How to Save $6000 on a 28-core Flagship Intel Xeon: Platinum 8280 vs Gold 6258R
by Dr. Ian Cutress on August 7, 2020 8:00 AM ESTDoes the 6258R Make Sense for Intel?
For this test, I wanted to compare the difference between Intel’s Xeon Platinum 8280 and Intel’s Xeon Gold 6258R. These processors are practically identical on paper for any regular 1P or 2P server, offering up 28 cores at 2.7-4.0 GHz, however the Gold 6258R has a list price that saves over $6000 compared to the Platinum 8280.
As per our regular testing procedure, I put both CPUs in our 1P LGA3647 test system and ran through our performance benchmarks. We also took power measurements, latency measurements, and idle-to-turbo measurements. Everything came out the same. Without the name of the CPU on the heatspreader, or a different CPU string when probed, no-one would be able to tell the difference in a 1P or 2P environment.
So if anyone is thinking of deploying Intel’s high-end Xeon Platinum 8280s in anything less than a eight-socket system, don’t bother. Save a few grand per CPU and gain the plaudits of your boss – unless they start asking questions about why the infrastructure doesn’t have the latest ‘Xeon Plutonium’ things they’re heard about.
Jokes aside, the pragmatic question to ask is:
Has Intel shot itself in the foot with the 6258R?
Intel often repeats (as does AMD) that the majority of its server customers exist in that 1P and 2P spectrum. An offering like the 6258R replaces the 8280 in all aspects for that, giving Intel an effective performance-per-dollar improvement of 2.5x, while at the same time lowering its selling price - when we compare the prices, Intel stands to lose $6000 per processor sold.
However, Intel launched the 8280 in April 2019 as the flagship – the 6258R only came out in February 2020. Anyone who wanted the perfromance of the 8280 in that time frame already purchased one. At the same time, a few months later, the company has launched its 3rd Generation Xeon Scalable platform, known as Cooper Lake. We’ve covered Cooper Lake in detail, but the short information is that it is an OEM platform designed for 4-socket and 8-socket servers. Any customer who needs servers that large are now going to look at Cooper Lake as the leading product, meanwhile the 1-socket and 2-socket customers are still on the Cascade Refresh options.
At this point, the 8280 is a dead product for Intel.
- Users who want the 4-8 socket compatibility and performance can now get the 8380H/HL.
- Users who want the 1-2 socket compatibility and performance will go for the 6258R.
If you’re wondering where the 6258R stacks up against AMD, we’re in the process of re-testing the parts we have on hand as we go through our regression testing. The EPYC 7542 is probably the best comparison point (32C, 2.9-3.4 GHz, 225W, $3400), however we’ll have to look into getting one of those.
Facebook
Twitter
RSS
81 Comments
View All Comments
benedict - Friday, August 7, 2020 - link
The question is not whether Intel is shooting itself in the foot.It is whether anyone buying Intel is shooting himself in the foot.
In the past no one got fired for buying Intel. Maybe it's time to change that.
ZoZo - Friday, August 7, 2020 - link
It's not just about performance or performance/watt, it's also about platform features and robustness. If you buy Intel you're a bit more certain that things will "just work".For example, I bought a 3rd gen Threadripper and have encountered incompatibilities with Linux KVM-based virtual machines, whether it's the FLR reset bug on the USB controllers when passing them through (fixed late June in Linux kernel, 7 months after TR was released), the fact that a Windows guest doesn't yet support nested AMD virtualization (coming in 2H20), or strange performance behaviors that don't happen on Intel (no resolution in sight). These quirks are bound to exist on Epyc too, but I'll admit that virtualization is probably the most tricky thing you can put the platform through. If you're building a server that just takes a bunch of containers without any virtualization gimmicks, it should be fine.
For workstations, the Intel platform supports RDIMMs and therefore much more RAM, unless you buy from the very few OEMs that sell the Threadripper Pro.
Revv233 - Friday, August 7, 2020 - link
You put that very well about the expectations of not having wierd issues.I always kind of figured it was due to chipset more than CPU but I've experienced that from my old Barton & A64's in spades...
To be fair I once had a P4 Northwood on a VIA chipset and I felt that same pain.
