Slashdot Mirror


Intel Launches Xeon Scalable CPUs: Dual Xeon Platinum 8176, 112 Threads Tested (hothardware.com)

MojoKid writes: Intel announced its new Xeon Scalable processor family based on the 14nm Skylake-SP microarchitecture a few weeks back, though today marks the official launch of the platform. Not only do these processors feature a new microarchitecture, but Intel has also revamped the naming convention and arrangement of the Xeon product stack, branding them with Platinum, Gold, Silver, and Bronze model families. Intel Xeon Scalable series processors feature core counts ranging from 4 to 28, with varied frequencies and cache configurations. Workstation processors and lower-core count server chips top out in the 3.2GHz -- 3.6GHz range, while the higher-core count products typically fall in the 2GHz -- 2.7GHz range. Six memory channels are supported and the chips have 48 lanes of integrated PCIe 3.0 connectivity. Power envelopes range all the way from 70W on up to 205W. The Xeon Scalable series also introduces new security, virtualization, and storage-related features, more memory bandwidth, support for AVX-512 extensions, a mesh interconnect, and enhanced hardware controlled power management, among a host of other architectural improvements. Testing of a 2P Xeon Platinum 8176 system, sporting 56 physical cores / 112 threads shows significantly increased performance and bandwidth, with only moderately higher power consumption versus a previous-gen 2P Xeon E5-2679 v4-based system.

54 comments

  1. finally, Xeon by turkeydance · · Score: 1

    8176+112=8288. nerd stuff.

    1. Re:finally, Xeon by U8MyData · · Score: 1

      Be really funny if it added up to 8088.

  2. Imagine a Beowulf cluster... by Red+Herring · · Score: 5, Funny

    Too old a joke? Anyone? Anyone? This thing on?

    --
    #include "standard_disclaimer.h"
    1. Re:Imagine a Beowulf cluster... by sl3xd · · Score: 1, Troll

      Beowulf is an ancient story, whether we're talking about literature or technology.

      It was too old in 2005.

      At least in 2005, AMD had a processor that was genuinely faster than anything from Intel. Those were heady days.

      Now we just get fanbois droning on endlessly about how awesome it is that AMD's next generation will be slower but cheaper than Intel's latest offering.

      --
      -- Sometimes you have to turn the lights off in order to see.
    2. Re:Imagine a Beowulf cluster... by Anonymous Coward · · Score: 0

      not too old - just not funny like all the "jokes" on /.

    3. Re:Imagine a Beowulf cluster... by Anonymous Coward · · Score: 0

      It's not a matter of 'slower but cheaper' on a linear scale, that's the point you're missing. It doesn't take a fanboy to point that out nicely either.

      You'd rather throw your investment into a company that is renown for repeatedly gouging as-able rather than a company actually innovating substantially.

      I suppose I understand, you're just lazy intellectually and perhaps spiritually. If it's slightly faster that's 100% of your criteria, and nothing else matters.

      You're a techno-nihilist, no need to apologize for your narrow values.

    4. Re:Imagine a Beowulf cluster... by alvinrod · · Score: 1

      Now we just get fanbois droning on endlessly about how awesome it is that AMD's next generation will be slower but cheaper than Intel's latest offering.

      Actually, AMDs newest chip is pretty good. Earlier today, Anandtechan article comparing Intel and AMDs latest offerings, and puts the 8176 up against AMDs best CPU. The overall takeaway is that the AMD system is faster in a lot of workloads (databases being the one big notable area where it loses badly) than the Intel system and at a much lower cost. It isn't a complete ass-whooping to the degree it was in the Opteron days, but now AMD actually has a product that won't be relegated to the bargain bin.

    5. Re:Imagine a Beowulf cluster... by Anonymous Coward · · Score: 0

      Yes, John Connor, it is on.

    6. Re:Imagine a Beowulf cluster... by Anonymous Coward · · Score: 0

      Michael Jackson has returned as an alien and has given AMD technology from the future to battle against intel now... and thats an ALTERNATIVE fact!

    7. Re:Imagine a Beowulf cluster... by Anonymous Coward · · Score: 0

      Yes! And with Sega Dreamcasts too, 300 of 'em.
          Now to render that 3D scene it'll only take two days!

