Comments Locked

67 Comments

Back to Article

  • DanNeely - Tuesday, April 26, 2011 - link

    Do you have any confirmation from Intel about x79 using a QPI link instead of DMI? The older rumor/leaks I've seen on LGA2011 all listed it as using the latter to connect to the southbridge.
  • JarredWalton - Tuesday, April 26, 2011 - link

    The information we have doesn't specify the link type/speed, which is why we say "assuming it uses a similar QPI link..." If it's DMI, then the additional storage-dedicated lanes make a lot more sense, but then the PCIe links would have to come straight from the CPU or they'd saturate the DMI link, right?

    The Sandy Bridge and Ivy Bridge platforms both use DMI, because the CPU has the PCIe graphics link, but I believe that's the whole point of LGA2011. Also, as noted above there's a huge increase in the pinout of LGA2011 relative to LGA1366, and I can't think of where they would use all those extra pins if they're not doing a full QPI link.
  • JarredWalton - Tuesday, April 26, 2011 - link

    Okay, discussed with Anand and updated. It's no *entirely* clear whether SNB-E will use QPI or DMI. Given the extra four lanes for storage, DMI would be very likely. What does this mean for Xeon, though? Maybe the chips have the option to do QPI or DMI, but they're using DMI on single socket configurations and consumer models, and QPI will only be on multi-socket. Or maybe they'll use a different die for Xeon. Lots of possibilities, but the text now reflects the fact that some of this isn't entirely clear, and we make some educated guesses as to what's going on.
  • DanNeely - Tuesday, April 26, 2011 - link

    I've been scratching my head about the pincount as well. LGA 1567 is quad channel and 4x QPI, so the only new on die features adding pin count would be the PCIe lanes, and DMI link (4x PCIe). Each lane is 4 data pins, so 40 pcie lanes (32 +4 sata + 4 dmi) is 160 pins for data, and a few more for control/clock/etc (Wikipedia isn't clear which control signals are per slot which are shared), for maybe 200ish total. I'm going to assume that intel will also be using this socket for their forthcoming 4/8way boards and will keep the four QPI links. If they drop back to only two there're another 164 pins unaccounted for.

    Maybe the remaining 250ish are just extra power/ground connections in order to support bigger/faster chips?
  • Kristian Vättö - Tuesday, April 26, 2011 - link

    http://vr-zone.com/articles/intel-x79-roadmap-leak...

    That says that the SATA ports will have total bandwidth of 8GB/s (64Gb/s) due to the four extra PCIe lines. That would be enough even for 10 SATA 6Gb/s SSDs. A big upgrade from X58 though, since it tops out at around 660MB/s from what I have heard.

    This is starting to be very complex lineup for Intel if some parts use QPI and others DMI, let alone what LGA 1356 will add to the mixture :D
  • DanNeely - Tuesday, April 26, 2011 - link

    some parts using QPI and others DMI is the same situation as we had with LGA1156/1366, so I don't see anything changing.

    LGA 1356 rumors are operating at a much lower level than LGA1155/2011 rumors have. It's an open question if it's a real product or not. My guess is that it either always was vaporware, was a real part at one point but completely discontinued to only have 2 sockets vs 3 for the last architecture, or is a plan B part if bulldozer turns out fast enough that quad core/dual channel Ivy Bridge CPUs aren't fast enough to compete in the $200-300 range, and hexcore dual channel DDR3 IB chips would be badly memory bottle necked.
  • Kristian Vättö - Tuesday, April 26, 2011 - link

    The LGA 1356 is pretty much confirmed from what I have read and heard from my sources. Currently it looks like it will be meant for low-end DP systems so it is unlikely that we will see any consumer parts for that socket (i.e. only Xeons).

    That would mean that there will be three sockets of SB Xeons (don't forget E3 lineup for LGA 1155 socket).
  • DanNeely - Tuesday, April 26, 2011 - link

    Anything publicly available on highly trustworthy sites? I don't think I've seen anything on sites I consider A list for reliability. This isn't to say I think all the sites reporting the rumors are questionable. some might quality, but there're far more sites reporting rumors than I normally follow and can evaluate.
  • Casper42 - Tuesday, April 26, 2011 - link

    I was in a briefing by Intel in Feb and one of the things I remember complaining about was the complexity of the SB lineup in servers.

