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

 



Forgot your password?
typodupeerror
×
Intel Upgrades Hardware

Looking at Intel's New-ish Desktop Socket, LGA 1366 100

Slatterz writes "LGA 1366 is Intel's first new desktop socket in four years. It uses the same ZIF design as the familiar LGA 775 architecture, but it incorporates many more contacts. These big architectural changes are backed up by some less visible advances. Until now, Intel's quad-core processors have been constructed from two dual-core dies, but now Core i7 brings together four cores on a single die. It's also Intel's first processor design to use an L3 cache, shared between all four cores."
This discussion has been archived. No new comments can be posted.

Looking at Intel's New-ish Desktop Socket, LGA 1366

Comments Filter:
  • by sznupi ( 719324 ) on Monday June 01, 2009 @05:40AM (#28165823) Homepage

    ...but c'mon.

  • Uh, what? (Score:5, Funny)

    by pathological liar ( 659969 ) on Monday June 01, 2009 @05:44AM (#28165841)

    This might have been news 7 or 8 months ago when the chips were released.

    What's next? ATX Power Supplies Explained? "Plugs into any ordinary wall socket! Flick the switch and it turns on! Use it to power your computer! You'll see them turning up in shops any day now!"

    • by Norsefire ( 1494323 ) * on Monday June 01, 2009 @06:11AM (#28165981) Journal
      Right after they announce Tim Berners-Lee's new invention. I heard it's going to change the world.
      • I heard it's going to change the world.

        Make it "could change the world", pull out some key facts, misspell the title and you've got a story!
        kdawson

    • by LordKronos ( 470910 ) on Monday June 01, 2009 @06:31AM (#28166077)

      Flick the switch and it turns on!

      I've heard some of the newer designs are using pushbuttons instead of switches. I know....it sounds absurd, but apparently it's true.

      • Since the soft power pushbutton goes through the OS, you can be stuck there with a computer running and no way to turn it off but pull the power cable. Not having an actual switch on the power supply is frustrating.

        • Re: (Score:1, Funny)

          by Anonymous Coward
          Try pushing and holding the button for 3-5 seconds, JACKASS!
          • Depends on BIOS, JACKASS!

            ALWAYS have a physical power switch.
            ALWAYS.

            • Uhhh, no. In my 15-ish years as a PC tech, I've not once run into an ATX machine that didn't turn off when the power switch was held unless there was a malfunction in the power supply or the switch itself. Some machines take a few seconds longer to power off than others, but they *all* work.

              • We have a rack full of dell servers that let us disable this in the BIOS.

                If the thing hangs, you have to pull the plug.

            • Any decent ATX power supply includes a physical power switch to cut the juice. Also comes in handy when working in the PC - you can turn the switch off to kill the power but leave the cord plugged in so the chassis is grounded. I hate the cheap ones with no switch on the back.

              • by Matt_R ( 23461 )
                Here in the developed world, we can turn the power off at the wall socket.
                • I HATE switches at wall sockets.
                  So easy for some fool to bump it and take shit offline.

                  That's what a circuit breaker box is for.

    • by L4t3r4lu5 ( 1216702 ) on Monday June 01, 2009 @06:49AM (#28166187)
      I'm afraid ATX supplies won't be needed when the new wireless platform that never needs to be charged [penny-arcade.com] is released.
  • Metaphor (Score:4, Funny)

    by goldaryn ( 834427 ) on Monday June 01, 2009 @05:47AM (#28165851) Homepage
    "LGA 1366 is Intel's first new desktop socket in four years. It uses the same ZIF design as the familiar LGA 775 architecture, but it incorporates many more contacts

    Bigger number! Woo-hoo! With the POWER of MATHS I can tell you ...1366 over 775 = 76% better!