  3. Hilarious by Anonymous Coward · · Score: 5, Informative

    Very expensive (the top one is $11722) and very fragmented portfolio. There actually is a review comparing one of the top Xeon Platinums 8176 ($8719) to the AMD EPYC 7601 (4200$) - http://www.anandtech.com/show/11544/intel-skylake-ep-vs-amd-epyc-7000-cpu-battle-of-the-decade

    The results might be surprising and speak of desperation in pushing the Core-derived architectures too far. Another indicator is the recent HEDT platform's problems on Tom's Hardware - http://www.tomshardware.com/reviews/-intel-skylake-x-overclocking-thermal-issues,5117.html

    Not to mention another security nightmare on the level of Intel ME/AMT:

    "The chipset will also include a new feature called Intel’s Innovation Engine, giving a small embedded core into the PCH which mirrors Intel’s Management Engine but is designed for system-builders and integrators. This allows specialist firmware to manage some of the capabilities of the system on top of Intel’s ME, and is essentially an Intel Quark x86 core with 1.4MB SRAM."

    Because we all know the motherboard manufacturers are known for stable and secure code, so let's let them botch yet another thing.

    It's high time AMD kicked Intel's butt into actually trying and not milking their customers constantly.

    1. Re:Hilarious by gweihir · · Score: 3, Interesting

      I think this is a sign of pretty extreme desperation: They do not have anything on par with what AMD offers and they will not have anything for years to come, as developing new architectures takes a lot of time, regardless of how much money you throw at it. AMD, meanwhile can optimize their new design for the next 5 years or more before they are even remotely threatened by Intel. Will take all the mindless sheep a while to understand, but eventually it might even dawn on them how thoroughly Intel has fucked them over the last few years..

      --
      Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
    2. Re:Hilarious by Anonymous Coward · · Score: 0

      Platinum, Gold, Silver, and Bronze?

      I will wait for the Lead model.

    3. Re:Hilarious by kfh227 · · Score: 1

      And your source from Intel's R&D department saying they having started years ago is?

    4. Re:Hilarious by thegarbz · · Score: 1

      Not to mention another security nightmare on the level of Intel ME/AMT:

      Yeah sorry but no one cares about this in this given market. Pretty much every Xeon chip ever sold has been put into a motherboard with some equivalent of this as a value added extra, and at a premium price too.

      If you're in the market for these and you're trying to control access at this level then you're doing your security wrong.

    5. Re:Hilarious by Anonymous Coward · · Score: 0

      Wow. You apparently do NOT work in Enterprise computer OR HPC.

      Enterprise:
      -speed of code
      -$

      HPC (seriously, no one has put all their eggs into accelerators outside 2 sites WORLD WIDE)
      -memory bandwidth
      -FP
      -TCO so energy/$

      EPYC actually looks great, but only compared to Skylake SP. We are hearing that Skylake SP stinks compared to existing Broadwell, so time will tell.

    6. Re:Hilarious by tlhIngan · · Score: 1

      I think this is a sign of pretty extreme desperation: They do not have anything on par with what AMD offers and they will not have anything for years to come, as developing new architectures takes a lot of time, regardless of how much money you throw at it. AMD, meanwhile can optimize their new design for the next 5 years or more before they are even remotely threatened by Intel. Will take all the mindless sheep a while to understand, but eventually it might even dawn on them how thoroughly Intel has fucked them over the last few years..

      Here's the thing - AMD cannot die. Intel does NOT want AMD to die. Because if AMD dies, chipzilla will become the focus of a LOT of government scrutiny, and a lot of sweetheart deals (like patent cross-licensing between AMD and Intel) goes away. Hell, AMD's patents may not even be available for purchase by Intel. - instead sold to competitors like Apple, Qualcomm, ARM, etc. who may or may not want to enter in a cross-licensing deal with Intel. Follow that with probably tons of government regulation on business practices and it's not a place where Intel wants to be. Hell, Intel might be forced to break up under government rule.