    I remember seeing the following:
    1 way with Dual Channel
    1/2 way with Triple Channel
    2/4 way with Quad Channel (4 way design here would need 2 hops for some chip-to-chip communication due to limited QPI links so it was like the poor man's 4 way box)
    4/8 way with Quad Channel (this had more QPI links thus making it a proper 4 way box)

    I wanna say they were socket H, B2, R and lastly EX-B for the big boys. (same order as specs above)

    I straight out called them out on it as AMD has boiled their lineup to a simple 1/2 Dual Chan socket and a 2/4 Quad Chan socket. Much easier to build servers when the chips are clearly delineated.

    So assuming 2011 is R and I know 1156 is H, that certainly leaves room for 1356 to come in as B2. EX-B will likely be the last addition as it will supplant the soon to be released Westmere-EX chips. The monster 10 core 30MB cache ones.
  • Kristian Vättö - Wednesday, April 27, 2011 - link

    http://www.cpu-world.com/news_2011/2011020803_Deta...

    Socket B2 is same as LGA 1356. Obviously, I cannot confirm that this is 100% accurate. It does make sense though. Currently Intel doesn't offer much in the low-end DP market due to their high prices. Maybe LGA 2011 parts are simply too expensive for Intel to manufacture, so they couldn't replace the E5502 for example.

    LGA 1356 with crippled features would let Intel stay competitive, even in lower-end markets. Cut down the PCIe lanes, stick with triple-channel, possibly no ECC and stuff like that.

    Server grade stuff is not usually reported by these regular news sites so getting information that is available for everyone is harder. Subscription is usually required to get access to information like this. I don't have access like that so this is just what I heard from a friend of mine who has access to this kind of information.
  • SteelCity1981 - Tuesday, April 26, 2011 - link

    "USB 3.0 support is still missing,"

    Seriously Intel? I mean, seriously intel this is your new highend chipset and you still aren't going to support USB 3.0. Freaking stupid........
  • Kaboose - Tuesday, April 26, 2011 - link

    Intel is probably trying to get people to adopt their lightpeak/thunderbolt technology instead, but i agree USB 3.0 WOULD be a nice thing to have.
  • Ammaross - Tuesday, April 26, 2011 - link

    It's not because they're pushing LightPeak, since they'd have onboard support for LightPeak instead of USB3 to actually make that assumption work. My bet is they're just behind on specs.
  • Marlin1975 - Tuesday, April 26, 2011 - link

    Thats because they are trying to push their new connection.

    AMD will have native USB3 on the Llano chips.
  • JarredWalton - Tuesday, April 26, 2011 - link

    I think it has more to do with the fact that X79 was probably specced out and started at the same time as the 6-series chipsets. The upcoming Ivy Bridge platform should have both native USB3 and ThunderBolt support.
  • greylica - Tuesday, April 26, 2011 - link

    Gazzilions and Bazzilions....
  • Kristian Vättö - Tuesday, April 26, 2011 - link

    X79 is an enterprise level chipset and they are more complex than consumer (LGA 1155) stuff. Since they are more complex, they require more money and time to be done right, thus it is easier and cheaper for Intel to not include USB 3.0 in X79.

    As consumer stuff isn't as complex, adding USB 3.0 to Ivy Bridge is easier
  • Spivonious - Tuesday, April 26, 2011 - link

    Aside from external hard drives, what needs the speed of USB 3? I think Intel is playing it smart here and waiting to see what the market response to USB 3 will be.
  • name99 - Tuesday, April 26, 2011 - link

    "Aside from external hard drives, what needs the speed of USB 3?"

    That's a pretty big aside...

    Yes, no-one is asking for USB3 printers. So what?
    We want faster external and portable drives. This is a perfectly reasonable request. For most people, the speed of their storage is THE determining factor as to the perceived speed of their computers, not the CPU specs.

    At some point, there will probably also be a transition period between 1G and 10G ethernet, where it would be nice to retrofit 10G ethernet to an older machine through USB3.