    Reminds me of when I upgraded my disposable razor from one with 3 blades to 4 FREAKING BLADES!!!11!1

    I get 33% more hot babes now.. FUCK YEAH

    (Yeah, yeah, 33% of 0....)
    • Re: (Score:2, Funny)

      by faragon ( 789704 )
      From 3 to 4 blades? Come on! [theonion.com]
    • Re: (Score:1, Offtopic)

      by TeknoHog ( 164938 )

      Reminds me of when I upgraded my disposable razor from one with 3 blades to 4 FREAKING BLADES!!!11!1

      I get 33% more hot babes now.. FUCK YEAH

      (Yeah, yeah, 33% of 0....)

      And that's why they call it "machturbo".

    • Re: (Score:1, Funny)

      by cripkd ( 709136 )
      4 blades? Why are you leaving in the stone age? Why don't you use a knife instead? We have upgraded to 6 (yes, SIX) blades like for... months!!! Best sell your flintstone car and buy an SUV. They will be the next big thing.
    • More contacts -> more data transfer at same clock rate. More blades on a razor -> not much difference. (Also re:title: "Metaphor"? Don't you mean "analogy"?)
      • More specifically, many of those extra pins are there for the on-die memory controller that Intel was just too cool for during the period when the A64s where kicking their asses. Now they have one as well, and it provides just as big a boost, for a rather nicer core.
    • Re: (Score:3, Funny)

      by L4t3r4lu5 ( 1216702 )
      I can shave so quickly with my new R^hAzor [shave.com] that the last four letters of the name accelerate faster than the first one!

      Whooosh!
    • I thought myself, "what the fuck, why am I paying 32 euro for a couple of razor blades!" and I bought the cheap generic brand 2 blade stuff.

      OUCH! Not saying it cut my face, it didn't but there really is a HUGE difference between the "quality" 5 blade razors and the cheap 2 blade kind. It is a smooth shave versus having the hair torn from your face.

      Crybaby you might say, or I might just have had a really bad 2 blade razor but still. I ain't going back again.

      So if you got a baby soft skin and a 5 o'clock s

      • You're better off with a two-blade Bic than a four-blade Trim. The biggest difference between good razor and shitty razor is the steel.

      • by Anonymous Coward
        That is funny. You paid 32 euros for a disposable razor, when about the same amount of money could have gotten you a quality straight razor. Once you've learned how to use a straight razor, you'll never want to use any other kind. Close shave, but unlike the (dear god are we actually up to five blades? That was a joke I saw in a cartoon years back) disposable razors, you won't have any ingrown hairs.
      • I thought myself, "what the fuck, why am I paying 32 euro for a couple of razor blades!" and I bought the cheap generic brand 2 blade stuff.

        OUCH! Not saying it cut my face, it didn't but there really is a HUGE difference between the "quality" 5 blade razors and the cheap 2 blade kind. It is a smooth shave versus having the hair torn from your face.

        I had the exact same experience when I tried to buy cheap generic 2-blade razors to replace my expensive name brand 2-blade razors. Because it's about the quali

      • by piojo ( 995934 )

        After using a 3-blade Gilette razor for a couple years, I tried the older 2-blade version and liked it a lot better. I can't really say why, but it seemed to give a cleaner shave and have less irritation. I consequently believe that all this "now with even more blades" nonsense is just marketing.

    • Bigger number! Woo-hoo! With the POWER of MATHS I can tell you ...1366 over 775 = 76% better!

      I'm certain the number is just the pin count which is a typical way to name sockets. More pins isn't necessarily better but since in this particular case the explosion in pin count is due to adding three DDR3 channels on-chip I'm willing to bet it is.

  • About time! (Score:1, Interesting)

    by Anonymous Coward

    About time they copied HyperTransport! If I remember correctly AMD had a leg up on Intel for a few quarters in the multi-core war because of HyperTransport.

  • by nxtw ( 866177 ) on Monday June 01, 2009 @06:15AM (#28165991)

    LGA 1366 is intended for servers, workstations, and high-end gaming PCs. LGA 1156 will be the mainstream
      desktop socket.