When you are talking mission critical stuff. One bad taste from 20 years ago is going to make you hesitate before you give something another try.
eek2121 - Sunday, August 9, 2020 - link
I agree with him. I run AMD all day long, but at work we tried that with Rome, apparently. Something about our application stack didn’t work well with EPYC. I can only assume it was memory latency, however I am not in that department so I don’t know. What I DO known is that AMD has an amazing product, but the platform isn’t there yet. They either need to clamp down on OEMs or release a first party platform to push the OEMs to release better quality stuff.yeeeeman - Friday, August 7, 2020 - link
I can't say this enough times....I see a lot of comments about AMD being now all of a sudden THE CHOICE.
Hold on for a bit, since there is much more to it than performance or efficiency.
Performance wise, sure, AMD has an advantage, but it is not out of the ordinary. Efficiency wise, the same.
Quality wise, there is no comparison between Intel and AMD platforms. AMD has a lot of work to do on the software and even hardware/firmware front. Many people are not aware of this maybe, but Intel during all these years has invested a lot of effort and money in creating streamlined platforms, with quality software, quality firmware, that it is almost plug and play.
AMD on the other hand, being the underdog and so far away of Intel for so long, nobody focused on their products and they lack badly in optimization and compatibility. Sure, things will change with time, but it will take at least 5 more years of actual work and $$$ from AMD to make it happen.
Same is valid for notebook space. Everyone is crying about how AMD laptops are crippled and how OEMs love Intel. Well, Intel has invested a lot of money and effort in creating design templates and making it as easy as possible for OEMs to create a new laptop.
AMD....well, they just have a good cpu and that is it. There is a lack of field engineers, a lack of streamlined process, a lack of clear BOM and especially a perception from the market that still sees AMD as the cheaper option. Non hardware guys, which is basically 90% of the market cannot deduct from a sticker that 4th gen Ryzen is much better than Buldozer or whatever. AMD needs to invest $$$ into publicity, into OEM partnerships, into creating something similar to project Athena, something to give them the premium feel so that the market perception will change. Otherwise, it will take years before they actually reach majority of market share and by that time Intel could come back.
Anyway, back to the topic of this review, 10k $ is in any case a ridiculous amount of money for an old CPU. 3k $ is even stretching it, so you say it right. If someone wants a top Intel CPU, they should buy this Gold version.
duploxxx - Saturday, August 8, 2020 - link
I can't say this enough. You talk bias. You don't need field engineers. Bom are delivered by oem like Dell hpe Cisco that increase there portfolio for amd on every release. Intel is not investing a lot of money in business for stability all they do is making sure business dies not get fed up with all the CVE bugs they need to patch and deliver in fact the long lasting issues with supplies had them against the wall big time. All intel does is paying money to oem in r&d to keep designing base lines so that they can keep selling the masses. This is done in both WS as Server area. But let's be honest AMD would never be capable to deliver far more. But it's thx to AMD that things like Skylake R exist don't forget that.Smell This - Saturday, August 8, 2020 - link
Now with new glue and Omni-Path v4.0.89.42.000
(Two Dies Are Twice As Nice As One!)
The Cascade Lake Xeon Scalable platform failed last year, and the refresh will fail, again, today.
ProDigit - Saturday, August 8, 2020 - link
Who cares about optimization, when their processors are 25% more efficient, and host between 25-150% more cores?DominionSeraph - Monday, August 10, 2020 - link
This.When the Ryzen 2700 dropped to $150 I jumped on it, figuring I could retire my i7 4790. I put it together and its encoding speed was a very impressive improvement. I couldn't swap out my main immediately as the mobo didn't have enough SATA ports for all my drives so I was using them side-by-side for a while, and I noticed the AMD just lagged and had weird quirks where the 4790 was perfection 24/7/365. I couldn't live with it and gave up on the idea of using it as a replacement, and I eventually sold it off.
I'm still on the i7 4790 as my main even though I now have a 3950X too. AMD is ok for a secondary crunching machine but it's just not suitable for human use.
WizardMerlin - Wednesday, August 26, 2020 - link
Having been using my 3900X for many months now for 16 hours a day for a varied workload, I'd disagree.