      What Intel wants is AMD to be somewhat financially stable - AMD has a good steady income source right now which they deployed for Ryzen (courtesy Microsoft and Sony). Heck, we don't know, but it's probably likely Intel told Sony and Microsoft to not bother and give the money to AMD instead.

      And AMD is right where Intel wants them to be - a tiny competitor that's small enough to be of little mind to Intel, but big enough to appear as a worthy competitor so Intel will avoid government scrutiny. Intel's not threatened - they're wanting people to go AMD.

      Face it - Intel's not expecting to sell many of these chips - they're expecting people to go AMD instead. And if that becomes a problem, they can always drop the price a bit. It's not a panic response. It's just to have something "to compete" and again, make the market appear to be solid.

      Intel's not threatened by this (they don't call them chipzilla for nothing), they're not panicking, they're just offering enough so AMD can be self-sustaining and off Intel's life support.

      Compare how Intel's mis-steps don't hurt it at all (remember, AMD being better is not a new thing - the Pentium 4 was outclassed by AMD"s offerings which were faster and cheaper on the whole). But AMD"s mis-steps with Bulldozer nearly killed them.

    7. Re:Hilarious by gweihir · · Score: 1

      Do you have a reading comprehension issue?

      --
      Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
  4. Scalable price tag. by Anonymous Coward · · Score: 0

    Will it have *scalable* prices as well? Like they tried last time where if you wanted an additional feature turned on you had to pay more for it?

  5. I also wonder what other by Anonymous Coward · · Score: 0

    Backdoors they baked into the dies this time.

  6. Obsolete computer architecture by Script+Cat · · Score: 1

    Threads, system buses
    It's time for networked processors (not ethernet)
    You give each processor its own memory space with a large amount of the space dedicated to its own private ram. Other sections of this memory space can intersect the memory space of a set of neighboring processors, shared ram. Get rid of the cashing and swapping and thrashing. Extremely complex operations can pipeline from processor space to processor space. One could make Neural networks, Function networks, Data flow programs as they wish. Have thousands of these processors. Note this is not the same as the single instruction multiple data (SIMD) GPUs being hoisted on us right now.

    1. Re: Obsolete computer architecture by Anonymous Coward · · Score: 0

      And what is the magical interconnect that you are going to use? Infiniband? Fibre Ring?

    2. Re:Obsolete computer architecture by Anonymous Coward · · Score: 0

      >cashing

    3. Re:Obsolete computer architecture by Anonymous Coward · · Score: 0

      In other words, The Machine that HP Labs is working on.

    4. Re:Obsolete computer architecture by __1200333 · · Score: 1

      So basically the Inmos Transputer from the 1980s?

    5. Re:Obsolete computer architecture by Anonymous Coward · · Score: 0

      This idea was invented by Shampoo.

    6. Re:Obsolete computer architecture by Rockoon · · Score: 1

      Shared memory is superior in general because the code can simply avoid sharing to avoid nearly all of the negatives. Its called Non-Uniform Memory Access (NUMA) .. maintaining cache coherency with simple algorithms while giving each thread its own pool of fast memory.

      The advantage is that when you do need to share, its not nearly as bad as trying to push data across a cluster.

      --
      "His name was James Damore."
    7. Re:Obsolete computer architecture by gweihir · · Score: 1

      That bullshit again. Yes, again, it was up around 20 years ago, complete failure, and it might have been up before that, also complete failure if it was.

      --
      Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
    8. Re: Obsolete computer architecture by Rockoon · · Score: 2

      Its powered by Wishful Thinking. The fastest data lines ever devised.

      --
      "His name was James Damore."
    9. Re: Obsolete computer architecture by Script+Cat · · Score: 1

      It's called metal lines on the die. You put either Dual port ram between the units or interleaved ram. Many identical units within the same chip. Then put pin to pin connections to neighboring chips of the same type if you need more.

        The main point is to leave the data with its processor and quit shuffling data pages up and down different levels of cache.

    10. Re:Obsolete computer architecture by bigtreeman · · Score: 1

      WTF? Up to 205W, man that would glow without liquid cooling.
      Fully agree they're stuck on a track of blindly making it bigger.
      networked processors
      http://www.greenarraychips.com...
      144 separate processors in a chip, all running, well under 1W
      Chuck Moore's been doing it for years, nothing new, and it doesn't glow in the dark.