    What's the difference? IIRC, LGA 1366 has a tripe-channel memory controller and support for dual CPUs (via another QuickPath link). LGA 1156 has dual-channel memory support with support for only one CPU.

    The desktop CPU for LGA 1156 will be called Core i5.

    • Re: (Score:3, Funny)

      by Anonymous Coward

      LGA 1366 has a tripe-channel memory controller

      Can't be that bad, surely.

    • Re: (Score:2, Informative)

      by vyvepe ( 809573 )
      LGA 1156: Also, there is 16 PCIe links on the processor die directly and there is only 2 GB/s connection from CPU to P55 PCH. No QPI. More information is here: http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=3570&p=2 [anandtech.com]
    • by asliarun ( 636603 ) on Monday June 01, 2009 @06:59AM (#28166259)

      Please refer to the excellent Anandtech preview article on Lynnfield [anandtech.com] that will be the first family of CPUs to use LGA 1156. Lynnfield has uses a dual-channel DDR3 controller instead of using triple-channel integrated memory controller in its uncore like Nehalem does. However, the dual-channel controller should still provide enough bandwidth for most desktop apps (the Nehalem architecture is not bandwidth constrained at all, unlike all previous generation CPUs including Core2 that used massive L2 caches to offset the memory bandwidth bottleneck due to the FSB).

      However, the main difference between Lynnfield/LGA 1156 and LGA 1366 used in servers is the fact that it doesn't use QuickPath at all. Instead, it uses a combination of integrated PCIe 2.0 x16 controller (to talk to the graphics subsystem) and a (much slower) DMI controller to talk to everything else. Its an interesting alternative to QuickPath which is frankly expensive overkill for desktops anyway. The key advantage of the new socket will be significantly lower prices of motherboards and CPUs, which will allow Intel to provide some credible alternatives to AMD's current offerings that may be slower than Nehalem but are also much cheaper.

      • Re: (Score:2, Informative)

        by Anonymous Coward

        the Nehalem architecture is not bandwidth constrained at all

        Don't be silly. Total bandwidth to all 4 cores of an i7 920 on triple-channel DDR3/10333 is only 20GB/s. That's about 1.8 bytes per core per cycle. Given the existence of SSE instructions reading 16 bytes per cycle you're about 90% short of unconstrained bandwidth.

        Even with the best DDR3 money can buy you're still 80% short.

        I think you meant to say the Nehalem architecture is not bottlenecked at all.

        • Re: (Score:3, Interesting)

          by asliarun ( 636603 )

          Thanks for the correction. Yes, I meant to say that Nehalem is not bottlenecked.

      • just one x16 link and a slow dmi link is too small

        as 1 / 2 video cards cards will eat up the x16 link. Making sound / ide / sata / network / system and bios IO / other pci-e slots / firewire / usb all share the slow DMI link or force MB makers to put pci-e switches (driving costs up) one the x16 link.

        AMD wins aging all cpus use HT driving MB costs down as well letting the same MB run low and high end cpus.

    • by asc99c ( 938635 ) on Monday June 01, 2009 @07:06AM (#28166281)

      The socket for Core i7 equivalent Xeons is the same LGA-1366, but the standard Core i7 only has one QPI link, so you can't use them in dual CPU configs.

      I hadn't heard about LGA-1156, but I'm a bit suspicious whether it will really take off. I don't really understand now that Intel have launched LGA-1366 where is the room for a slightly lower spec socket. I wouldn't have thought a few extra pins in the socket is that expensive, and buying RAM in packs of 3 isn't that much of a problem - and it's optional anyway.

      By Q3 when LGA-1156 is due, Core i7 will be already heading down into the mainstream.

      • by nxtw ( 866177 )

        i5 systems will be much cheaper - turns out that they won't have QuickPath at all, but will have an integrated PCI-e controller instead. This should allow for cheaper/simpler system boards. I think this will lead to lower power use as well.

        LGA 1156 wil replace LGA 775 as the mainstream socket. And LGA 1366 won't take over as the mainstream desktop socket when all the CPUs are well over $100 and there are no chipsets with integrated graphics available.

