Become a fan of Slashdot on Facebook

 


Forgot your password?
Close
typodupeerror
×
AMD Intel

AMD Poses 'Major Challenge' to Intel's Server Leadership (eweek.com) 75

Posted by EditorDavid from the chipping-away dept.
Rob Enderle reports on the excitement at AMD's Epyc processor launch in San Francisco: I've been at a lot of AMD events, and up until this one, the general message was that AMD was almost as good as Intel but not as expensive. This year it is very different; Intel has stumbled badly, and AMD is moving to take the leadership role in the data center, so its message isn't that it is nearly as good but cheaper anymore; it is that it has better customer focus, better security and better performance. Intel's slip really was around trust, and as Intel seemed to abandon the processor segment, OEMs and customers lost faith, and AMD is capitalizing on that slip...

AMD has always been relatively conservative, but Lisa Su, AMD's CEO, stated that the company has broken 80 performance records and that this new processor is the highest-performing one in the segment. This is one thing Lisa's IBM training helps validate; I went through that training myself and, at IBM, you aren't allowed to make false claims. AMD isn't making a false claim here. The new Epyc 2 is 64 cores and 128 threads and with PCIe generation 4, it has 128 lanes on top its 7nm technology, which currently also appears to lead the market. Over the years the average performance for the data center chips, according to Su, has improved around 15% per year. The last generation of Epyc exceeded this when it launched, but just slightly. This new generation blows the curve out; instead of 15% year-over-year improvement, it is closer to 100%...

Intel has had a number of dire security problems that it didn't disclose in timely fashion, making their largest customers very nervous. AMD is going after this vulnerability aggressively and pointing to how they've uniquely hardened Epyc 2 so that customers that use it have few, if any, of the concerns they've had surrounding Intel parts. Part of this is jumping to more than 500 unique encryption keys tied to the platform.
Besides Google and Twitter, AMD's event also included announcements from Hewlett-Packard Enterprise, Dell, Cray, Lenovo, and Microsoft Azure. For example, Hewlett Packard Enterprise has three systems immediately available with AMD's new processor, the article reports, with plan to have 9 more within the next 12 months. And their CTO told the audience that their new systems have already broken 37 world performance records, and "attested to the fact that some of the most powerful supercomputers coming to market will use this processor, because it is higher performing," calling them the most secure in the industry and the highest-performing.

"AMD came to play in San Francisco this week," Enderle writes. "I've never seen it go after Intel this aggressively and, to be frank, this would have failed had it not been for the massive third-party advocacy behind Epyc 2. I've been in this business since the mid-'80s, and I've never seen this level of advocacy for a new processor ever before. And it was critical that AMD set this new bar; I guess this was an extra record they set, but AMD can legitimately argue that it is the new market leader, at least in terms of both raw and price performance, in the HPC in the server segment.

"I think this also showcases how badly Intel is bleeding support after abandoning the IDF (Intel Developer Forum) conference."
This discussion has been archived. No new comments can be posted.

AMD Poses 'Major Challenge' to Intel's Server Leadership More

AMD Poses 'Major Challenge' to Intel's Server Leadership

Comments Filter:

  • How DARE they! (Score:4, Insightful)

    by dpilot ( 134227 ) on Saturday August 10, 2019 @02:38PM (#59074428) Homepage Journal

    After all, isn't server leadership Intel's by right?

    More seriously, the market is better served when there is serious competition. Intel has a habit of going a bit off the rails whenever it doesn't have enough competition, finding ways to improve it's profit margins at its customers' expense.

    • Re: (Score:2)

      by Kokuyo ( 549451 )

      Couldn't agree more.

      Now let's just hope AMD is able to catch up to Nvidia in the top performance segment as well.

  • Anyone who has been following tech reporting for any length of time knows that right there is the place to stop reading the article. Enderle is a paid shill. Over the year's he has correctly predicted 34 of the 0 instances of Apple going out of business [macobserver.com]...

    • Re: (Score:2)

      by rnturn ( 11092 )
      Ah, yes. Bob Enderle. The reporter with the never-ending grudge against open source software going back about twenty years ago. Has he softened his tone about OSS? Or was it just not something he needed to expound upon since this was news about hardware?

    • Paid shill? Who is paying him? AMD is getting plenty of free press from reviews of Rome:

      https://www.phoronix.com/scan.... [phoronix.com]

      https://www.phoronix.com/scan.... [phoronix.com]

      Unless you trust Francois Piednol of course. Then AMD is in serious trouble! Or something.

      • How well does Rome do on database loads?

    • Yeah, I'll never forget his shameless shilling for SCO/Microsoft back in the Groklaw days. Man it makes me feel old realizing that was 15 years ago now. There were countless episodes of his shilling and dishonesty, but without spending more time than I'd like here are a few links that were easy to find. Here [groklaw.net], here [linux.com], and here [lwn.net].