    It's nice to hope that one can get USB3 through Thunderbolt, but right now there are, as far as I can tell, precisely ZERO USB3<-->Thunderbolt bridges available for purchase.
  • GullLars - Friday, April 29, 2011 - link

    "For most people, the speed of their storage is THE determining factor as to the perceived speed of their computers, not the CPU specs."
    I can confirm this as absolutely true.
    For >90% of consumers, a 2-2,5GHz C2D is fast enough, 2-3GB RAM decent, and NV x20/x30 or AMD x3xx/x5xx sufficient. Give them a decent SSD for OS + Apps and they'll sing you praises.

    I've been using SSDs since 2008, and whenever i have to use a computer with a HDD (system drive), regardless of CPU and RAM, i feel like i want to smash my head against the wall or throw the computer out the window.
    Since i upgraded my family and friends computers with SSDs, there have been virtually no complaining about them slowing down, and those who used to buy a new computer every 2-3 years now have no desire to do so (other than a couple of gamers :P).
  • MrBlonde - Tuesday, April 26, 2011 - link

    If Intel planned USB 3 on this generation of I/O controllers, they'd have started before USB 3 was finished. And if they tried to tack it on afterwards... they could have something like the Cougar Point flaw. This would also be unfair to AMD because they have to wait for Intel to release USB 3 specs then design the specs into AMD chipsets.
  • slyck - Tuesday, April 26, 2011 - link

    And we all know Intel wouldn't want to be unfair to AMD.
  • Shadowmaster625 - Tuesday, April 26, 2011 - link

    I have a great idea! Let's all rush out and spend $700 on a cpu/mobo/RAM and then have to bend over for some third party's usb 3.0 card or chipset, for absolutely no reason whatsoever. It's funny, because one of the biggest bottlenecks on today's computers is usb transfer speeds. It is really becoming quite annoying.
  • ggathagan - Tuesday, April 26, 2011 - link

    ...which is why Intel developed Thunderbolt.
  • skrewler2 - Tuesday, April 26, 2011 - link

    That quad channel is awesome.. prices have dropped to where putting 32GB in it wont break the bank.

    Any idea what onboard storage controller we're talking about?
  • skrewler2 - Tuesday, April 26, 2011 - link

    Also.. I assumed the boards will have 8 DIMM slots. Any word on whether boards will have 4 vs 8?
  • DanNeely - Tuesday, April 26, 2011 - link

    Rumors I've seen earlier indicated that LGA 2011 systems using unbuffered (normal) ram would only support a single dimm per channel in order to increase the maximum validated speed. A few blurred mobo pictures leaked months ago showed 2 memory sockets on either side of the CPU.
  • JarredWalton - Tuesday, April 26, 2011 - link

    I would guess eight slots, though it's up to the motherboard vendor as well. We know it's easy enough to put two DIMMs on a single channel, so it's just a matter of finding space. I expect we'll see 4-slot mATX and 8-slot eATX, and the regular ATX will be a mix of both.
  • cactusdog - Wednesday, April 27, 2011 - link

    Could be, but with 8GB Dimms coming i dont see a need for any more than 4 slots on consumer boards. Especially when 1 dimm per channel means higher speed and lower latency. But yeh i agree server boards will probably have 8 slots.

    Thanks for the info, cant wait for this platform.
  • GullLars - Friday, April 29, 2011 - link

    8GB DIMMs won't be cheap for consumers. 4GB DIMMs are finally approacing 2GB ones in $/GB, so i'm guessing early (first 6 or so months) 8GB DIMMs will be about twice the $/GB of 2GB DIMMs for the same speed and latency.
  • evilspoons - Tuesday, April 26, 2011 - link

    I'm dying just waiting for Z68 to show up so I can buy myself a 2500k or 2600k. I think LGA2011 would be a better fit for me but dang, my Q6600 is a bit old and creaky.

