Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!


Forgot your password?
Intel Hardware

Intel X38 High End Chipset Launch and Benchmarks 87

MojoKid writes "Though many leaks of the product have been circulating for some time, Intel officially took the wraps off and launched their new X38 Express chipset for the high-end desktop motherboard market. With this launch, the Intel desktop chipset line-up gets a new flagship. Intel's new X38 chipset encompasses all of the technology advances that have made the P35 a success and adds a slew of new features designed to increase memory and graphics subsystem performance, like PCI Express 2.0 SerDes and Intel Extreme Memory technology in the new X38 MCH. The Asus motherboard tested by HotHardware even features an embedded Linux-based OS that boots in a matter a seconds."
This discussion has been archived. No new comments can be posted.

Intel X38 High End Chipset Launch and Benchmarks

Comments Filter:
  • The article description says "High end", but the link describes a single processor configuration. Am I missing something?
  • by crowbarsarefornerdyg ( 1021537 ) on Saturday October 13, 2007 @02:38PM (#20967683)
    Windows AfterVista®?
    • Yes (Score:1, Funny)

      by Xiph ( 723935 )
      It'll run xwindows quite nicely

      After having tried Vista at work, the partition, i had set up for my third os, soon became infected with ubuntu.
      Not all infections are bad :)
  • Apple next? (Score:1, Interesting)

    by Anonymous Coward
    This may mean we'll soon see a Mac hardware announcement from Apple that uses X38.
    • Re:Apple next? (Score:4, Interesting)

      by Joe The Dragon ( 967727 ) on Saturday October 13, 2007 @02:54PM (#20967777)
      we more likely to see a g33 / g35 system from apple to replace the mini with on board video and x16 and x4 pci-e slots.

      And apple will making a dumb move by going with over priced DDR3 that is not much faster then much cheaper DDR2 ram.
      • we more likely to see a g33 / g35 system from apple to replace the mini with on board video and x16 and x4 pci-e slots.

        I think you're dreaming. Apple wants people to use minis for home entertainment centers, not to use them for anything that requires oomph. I'd love them to do something like that as well, but once bitten...
      • Re: (Score:3, Interesting)

        by p0tat03 ( 985078 )
        The same Apple that has stuck with DD2-667 even on their high-end workstations, despite there being faster DDR2 clocks? Apple doesn't look like they're the type to cash in on the newest buzzword-tech, which is sometimes good, and sometimes bad (I do wish the Mac Pros would ship with something a tad faster than DDR2-667...)
        • by teg ( 97890 )

          The same Apple that has stuck with DD2-667 even on their high-end workstations, despite there being faster DDR2 clocks? Apple doesn't look like they're the type to cash in on the newest buzzword-tech, which is sometimes good, and sometimes bad (I do wish the Mac Pros would ship with something a tad faster than DDR2-667...)

          Mac Pros use FBD, not DDR2.
    • This may mean we'll soon see a Mac hardware announcement from Apple that uses X38.

      We may be dreamin' (as another comment in this thread said), but I think this would be a great chipset for the mythical "xMac" or "Mac XL" for all those desktop buyers who don't want an underpowered SFF (Mac mini), an all-in-one with few options (iMac), or an overpowered dual-processor workstation (Mac Pro).

      The X38 chipset offers Apple a choice of implementing four DDR2 or DDR3 slots, which are both better options than the iMac's two SO-DIMM slots or the Mac Pro's FB-DIMM slots. The CPU socket will accep

  • And someone will buy this, then load Windows Vista on it, where all its uber spiffy performance will be limited by a broken USB driver, or the fact that their DSL link is only 768kbps, or they chintzed out and bought a 5400 rpm drive.
  • DDR3 ECC supported! (Score:4, Informative)

    by DigiShaman ( 671371 ) on Saturday October 13, 2007 @03:05PM (#20967861) Homepage
    It's about damn time. I've looked all over the net to find an official answer. Not even Intel's webpage mentions anything about the X38 supporting ECC. So, I downloaded the manual for the Asus P5E3 Deluxe board. From page 2-13 of the PDF document.

    You may install 512, 1GB, and 2GB unbuffered ECC, non-ECC DDR3 DIMMs into the DIMM sockets.

