link
Optane and Thunderbolt.Azix - Tuesday, April 6, 2021 - link
There's a reason intel's data center revenues are still massive compared to AMDs. These will sell in large quantities because AMD can't supply.