    (FWIW it's pretty amazing I've forgone upgrading for the last 3.5 years and hardly even noticed until the last couple months. Not bad for someone who used to upgrade twice a year!)
  • 7Enigma - Tuesday, April 26, 2011 - link

    Same here! Currently on a E8500 with ATI 4870 and really starting to get the upgrade itch. The whole Sandy Bridge fiasco with overclocking or on-board graphics but not both kept me calm the last couple of months (not to mention the SATA mobo bug) but it's starting to get tough not to bite the bullet and rebuild.

    choices choices.
  • GullLars - Friday, April 29, 2011 - link

    Both you guys, if you don't have SSDs already, that could tide you over to Q4. If you do, i can understand your itch to upgrade. I gave into it on an AMD system this fall, and couldn't wait for bulldozer. It did earn me the PCMark Vantage WR for AMD systems though :D (with only air-cooling)
  • JMS3072 - Tuesday, April 26, 2011 - link

    One important thing to note is that you'd need 4 DIMM slots populated to get full quad-channel badnwidth. I also suspect that we're going to see a return to 4 slots on most boards from the 6 present on most 1366 boards- which means that if you build a rig to take full advantage of the quad-channel architecture, upgrading your RAM later on is going to get pricey.
  • Casper42 - Tuesday, April 26, 2011 - link

    Except that you can get 16GB of RAM these days for like $150
    I can only imagine that unless the Japan situation has a major impact on prices, that this could be even cheaper by the launch date.
    And by then, the enthusiasts who can afford a $500/1000 CPU and $300 mobo could probably afford to load that thing with 8GB DIMMs as well.
  • cactusdog - Wednesday, April 27, 2011 - link

    Also, we are at a point where memory upgrades are pointless. Everyone has more than enough memory these days.

    A 4x4GB kit is going to be plenty for a long time to come. Even a 4x2GB kit is more than enough for the home user for the next 3 years at least.
  • JMS3072 - Thursday, April 28, 2011 - link

    1981 called. Bill Gates wants his quote back.
  • cactusdog - Thursday, April 28, 2011 - link

    haha i didnt say ever.....the thing is people will move on to the next platform before there is a need for more than 32GBs of memory. They'll probably upgrade 3 or 4 platforms before then.
  • DanNeely - Thursday, April 28, 2011 - link

    The bigger issue for most users is that w7 home premium can only use 16gb. To go higher you need to upgrade to pro/ultimate which nets you 192gb.
  • GullLars - Friday, April 29, 2011 - link

    Well, not entirely pointless. If you have a >4GHz Hexacore, a RAM-disk to install all apps or even some games on would certainly net a noticable bump, even over the insane SSD RAIDs you can make on X79.
    I'm on a AM3 rigg currently holding the PCMark Vantage WR for AMD systems (only about 17K marks though :P), and my bootdrive is a 256GB SSD RAID (4x64GB) pushing well over 1GB/s and 100.000 IOPS, but still i can notice a bump in speed with a RAM-disk for certain apps. Most of the time though, my 4,4GHz 1090T just maxes one core or maybe two, since most programs still are not heavily threaded.
  • mfago - Tuesday, April 26, 2011 - link

    I had recalled LGA2011would be 40 lanes of PCIe 3.0? Any news on this?
    Thanks!
  • DanNeely - Tuesday, April 26, 2011 - link

    32 main slots, 4 sata extra BW, 4 DMI = 40 total. (About half the LGA1155/1156 prerelease reports counted the DMI lanes in the total PCIe count).
  • JMC2000 - Tuesday, April 26, 2011 - link

    I think PCIe 3.0 will be included in Ivy Bridge chips, but the spec was released too late to make it into Sandy Bridge.
  • Filiprino - Tuesday, April 26, 2011 - link

    I thought 8 core Sandy Bridge-E were planned for LGA2011.
  • DanNeely - Tuesday, April 26, 2011 - link

    Same here and I've seen them on recently leaked roadmaps. Maybe they're not launching at release time, or they'll be coming out as XEON's not consumer branded parts.
  • JarredWalton - Tuesday, April 26, 2011 - link

    I believe those are going to arrive some time in 2012, but don't quote me on that. LOL
  • hansmuff - Tuesday, April 26, 2011 - link

    That's a lot of extra traces to put between CPU and memory. It seems to me that this will be an expensive platform compared to 1155.

    Couple that with the "limited unlock" of that 4-core i7 they will put into the 2600K price range, and I'm not entirely sure of the value of the "low end" 2011. The 2600K seems to have no trouble running at 4.4GHz for a lot of people, so I'd set that as a benchmark.

