Intel preps 28-core Nova Lake-S CPUs for Dunlow workstation platform — Entry-level Xeon chip features LGA1954 socket

Intel Core Ultra
(Image credit: Intel)

Intel is working on a version of its Nova Lake-S processor platform codenamed Dunlow that will offer up to 28 cores and will target entry-level server and workstation applications, according to shipment manifests located in the NBD database by @x86deadandback.

Formally, Intel's codenamed Dunlow platform will succeed the company's Catlow platform with Xeon 6300P-series CPUs and will support Xeon E-class Nova Lake-S processors (presumably) with up to 28 cores that feature a dual-channel memory subsystem, come in an LGA1954 form-factor, and have a processor base power of 95W, according to shipments manifests at NBD data.

Latest Videos FromTom's Hardware
Anton Shilov
Contributing Writer

Anton Shilov is a contributing writer at Tom’s Hardware. Over the past couple of decades, he has covered everything from CPUs and GPUs to supercomputers and from modern process technologies and latest fab tools to high-tech industry trends.

  • bit_user
    Note: the embedded chart doesn't have a "zoom" button. If you click this one, you can zoom it to full size:

    The article said:
    Therefore, unless Intel plans to offer energy-efficient cores in its next Xeon CPU aimed at entry-level servers and workstations, we may be dealing with a very special processor that features 28 P-cores that is designed to beat all desktop-grade platforms in demanding applications.
    Don't go creating silly rumors!

    99% likelihood this is just a Nova Lake with single compute tile. Especially at 95W!!!

    The article said:
    it should also be noted that 28 cores do not naturally derive from a 16P+32E desktop design and are impossible to derive from a notebook-grade 8P+16E design.
    How did you not get the memo that Nova Lake's die configuration is one of:
    8P + 16E + 4LPE
    2 * (8P + 16E) + 4LPE???

    The first clearly shows how they arrive at 28 cores.

    Source: https://www.tomshardware.com/pc-components/cpus/intel-nova-lake-specs-leaked-up-to-52-cores-and-150w-of-tdp-for-intels-amd-zen-6-rival
    The article said:
    the abandoned Raptor Lake-32C, which featured an all-P-core design aimed at workstations and entry servers before being canceled.
    But, that would never have gone in a desktop socket!! It must've been planned to use the same socket as Xeon-W2400.

    Either that, or it was a 0P + 32E chip. That I could easily see slotting into LGA1700 and would make a lot of sense for light-duty servers.
    Reply
  • thestryker
    It seems very odd that Intel would be bringing hybrid design into Xeon unless they're rebranding the edge processors. The other super cynical guess would be dropping ECC support from desktop parts, but that doesn't seem particularly likely.
    bit_user said:
    But, that would never have gone in a desktop socket!! It must've been planned to use the same socket as Xeon-W2400.
    I think the article was referring to this which I don't recall ever seeing any real clarification on: https://www.tomshardware.com/news/intels-unannounced-34-core-raptor-lake-cpus-displayed-on-wafer
    Reply
  • Gururu
    bit_user said:
    Note: the embedded chart doesn't have a "zoom" button. If you click this one, you can zoom it to full size:


    Don't go creating silly rumors!

    99% likelihood this is just a Nova Lake with single compute tile. Especially at 95W!!!


    How did you not get the memo that Nova Lake's die configuration is one of:
    8P + 16E + 4LPE
    2 * (8P + 16E) + 4LPE???

    The first clearly shows how they arrive at 28 cores.

    Source: https://www.tomshardware.com/pc-components/cpus/intel-nova-lake-specs-leaked-up-to-52-cores-and-150w-of-tdp-for-intels-amd-zen-6-rival

    But, that would never have gone in a desktop socket!! It must've been planned to use the same socket as Xeon-W2400.

    Either that, or it was a 0P + 32E chip. That I could easily see slotting into LGA1700 and would make a lot of sense for light-duty servers.
    Didn't Anton Shilov write this article? :P
    Reply
  • bit_user
    thestryker said:
    I think the article was referring to this which I don't recall ever seeing any real clarification on: https://www.tomshardware.com/news/intels-unannounced-34-core-raptor-lake-cpus-displayed-on-wafer
    In that article, it says:
    "The die appears larger than the die that drops into the standard desktop PCs with the LGA 1700 socket, so it appears to be too large to fit inside the package for desktop PCs. That means this is likely a CPU destined for the workstation market. "
    So, it was some sort of HEDT or Xeon-W die, most likely. That's why I said it would never slot into LGA1700 and must've been destined for LGA4677.

    This article is specifically talking about Dunlow being LGA1954. So, there's no way it's 28 P-cores. Not in 95W. That's just crazy talk.
    Reply
  • bit_user
    Gururu said:
    Didn't Anton Shilov write this article? :P
    I think Anton doesn't track desktop PC products as well as he tracks silicon engineering. Like everyone, he has his niches and areas of interest.

    The fact that he didn't know about the single compute tile Nova Lake die configuration is understandable, from that perspective, but still something that should've been caught by an editor.
    Reply
  • Jaykihn0
    Yeah, this is a standard Nova Lake -S 8+16+4 silicon stepping derivative.
    Reply
  • thestryker
    bit_user said:
    So, it was some sort of HEDT or Xeon-W die, most likely. That's why I said it would never slot into LGA1700 and must've been destined for LGA4677.
    It's a wafer full of EMR tiles, but at the time I don't believe it had been announced. That was just the only thing I could think of which ever carried the "RPL-S" title which had 3x cores.
    Reply
  • Stomx
    A: It is 8P + 16E + 4LPZ
    B: No it is 3P + 4B + 12E + 4LPE+14XP
    A: At the maximum it is 8P + 12E + 4LPE + 4Z
    B: As to the maximum then it is 3P + 4B + 12E + 4LPE+14XP+3P + 4B + 12E + 4LPE+14XP
    A: This is what I spent 3 years on and uncovered that in my PhD thesis. And do your math fool.
    B: And who are you then?
    Reply
  • usertests
    Stomx said:
    A: It is 8P + 16E + 4LPZ
    B: No it is 3P + 4B + 12E + 4LPE+14XP
    A: At the maximum it is 8P + 12E + 4LPE + 4Z
    B: As to the maximum then it is 3P + 4B + 12E + 4LPE+14XP+3P + 4B + 12E + 4LPE+14XP
    A: This is what I spent 3 years on and uncovered that in my PhD thesis. And do your math fool.
    B: And who are you then?
    C: Don't quit your day job!
    Reply
  • TerryLaze
    bit_user said:
    This article is specifically talking about Dunlow being LGA1954. So, there's no way it's 28 P-cores. Not in 95W. That's just crazy talk.
    I have no opinion either way but why is 95W such an issue for you?!
    It's entry level so it could be the equivalent of the -T desktop models, significantly lowered PBP to run cooler and be much more efficient (and you can still give them the same amount of power as the normal ones to get the same speed and inefficiency) .
    Reply