      --
      Go well
    11. Re:Obsolete computer architecture by Anonymous Coward · · Score: 0

      Like a microbee gamma (2 z80s and a 68010) on steroids?

    12. Re:Obsolete computer architecture by Anonymous Coward · · Score: 0

      Not sure if you have ever stepped into a server room, but they dont care about quiet in the rack. They practically shove 4 80mm tornadoes into a 2u unit. cooling wont be an issues.

    13. Re:Obsolete computer architecture by Script+Cat · · Score: 1

      Looks cool. Are these available? It seems the nodes are separated by comports and not shared ram buffers. I was hoping for in addition to local ram for each processor core, a few shared memory ranged with a group of other processors. Maybe in a bipartite graph configuration.

    14. Re:Obsolete computer architecture by Script+Cat · · Score: 1

      Shampoo? Like all sudsy?

    15. Re:Obsolete computer architecture by Script+Cat · · Score: 1

      How did it fail?

    16. Re:Obsolete computer architecture by Script+Cat · · Score: 1

      Serial links

  7. was there an earthquake in silicon valley? by Anonymous Coward · · Score: 0

    i swear there was, because intel is shaking in their boots in their mad rush to beat amd to the punch with the next versions of processors.

    1. Re:was there an earthquake in silicon valley? by gweihir · · Score: 1

      It is pretty obvious they know they cannot. So they are not trying to fake it.

      --
      Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
  8. Sounds great for supercomputers... by Gravis+Zero · · Score: 1

    and really shitty for everyone else who runs servers on a reasonable budget. It still appears AMD is going to take the server market.

    --
    Anons need not reply. Questions end with a question mark.
  9. AS/400 by ArchieBunker · · Score: 1

    Read up on the AS/400 architecture. Makes much more sense than the 8088 kludge we've had for decades.

    --
    Only the State obtains its revenue by coercion. - Murray Rothbard
    1. Re:AS/400 by Anonymous Coward · · Score: 0

      I would say look at the POWER architecture since AS400 got rebranded and melded with its brother RS6000. Power9 servers are looking great this year.

  10. how do the amd EPYC systems hold up? by Joe_Dragon · · Score: 1

    how do the amd EPYC systems hold up?

    With Intel you need 3 cpus to get the same pci-e as 1 amd.

    Also that 1 AMD cpu has more ram channels then 1 intel cpu.

    1. Re:how do the amd EPYC systems hold up? by Anonymous Coward · · Score: 0

      But funnily 1 AMD cpu has 128 PCIe lanes, and 2 AMD cpus have 128 PCIe lanes, as half the I/O is reconfigured for inter-socket "Infinity Fabric" communications instead of PCIe.

  11. software Virtual RAID for pci-e?? intel only ssd? by Joe_Dragon · · Score: 1

    software Virtual (AKA fake) RAID for pci-e?? does that only work with intel pci-e ssd? and does it work with drives behind pci-e switches? hot swap?

    At least AMD gives you the pci-e lanes to do a system with lot's of pci-e storage and stuff like ZFS / CEPH / etc works good with storage in HBA mode not with all kinds of overhead like JBOD or 1 disk raid 0 setups.

  12. Databases by nojayuk · · Score: 3, Insightful

    The overall takeaway is that the AMD system is faster in a lot of workloads (databases being the one big notable area where it loses badly) than the Intel system and at a much lower cost.

    Databases are what the Big Iron servers live to support so AMD losing badly against Skylake on that front means they've lost the sales war. Web servers are databases, order processing systems are databases, pretty much everything that's computationally intensive has a database or six on the backend.

    High Performance Computing (HPC) is shiny and prominent but the sales are limited and a lot of new HPC kit is based around non-CPU computation elements derived from GPUs rather than general-purpose CPUs so even good performance in that area won't save AMD in the datacentre markets.

    1. Re:Databases by Anonymous Coward · · Score: 0

      You don't know what you're talking about, server market share is going increasingly to AMD now.

    2. Re:Databases by epine · · Score: 2

      Databases are what the Big Iron servers live to support so AMD losing badly against Skylake on that front means they've lost the sales war.