    The memory throughput is of course fantastic, and I believe that to be a very significant development. It is important to give 4+ cores enough bandwidth to actually work their magic.
  • DanNeely - Tuesday, April 26, 2011 - link

    2 channels of DDR3-1333 were fast enough to keep lga1156 CPUs from bottlenecking on memory IO in anything Intel considered normal use (vs a memory bench), while SB is faster than Lynnfield, the addition of official DDR3-1600 support should largely counterbalance it. The LGA2011 quad's extra bandwidth will be about as useful as the 3rd channel of the LGA1366 quad's were.
  • hansmuff - Tuesday, April 26, 2011 - link

    Interesting. Makes even more of a case for the platform to be expensive, if the "low end" quad doesn't outperform the 2600k significantly.
    I suppose that on a 6 core chip the quad channel (or at least a triple channel) would make sense though, no?
  • cactusdog - Wednesday, April 27, 2011 - link

    Exactly, things change when you have a 6 or 8 core CPU. Plus 1 dimm per channel is going to make it much easier to reach those high speeds and with lower latency and less voltage.
  • Concillian - Wednesday, April 27, 2011 - link

    Couple that with the "limited unlock" of that 4-core i7 they will put into the 2600K price range, and I'm not entirely sure of the value of the "low end" 2011.


    Intel is not in the business of having a "value low end" processor anymore. They are very much insisting you spend money for the performance at this point. That much was already clear with 1155 and shouldn't come as a surprise that they're doing it again with 2011. They have specifically engineered this generation to separate as much money as possible from enthusiasts.

    It appears to be working well for them so far, lots of people are buying more processor than their demographic would have in the past. They're spending more on motherboards too. Intel is raking in the benefit of having so little competition on the high end.
  • Michael REMY - Tuesday, April 26, 2011 - link

    i really wonder why the number of usb port increase at each chipset generaton !

    in 20 years of computer science, it hardware, i never knew a person need such ports !

    i'd prefer more memory cache instead more usb ports !

    my notice is valuable also for sata port number !

    hard drive capacity grows up so ! number of hard drive decrease in the same time !!!
  • DanNeely - Tuesday, April 26, 2011 - link

    USB2 controllers are tiny, even on the legacy processes many southbridges are built with. As legacy IO ports are going away they need to pile on more USB2 controllers to keep the IO panel mostly filled up and to still have enough headers for cases with 4 built in ports and a card reader in one of the drive bays.
  • Lonyo - Tuesday, April 26, 2011 - link

    You can use that many if you really want, e.g. external drive + keyboard + mouse + USB drive + phone cable + printer + USB wi-fi device + another external drive + another USB drive + card reader.
    That's 10 right there. This is an enthusiast board after all.

    The 14 SATA ports is a little excessive, although it might be useful for more enterprise oriented motherboards where you might want lots of drives.
  • DanNeely - Tuesday, April 26, 2011 - link

    As for the sata ports, Intel's trying to squeeze out the 3rd party controller most higher end mobos have added. With LGA2011 dropping the PCI bus entirely this might not just be a case of trying to make sure that the extra dollar goes to them, because the several other misc devices that used to be piled onto the PCI bus are now having to compete with everything else for the 8 PCIe lanes on the SB. WIth no native USB3 support I suspect most full ATX enthusiast boards are going to end up resorting to PCIe multiplexer chips, or a PCI-PCIe bridge in order to fit everything in, while still putting at least a 1x PCIe slot in every available spot on the board.
  • name99 - Tuesday, April 26, 2011 - link

    The SATA ports make no sense. The USB are used in many surprising ways.
    On Mac portables, for example, USB is used for at least the internal kbd and trackpad, the webcam, bluetooth, the IR port. With two or three user USB ports, that gets you to 7 or 8. On a desktop mac you have the same stuff and maybe 5 USB ports. (And that's not dumb --- I use all 5 USB ports on my Mac mini HTPC, for example).