    To Intel and vendors: How bloody hard is to include "supports ECC" in your online product summary?
    • Well, most people run Windows and attributes any crash to Microsoft, not memory errors.
    • by Anonymous Coward on Saturday October 13, 2007 @03:21PM (#20967983)
      It may support ECC DIMMs, but that doesn't necessarily mean it'll use the ECC feature of those DIMMs.
      • Re: (Score:3, Interesting)

        Mod parent up, informative! My cheapass motherboard supports ECC DIMMs, but does not make use of the ECC features. There's no way to know whether it'll use the ECC feature unless you either A) try it and see (hard to do), or B) Intel tells you it does (in which case you have to accept their word for it.)

        It seems to me that a high-end motherboard chipset should support ECC features, but that doesn't mean that it does.
        • I would imagine that Intel's memory controllers do support ECC. I don't know if it compares, but the Alpha version of Linux put out console warnings. I know that the workstation and server chipsets do, I've seen screen shots of Mac's System Profiler on the Mac Pro do show a count of ECC errors if they use FB-DIMMs with inadequate heat sinks.
    • Re: (Score:3, Informative)

      by Agripa ( 139780 )
      Intel's chipset comparison page and datasheet are somewhat informative. Apparently, the X38 chipset supports ECC when using DDR2 but not when using DDR3.
      • by DigiShaman ( 671371 ) on Saturday October 13, 2007 @10:58PM (#20970727) Homepage
        You sir, would be correct. I found the documentation on Intel's website. Found in 31846901.pdf section 3.3 titled ECC Support.

        For DDR3 the X38 Express Chipset does NOT support ECC, does not support ECC unbuffered
        DIMMs, and it does not support any memory configuration that mixes non-
        ECC with ECC un-buffered DIMMs.

        For DDR2 the X38 Express Chipset does support ECC and ECC un-buffered DIMMs but
        it does NOT support any memory configuration that mixes non-ECC with ECC unbuffered

        Damn, that blows. And shame on Asus for false documentation!
        • by Agripa ( 139780 )
          I examined Intel's documentation and did not find any specific reason for the lack of DDR3 ECC support. The memory controller uses the same burst length for both memory types and the 64,72 hamming code should work in either single or dual channel mode. None of the immediately preceding chipsets support ECC at all.

          Maybe the cause is cost or marketing?
          • Perhaps DDR3 ECC is not supported because the memory controller (North Bridge) can't up with the increase in calculation overhead that is necessary. I wouldn't think that would be the issue if all ECC calculations are a synchronous event with memory read/write operations.

            Just a guess.

            • by Agripa ( 139780 )
              That occurred to me but DDR2-800 supports ECC while DDR3-800 does not. DDR3 ECC is widely available so I also doubt it was for lack of modules to test.
              • by Agripa ( 139780 )
                Nix the DDR3 ECC comment. Crucual's dynamically updated page leave something to be desired. DDR3 ECC would seem to be rather rare so perhapse Intel was not in a position to test with it.
    • Intel allows the motherboard manufacturer to determine memory support. They created the chipset to handle DDR2 or DDR3 in whatever configuration the motherboard manufacturer desires. It is my understanding that on Intel's X38 desktop board configuration, they are supporting non-ECC DDR3 only.
    • It's not hard at all; Intel fully lists all supported memory on their motherboard support pages. When Intel releases the X38 chipset motherboard, they will list exactly what RAM their motherboard supports. It's up to the motherboard manufacturer to list what RAM a board supports, not the maker of the chipset.

      To Intel and vendors: How bloody hard is to include "supports ECC" in your online product summary?
    • by RMingin ( 985478 )
      Actually, you're still without an answer, because the mobo maker said ECC ram works, not that the ECC checking is done.