    • Re: (Score:3)

      by Guspaz ( 556486 )

      Maybe, but his general tone is right. AMD's EPYC chips beat Intel's Xeon lineup pretty much across the entire pricing spectrum in pretty much every way, as long as you don't need AVX512 support, in which case they merely match them.

      Intel's highest-end Xeon is a $13,000 28-core part, and it's going up against a $7,000 AMD EPYC chip that has 64 cores and more than double the performance.

      It's a combination of AMD executing really well with Zen 2 and 7nm and Intel being in the middle of a big stumble with both

      • Re: (Score:2)

        by Agripa ( 139780 )

        In the short term, this will force Intel to slash their prices by enormous margins and try to rush their next-gen parts to market. Their next-gen parts which will only hit 56 cores and still not offer as many PCIe lanes or as simple of a NUMA architecture...

        The last time this happened, Intel did not have to slash prices. They used a completely different method to handle AMD.

        • Re: (Score:2)

          by Guspaz ( 556486 )

          That same trick won't work again after Intel got caught doing it, and the new Epyc chips have seen widespread and enthusiastic support from the enterprise hardware manufacturers so far.

          If you're talking about the Athlon 64 days, they got ahead of Intel, but not by this much, not with four times the performance-per-dollar for the closest to feature-equivalent parts as you can get on the high-end.

  • But most people aren't going to be building their own servers - so the announcements from HP, Dell, et. al. are welcome news. We're a university department, so the math is a bit different... because we're not directly paying bandwidth or electricity costs, running our own servers (in most circumstances, anyway) still makes financial sense. It's not as if our overhead charges would change if we suddenly moved all our servers to AWS.

    I have my doubts that most people who aren't hobbyists actually care much whe

  • AMD has done a good job of narrowing the gap; however, the fact that AMD can produce sufficient quantity of chips would have threatened Intel. After all, if companies like Dell can’t sell their latest Intel servers because they don’t have enough Intel chips, they are going to look for alternatives.

    • OK, now look up how AMD gets their chips made, do their still have their own fabs, or not? And what is the capacity?

      That's how awful your comment is. It is something you remembered from decades past, not something that you're following and know something about.

      Just go look it up before commenting.

      • Ok now, look at how Apple or Qualcomm’s chips are made. Do they own their own fabs? Do you see how awful your comment is? It is a fact that in the last few years Dell and HP are deeply unhappy with Intel over yields of the latest chips. Seriously is this some sort of egotism that Intel can claim they own their fabs but have been so terrible at 10nm yields that their customers have to turn to alternate suppliers. This is not in decades past. I would encourage you to look this up before commenting.

    • TSMC fabs the chips. There is no capacity constraint.

      And AMD did more than "narrow the gap".

      https://www.anandtech.com/show... [anandtech.com]

      "For those with little time: at the high end with socketed x86 CPUs, AMD offers you up to 50 to 100% higher performance while offering a 40% lower price. Unless you go for the low end server CPUs, there is no contest: AMD offers much better performance for a much lower price than Intel, with more memory channels and over 2x the number of PCIe lanes. These are also PCIe 4.0 lanes. Wha

  • Other industry's (Often with more players.) are static by comparison. This is a great example of real competition, here we can see the benefits of not letting a single company buy, dominate or force an under the table deal with competition in order to stop real competition and progress. This should be a usefully model to help draw people intelligent conclusions in our present reality of business giants (who are often clearly monopolies) dominating world economics.

  • Intel has seemingly reached a performance standstill, allowing AMD to catch up and surpass them. So what went wrong at Intel? Could it be that AMD has been focused on making chips while Intel has been focused on social engineering? Here's an Intel video [youtube.com] showing their work int he field of "diversity." The video in no way explains how recruiting people based on race and sex leads to better products, yet "diversity" has become the primary focus at Intel.

    Intel said they want to [intel.com] "bring the number of female,

    • In principle nothing wrong with all that diversity.

      In the meanwhile AMD is being run by a woman of Asian origin.

      • There is a lot wrong with "diversity" in principle.

        Mainly because it emphasizes anything else but results, education, skills, talent and team coherence when evaluating workplace performance. If that comes as a surprise it means we misinterpreted our wish to make things equal with the fact of life that people are not born equal, that talents and behavioral preferences are learnt as well as inherited and not all groups and their traditions are able to produce equal results under similar conditions, and that t

      • Diversity is good in situations where having a different background gives you unique insights into the topic of discussion. It's great in social policy discussion and in the arts and I'm sure a great many other things. It doesn't really provide any meaningful benefit to designing computer chips and targeting it as a hiring metric doesn't make sense.

        Lack of diversity in professionals designing computer chips could indicate unintended social, educational, or economic barriers making entry into the discipli
    • at least not based on the benchmarks I'm seeing. It's not the crazy boosts we used to see, but my i5 7500 is about 66% slower than a 9400 in multi-core and about 12% slower in single core. Especially with the multi-core that's nothing to sneeze at.