      Big language. Been watching too much Bruno Ganz lately?

      I appears Ars tested MySQL Percona Server 5.7.0 as their chosen representative for the entire category. I wouldn't recall Rommel's tanks just yet.

      Typically when high response times were reported, this indicated low single threaded performance. However for EPYC this is not the case. We tested with a database that is quite a bit larger than the 8 MB L3-cache, and the high response time is probably a result of the L3-cache latency.

      I have about 30 different database products listed in my notes (many oriented at graphs or machine learning, along the entire sharing spectrum). Would they all suffer this much?

      What does this mean to the end user? The 64 MB L3 on the spec sheet does not really exist. In fact even the 16 MB L3 on a single Zeppelin die consists of two 8 MB L3-caches. There is no cache that truly functions as single, unified L3-cache on the MCM; instead there are eight separate 8 MB L3-caches.

      Well, that does make the present EPYC implementation suck for a popular worker-thread model used to concurrently access a single, large datastore.

      I suspect, however, that a database server server hundreds of small databases as part of a WordPress server farm would hardly suffer at all (so long as CPU locality is stable at the OS level).

      Web servers are databases, order processing systems are databases, pretty much everything that's computationally intensive has a database or six on the backend.

      Or six. You even said it yourself.

    3. Re:Databases by alvinrod · · Score: 1

      All of those things you've listed (and many you haven't) might connect to a database, but they're not always running on the same system or even the same hardware as the database. Pick Intel where it makes the most sense based on performance and choose AMD when they're better suited to the task. The only difference is that AMD hasn't had a server chip suited to any task beyond space heater for over half a decade, whereas now there's a compelling reason to buy AMD for some workloads.

  13. Well... by xforce · · Score: 1

    With 112 threads, can you compile Chrome ans still use the mouse in Windows 10?

  14. Just don't ... by CaptainDork · · Score: 1

    ... build Chrome.

    Moreover, the problem gets worse the more cores you have.

    --
    It little behooves the best of us to comment on the rest of us.
  15. king Midas's plantinum tip by epine · · Score: 1

    Xeon Scalable processor family is now designated by Platinum, Gold, Silver, and Bronze categories, with a single model number.

    This naming shift gives me the major heebs. It can't possibly be designed to aid comprehension. It's the end of an era, for sure.

    I'd buy AMD almost for that reason alone, once I'm sure AMD is solid in the ZFS camp (the story there has been spotty for some while).

    But more likely, I'll run my current dual E5-2620 NAS convergence box into the dirt (I figure on another five to seven years) and by then I'll get a turnkey NAS appliance for local bulk storage and everything else I've got will migrate into the cloud, where no-one cares what is under the hood, so long as $/mile is priced competitively.

    1959 planar process
    1963 complementary MOS
    2017 bronze, silver, gold, and platinum

    58 good years, RIP. Apparently, the generation who grew up on cherry iMacs are procuring cloud servers these days.

    Turns out, the King Midas story is a bit oversimplified for young audiences. While he didn't grow hair on his palms, he did wind up with a mixed bag and the ultimate shiner (as you'd pretty much expect when a royal figure is granted his wish by an androgynous saint to the oppressed who wanders about waving a thyrsus).

  16. Hilariously fickle by Anonymous Coward · · Score: 0

    Heh. People are a fickle bunch. Go back not that many months ago when AMD was on the ropes and everyone was predicting their demise, not to mention all the praise for Intel CPUs and how one would be a fool to stick with AMD. People were leaving in droves, AMD fans, and Intel fanboys arm in arm. Fast forward and now people are badmouthing the former posterboy because there's now some competition. Heaven help AMD if all the usual bench-marking and review places suddenly find something wrong because the praise-a-thon crowd will flip-flop faster than a politician.

    1. Re:Hilariously fickle by Anonymous Coward · · Score: 0

      It wasn't the same people saying both things... most likely.

      I suspect fans of Intel were the ones predicting AMD's demise while all the AMD fans were clammed up since they didn't have much to talk about.

      Now that the position of AMD has shifted the AMD fans are piping up and the Intel fans are closing their mouths.