      Them Taiwanese mobo guys are shiftier than a bad transmission, never ASSUME they mean anything.
    • by tokul ( 682258 )

      To Intel and vendors: How bloody hard is to include "supports ECC" in your online product summary?
      You looked at the wrong page. Check links under "Server Boards" and "Server and workstation chipsets". Intel supports ECC and Windows Server drivers only on pricey server broads and chipsets.
  • My computer boots in a matter of seconds also. It just so happens that it is in a matter of 120 seconds.
  • Intel and Linux (Score:5, Informative)

    by Simon (S2) ( 600188 ) on Saturday October 13, 2007 @03:21PM (#20967985) Homepage
    I just love Intels commitment to linux lately. They release open source drivers for their chipsets, and now an integrated linux os on the firmware of this mobo. The last notebook I had was all Intel chips (IPW2200 for wireless, GM945 graphics) and just everything worked out of the box without proprietary drivers. Really, thanks Intel, I am a happy customer.
    • Re:Intel and Linux (Score:5, Informative)

      by Justus ( 18814 ) on Saturday October 13, 2007 @05:31PM (#20968773)

      Although Intel's Linux driver support is pretty good on the whole, the integrated OS is a feature of the ASUS motherboard and isn't a product of Intel's good will toward Linux.

      There was a previous Slashdot feature specifically covering that [slashdot.org], if you want more information.

      • by niko9 ( 315647 )
        Mod parent up.

        Intel maybe be driver friendly when it comes to video and wireless devices, but they are sorely lacking in the LinuxBIOS dept or any other open BIOS for that matter. Most of the motherboards supported by LinuxBIOS are on AMD or NVIDIA chip sets. Intel is still trying to push EFI, which Linus, on the LKML, called: "this other Intel brain-damage (the first one being ACPI)".
        • Why should supporting an alternative bios have any priority for intel?
          It is far more important to have decent drivers to be able to use the hardware under a specific OS.
          For 99 percent of the users there will be absolutely no need to look at an alternative bios for their system. The bios supplied by the manufacurer is all they need.
          I am one of those people. I want my hardware to work and I see no reason to replace my bios with an open source version. The one Asus supplied does its job.

          Arguments like 'but it
    • I wonder if that isn't one of the major factors in AMD's recent decision to open up their GPU specs.

      Please, Intel and AMD, get into a war of openness one-upmanship. Everyone benefits.
  • by niteice ( 793961 ) <icefragment@gmail.com> on Saturday October 13, 2007 @03:36PM (#20968095) Journal
    does it run...oh
  • Express Gate appz (Score:2, Interesting)

    by Kensai7 ( 1005287 )
    If you launch the Express Gate applications, which are comprised of a web browser and Skype at this time, the embedded Linux-based OS is launched from a ROM and seconds later it's available for use. We found the Express Gate technology easy to use and quite handy. Say, for example, you need to download a driver or BIOS file and the hard drive-based OS isn't functioning properly. With Express Gate you can now access the web and integrated peripherals even if the system's full blown OS has a problem.

    This m
    • Intel has its own similar implentation and they've had it for a year now - it's called vPro. And you can do it remotely. However it's implemented, it's a great feature and I agree that every manufacturer should be building this into their boards.
    • by Ant P. ( 974313 )
      I wouldn't call that a "mini-OS" - they've got a full install of Firefox on there.
  • Benchmark result: (Score:3, Informative)

    by imsabbel ( 611519 ) on Saturday October 13, 2007 @04:07PM (#20968315)
    Same performance as any other board (inside the margin of error).
    Seriously, why even bother benchmarking?
    You buy those boards for compatibilty, or for their features, but not because they are _faster_....
    (extreme overclockers excluded. Some board may be better suited for FSB overclocking).
    • and considering this being marketed as a high-end mobo...

      extreme overclockers excluded. Some board may be better suited for FSB overclocking

      thats exactly who this product is primarily being marketed for. tho it remains a fairly valid question, why bother with benchmarks, unless of course the benchmarks are actually to see how far the mobo can overclock.

  • Has anyone tried Open Solaris with a board with an X38 chipset?

    I'm looking for a board that supports at least a 6 SATA drive RAID.
  • by Zymergy ( 803632 ) * on Saturday October 13, 2007 @05:18PM (#20968655)
    http://news.softpedia.com/news/Intel-039-s-X48-to-Come-in-Just-Another-5-Weeks-67604.shtml [softpedia.com]
    http://www.dvhardware.net/article22289.html [dvhardware.net]
    It appears the X48 chipset is actually the X38 chipset without the ECC support and for DDR3 Only? Great, just when we weren't confused!