        • by asc99c ( 938635 )

          Lack of integrated graphics is a good point. Anandtech's article linked to by asliarun above suggests that LGA1156 is well designed to run with integrated graphics, based on having an internal PCIe controller.

          But still, unless LGA1156 is really *much* simpler and cheaper to build, I would expect that the existing LGA775 / Core 2 Duo / Core 2 Quad platform would remain the cheapest option for quite a while. By the time you'd think about LGA1156, it seems LGA1366 would be the one to go for.

          • by Lvdata ( 1214190 )

            Not all I7's lack integrated graphics. I am running a http://www.newegg.com/Product/Product.aspx?Item=N82E16813182167 [newegg.com] in a low cost server application. It only goes to 1024x768 - not great with my monitor that does 1366x768 Wide Screen, but then most of the time I am using it with a remote desktop connection.

            • Usually "integrated graphics" refers to graphics on the northbridge, not a discrete 2D chip on the motherboard - that Supermicro board has a Matrox chip on there to handle 2D graphics.

    • >

      .... IIRC, LGA 1366 has a tripe-channel memory controller ....

      So ... it's optimized for WIndows Server?

    • Re: (Score:2, Insightful)

      by 313373_bot ( 766001 )

      For what it is worth, looks like artificial market segmentation = bad for upgraders. AMD is guilty of that, too.

      • by nxtw ( 866177 )

        No, it appears there are significant technical differences - such as the inclusion of QuickPath and triple channel on 1366 and an integrated PCI-e controller on 1156. I wouldn't want an inefficient desktop because I had to use a socket with an excessively high bandwidth link.

        • Re: (Score:3, Insightful)

          by 313373_bot ( 766001 )

          Point taken, perhaps "artificial" wasn't the best way to put it. But again, with few exceptions, the high-end technology of today becomes the mainstream of tomorrow.

  • by ricky-road-flats ( 770129 ) on Monday June 01, 2009 @06:37AM (#28166105) Homepage
    And I thought Australia was 12 hours ahead of us!
    • Re: (Score:3, Funny)

      by schwartzg ( 1089259 )
      It has nothing to do with timezones and everything to do with distance. Boats powered by magazine editors have only now reached Australia, but since we have the internet we learned of their arrival at the same time they arrived. It makes perfect sense.
  • Kittens (Score:1, Insightful)

    by Anonymous Coward
    Kittens are kinda awesome.
  • Pictures? (Score:3, Insightful)

    by AltGrendel ( 175092 ) <ag-slashdotNO@SPAMexit0.us> on Monday June 01, 2009 @06:47AM (#28166169) Homepage
    A picture of the socket instead of the chip would have been helpful.
  • I have noticed that motherboard prices for higher-end AMD Phenom systems seem to be a lot lower then MB prices for Intel I7 systems. AMD's high integration approach seems to be paying off there. The only real issue seems to be AMDs lack of support for larger 16G memory configurations in its desktop line.

    The whole-system price for a Phenom X4 system (using e.g. A Shuttle SN78SH7 as a base) is less then $600. Every year it seems I can buy a cheap off-the-shelf barebones system and completely replace severa

    • Phenom (and AM2 and above) have been around a lot longer. 1366 MB prices are starting to plummet now.
      • i wouldn't really say plummet. the cheapest i've seen is $180. down from $200 which it has sat at for a while. i got the MSI X58 Pro Openbox for $131 ...but then had to spend $48 on NB/SB coolers (2 x $24 for the HR-05 IFX; i probably could have gotten away with 1 + a cheaper SB cooler, but i was in a rush and didn't want to do even MORE searching and wait for it before powering on the machine)

        i THINK the X58M MSRP is supposed to be $160, but that hasn't trickled down, yet.

Isn't it interesting that the same people who laugh at science fiction listen to weather forecasts and economists? -- Kelvin Throop III

Working...