    I may have left some stuff out --- not sure how audio (out or in) is handled these days, for example. Also 802.11g I think used wifi (can't remember), but obviously n does not.
  • name99 - Tuesday, April 26, 2011 - link

    Oh, yeah, also include at least an SD slot for modern macs.
  • Blaze-Senpai - Tuesday, April 26, 2011 - link

    Hopefully Intel doesn't just start tacking on a pin to the socket every year now... would turn into more of a mess than it is now.
  • Hrel - Tuesday, April 26, 2011 - link

    This is getting rediculous. I really wish I had the influence to organize a global boycott on Intel until they get with the freakin times and put a USB 3.0 controller on their chipsets. I'm not buying another motherboard until EVERY USB port on it is USB 3.0.

    Seriously, WTF Intel?! WTF! So help me if you guys don't knock this shit off I'm going pure AMD out of spite. 50% profit margin is PLENTY! Stick with ONE socket for at least a couple generations. Like LGA775. As soon as it's physically possible, support the newest technologies and don't expect your customers to pay an arm and a leg for them. EVERY SATA port should be 6MBPS, EVERY USB port should be 3.0. Motherboards should NOT have major issues months after release.

    Gah, F it. At this point CPU power with just about anything is "good enough". Unless things change drastically in a year or so I'm now ONLY buying AMD.
  • slyck - Tuesday, April 26, 2011 - link

    That's why I'm going bulldozer. My Core2Duo has been great, and Intel's chips have the best cores. But their market segmentation games are beyond ridiculous, are are their naming schemes and planned obsolesence. LGA2011 is going to be expensive, just to get the same options you have had with more affordable AMD chipsets for years now.
  • rahvin - Wednesday, April 27, 2011 - link

    I bought an ASUS board with 4 USB 3.0 ports. Two rear and an internal header, the MB comes with a 3.5" breakout box for the front. The thing is the header on the 3.0 port is HUGE. It easily occupies the space of 2 SATA ports. Just guessing without measuring I'd say it's 3 times the size of a USB2 header.

    You aren't going to see a lot of USB3 headers on a board, they aren't going to have space for them. The shear number of pins and traces required is going to prohibit it. My guess is that in the future the most you will ever see is 2 rear and a single header for 2 front mount USB3. Anything more than that and you will pay serious money for the traces. Couple in that USB3 has a max cable length of about 1 meter (3 feet) and I just don't see widespread replacement of USB2.

    Given that USB3 is only really needed for large data transfers which is pretty much hard drives and cameras (debatable) I think Intel's plan on this is going to be the winning path. That plan is for thunderbolt to be the path for USB3 with a breakout box. Believe me, I felt just like you did until I saw the header size, I can't believe it is going to be deployed in large numbers on any board because it would cost a fortune to take up that much real-estate on a board.
  • GullLars - Friday, April 29, 2011 - link

    Well, since USB3 supports USB2, having just 4 native USB3 ports (that is full bandwidth and power on all at the same time) would be preferable to 14 USB2 IMO. You could have 2 full BW ports and 1 split into 4 ports (hub style) at the IO panel = 6 ports, and 1 internal header breaking out to 2/4/6 physical ports for the front.
    For support of older cases, an additional internal header for 2/4 USB2 ports can make sense.
  • sicofante - Wednesday, April 27, 2011 - link

    Any idea if the SNB-E Xeons might support overclocking by being unlocked as the Core i7 parts?
  • flyvog6 - Tuesday, May 17, 2011 - link

    sells clothing,footwear,handbags,Sunglasses
    Online Store,Get Name Brand Fashion From 12USD Now!
    Our Website: ===== www voguecatch com ====
    Our main product list is as follows:
  • flyvog6 - Tuesday, May 17, 2011 - link


    sells clothing,footwear,handbags,Sunglasses
    Online Store,Get Name Brand Fashion From 12USD Now!
    Our Website: ===== www. flyingstyle org ====
    Our main product list is as follows:
  • aa0101bb - Wednesday, May 18, 2011 - link

    sells clothing,footwear,handbags,Sunglasses
    Online Store,Get Name Brand Fashion From 12USD Now!
    Our Website: ===== w w w voguecatch com ====
    Our main product list is as follows:
  • jinmengml - Sunday, May 22, 2011 - link


    Wedding is every person's life the most important moment in the wedding the most beautiful wedding dress is every woman's dream in mind. In our case, we provide many types of wedding dress

Log in

Don't have an account? Sign up now