      As for diversity, they just want lower wages and a few more of the top end employees that are rare as hens teeth. Right now women are discouraged from going into tech. They're much more likely to go into medicine. Intel would love to change that since it would
      • It's not a significantly lower number though; as I recall, it's under 5%. The 79 cents on the dollar figure is apples and oranges bullshit; women aren't being paid vastly last for the same job.

        And even for a couple percentage "discount", it's not like the average "value" will be exactly the same. Women get paid maternity; men do not. It's not sexist to point out that this does affect the supposed financial incentive to hire women you're talking about. On average, men are more likely to be open for workin

      • Intel can't grow past Skylake-derived cores on 14nm++. That is the performance standstill. 10nm has failed them, and 7nm won't show up for anything besides Intel Xe (2021) before 2022. Sunny Cove, Willow Cove, and Golden Cove are all sitting by the sidelines waiting for process improvements so that Intel can actually implement them. The only Sunny Cove cores you'll see this year will be in very limited releases of Icelake in the mobile segment, featuring low core counts and curiously low clockspeeds. S

    • I have worked for 3 major name big tech companies from Silicon Valley (you know all 3). All 3 promote diversity. However, it's not a distraction to their business model. It's a faction they like to parade to make outsiders like us more. The core leadership is not distracted by a desire to ensure there are enough trans latinas in their ranks. A specific faction wants a "woke 2.0" diversity quota and every one else pays them enough lip service to shut them up. We don't change our roadmaps to accommodate

      • Re: (Score:1)

        by Khyber ( 864651 )

        "Either you're a professional troll or the economic equivalent of an incel."

        What is the economic equivalent of an involuntary celibate person, someone who has never been able to spend a penny in their lives?

        Your name-calling needs work. Your brain doubly-so.

      • I believe the internal rot you're looking for has a name. Brian Krzanich.

    • Comment removed based on user account deletion

  • Of course, there are a lot of idiots.

    • Re: (Score:1)

      by Zuriel ( 1760072 )

      There are a lot of places with existing Intel hardware, and for many of them it makes sense to continue adding Intel hardware rather than introduce AMD hardware and then be forced to deal with compatibility issues.

      It's not just stupidity or stubbornness, there are situations where continuing to buy the inferior kit is the sensible decision.

  • ''AMD was almost as good as Intel but not as expensive.''

    This has been the story for AMD since the release of the 386DX40. I've been rooting for them since 1991. They've underperformed as a company [and a security] since then.

    Major challenge? If historical performance was able to gauge future performance, I'd say not. Being almost as good, isn't something the marketing department should be selling. Which might
    be the issue in itself.

    • Re: (Score:3)

      by Guspaz ( 556486 )

      Right, and that's what changed with this product launch. It's much better than Intel and still not as expensive. AMD hit one out of the park at the same time as Intel broke an ankle.

      • Re: (Score:3)

        by dpilot ( 134227 )

        AMD hit one out of the park with K8, at the same time that Intel was making mistakes with NetBurst and IA64. However Intel arm-twisted it's customers to minimize the K8 success.

        • Intel isn't going to twist any arms while at the same time reducing its interest in the CPU market.

    • The only parts Intel had that could compete were literally 4x the price. AMD's only fault was they bet heavy on multi core and lost. Turned out multi-core was too hard to program for and there wasn't much demand for it. Especially on games since everything had to run on an XBox 360 and the 360's processor had gobs and gobs of single core performance, more than it's GPU could ever use.

      With the PS4 and XBone being AMD hardware (and underpowered AMD hardware at that...) developers had to step up their game
    • Nonsense. They kicked Intel's ass all over the place for something approaching 5 years in the early 2000s. (Netburst was inferior even though it had higher mhz.) They did this and immediately followed up with AMD-64 and forced Intel to copy it, instead of pursuing their non-backwards compatible Itanium strategy.

      Major challenge? If historical performance was able to gauge future performance, I'd say not. Being almost as good, isn't something the marketing department should be selling. Which might be the issue in itself.

      I mused on this recently: a big problem may have been they continued to price their processors cheaper than Intel's even though they were apparently faster. They should've priced them higher (while l

      • Re: (Score:2)

        by Zuriel ( 1760072 )
        The Itanium strategy was probably always doomed. When IA64 was introduced, everybody's existing software was 32-bit. A 64-bit chip that was awful at 32-bit was always going to struggle. Itanium needed the entire software world to cooperate and release IA64 builds, including obsolete or unsupported software and companies that had shut down.

    • This has been the story for AMD since the release of the 386DX40.

      This is simply not true.

      I don't recall the Opteron 24x from August 2003 running 64-bit code in a dual-socket configuration having any serious rival from Intel. Floor, this is mop; mop this is floor. Intel desperately tried to rectify this situation before AMD's decisive victory leaked into the consumer space a few months later.

      AMD's Opteron 148 and 248 processors [techreport.com] — 17 November 2003

      If you have been following these things at all, you proba

  • Truly! ...but sweet Slashvertisement, bro.

Slashdot Top Deals

If all else fails, lower your standards.

Close