    Here's another X38 review: http://www.anandtech.com/cpuchipsets/intel/showdoc.aspx?i=3120 [anandtech.com]
    A chipset comparison graphic: http://images.anandtech.com/reviews/chipsets/intel/x38-launch/memory-lg.png [anandtech.com]
    And another review: http://www.tomshardware.com/2007/09/26/intel_x38_chipset/ [tomshardware.com]
  • A Gamer's Dream? (Score:5, Insightful)

    by abirdman ( 557790 ) <abirdman@maine.rr . c om> on Saturday October 13, 2007 @06:20PM (#20969177) Homepage Journal
    Observations from reading the article:
    1. The two MB's they reviewed are going to be incredibly expensive (over $300) when they hit the streets,
    2. the only one that survived overclocking the already insanely fast CPU (the other one died and could not be revived), takes ridiculously expensive DDR3 memory, and
    3. when run at their max (optimal, i.e. stock) settings showed no difference in performance from their "hopelessly outdated" predecessors, the P35 chipset boards.
    4. Only one game benchmark (F.E.A.R.) showed any measurable difference at all from one to the other.
    5. All the video and audio encoding benchmark results were identical, which only proves those tasks are processor bound (all the test systems used the same CPU-- only the chipsets and/or memory differed).

    I find it odd the reviewers even recommended the board (the survivor-- they were skeptical of the dead one). I don't understand the attraction of a board/chipset like this! It's going to take another generation of hardware to take advantage of the 32 simultaneous 32 bit video data "lanes" on each PCI-E (or X or whatever) slot. And eventually, maybe DDR3 will drop in price when there's some demand for it. And all the I/O (8 USB 2.0 ports and external SATA ports and optical and coax digital AV) seem like they could come in handy. But seriously, why are they making these now? Is it for the quad-core support? Do other chipsets support quad-core Intels? Or is it because they allow plugging in not one but two $500+ dollar video cards?

    I look forward to lots of serious gamers buying these, devising new benchmarks to prove their efficacy, and bringing down the cost for this point of entry into the market for the rest of us. But gamers! Read the review and benchmarks. This chipset does not, at least based on this review, demonstrate a big leap forward.

    • by Ant P. ( 974313 )
      That's because this isn't aimed at gamers. It's aimed at rich idiots, who will pay to beta test new hardware.
    • Actually the key selling points of X-38 for an overclocker seem to be PCIe 2.0 and a heatspreader on the northbridge chip. PCIe 2.0 has twice the data transfer rate of PCIe 1.1. It's sort of like buying a motherboard with AGP8X versus AGP4X. It remains to be seen what difference, if any, it makes in game benchmarks. The heatspreader may be an advantage for cooling the northbridge which may allow for higher FSB clocks eventually, but all of this is still speculative. And these boards are awfully expensive.
  • ... it supports ECC DDR3 at 1333 and TWO Xeon quad core processors all in an ATX form factor.

  • Intel are noobs, everyone knows the new X38 are 48 less powerful then the x86.

  • What would be even cooler, IMO if it was embedded with VMWARE Virtual Infrastructure. Still, anybody know if this chipset will support ESX VI? (weould use (iSCSI SAN) Been looking to get some 'high end" equiptment that will run many VM's at once. Preferably building it. Yeah sorta of topic ... so atleast I am assuming that current Linux kernel supports this? or is the "embedded linux" highly modified? Either way , kewlness if I can get it for under $4000 with 4 GB DDR3 SAS RAID etc. Also RAID (real raid no
  • Not sure why the slashdot editor stripped out the links to the full version of the article but none of the full sized images don't work in it when you click the thumbnails.

    Please mod this up for all.
    Here is the full version article: http://www.hothardware.com/Articles/Intel_X38_Express_Chipset_Debuts/ [hothardware.com]
  • I have been dissapointed with Intel's two last motherboard releases. The P35 isn't any better than the P965 it is supposed to replace (check the article at techreport.com). They also have an article on the X38, and seems that the only difference is DDR3 (which at this time, it's SLOWER than DDR2 because of the high latency timings, despite the increased bandwidth, and of course, DD3 is double the price of DDR2 right now)

    So, if you are building a new system, get a cheap full featured P965 for just 100$.


Today is a good day for information-gathering. Read someone else's mail file.