Follow Slashdot stories on Twitter

 


Forgot your password?
Close
typodupeerror
×

Intel Takes Quad Core To the Desktop 191

Posted by Zonk from the and-stay-there dept.
Rob writes to mention a Computer Business Review Online article about Intel's official launch of the Kentsfield chipset. Their Quad Core offering, Intel is claiming, is up to 80% faster than the dual-core Conroe released this past July. From the article: "Kentsfield, a 2.66GHz chip with a 1066MHz front-side bus, is more for computational-heavy usage, including digital content creation, engineering analysis, such as CAD, and actuarial and other financial applications. Steve Smith, director of operations for Intel digital enterprise group, claimed rendering is 58% faster for users building digital content creation systems, for video, photo editing or digital audio. In other words, Kentsfield is for high-end desktops or workstations only. For the average office worker who uses their PC for general productivity apps, such as communications and garden-variety computing, Smith recommended the Core 2 Duo from 'a price point and performance perspective.'"
This discussion has been archived. No new comments can be posted.

Intel Takes Quad Core To the Desktop More

Intel Takes Quad Core To the Desktop

Comments Filter:

  • Why downplay it? (Score:5, Insightful)

    by Salvance ( 1014001 ) * on Tuesday November 14, 2006 @09:48AM (#16836886) Homepage Journal
    "Core 2 Extreme quad-core QX6700" - There's a mouthful. It's funny that Intel is continually trying to downplay the importance of this chip for the average user. They say it's best for "more for computational-heavy usage, including digital content creation, engineering analysis, such as CAD" ... sounds like gamers would flock to this. Maybe they realize it's a rushed product (to beat AMD to the punch), and it will be in short supply?

    • Re: (Score:2, Insightful)

      by Anonymous Coward
      Or it may be that most games are not optimized for multiple cores. If they target gamers, only for gamers to discover that there is little improvement over their previous processor, then Intel's image with gamers would be damaged. However, some game companies such as Valve have recently started to embrace multi-core processors, but it will be awhile before new games are published that take advantage of those extra cores.

    • Re:Why downplay it? (Score:5, Insightful)

      by GauteL ( 29207 ) on Tuesday November 14, 2006 @10:20AM (#16837160)
      Currently the quad-core is pretty useless for gamers unless you like to run video encoding apps at the same time as you play your game.

      The reason is of course, that most games are barely optimised for dual cores, let alone four cores. It is not simple either as balancing several cores to get the most out of them requires a redesign of the game engine.

      It will be significant for future games, but you are better off buying a high-end dual core now and replacing it with quad-core later on.

      • Re:Why downplay it? (Score:5, Insightful)

        by oojah ( 113006 ) on Tuesday November 14, 2006 @10:28AM (#16837234) Homepage
        you are better off buying a high-end dual core now and replacing it with quad-core later on.

        Right. The best bit about quad core for the moment is that it should drive the dual core prices down.

        Cheers,

        Roger

      • WTF? (Score:5, Funny)

        by LibertineR ( 591918 ) on Tuesday November 14, 2006 @10:36AM (#16837366)
        "Currently the quad-core is pretty useless for gamers unless you like to run video encoding apps at the same time as you play your game."

        What? I thought EVERYONE used WinDVR to encode MPEG2 files of Battlestar Gallactica from their TIVO while playing F.E.A.R., and turn it into H262 for uploading as a a killer torrent while kicking but in Call of Duty 2? I suck the life out of an X2 4400 bitch, and I am NOT alone.

        We cant all have a life, so I NEED that chip!

        You insensitive clod!!!

      • I've gone over this before and eventually posted on Kerneltrap [kerneltrap.org], and even when Slashdot reported on it [slashdot.org] nobody seems to have cared.

        Games are more than pumping FPS.

        (Ad yes Valve's babble about a "new programming strategy" to use multi-core CPUs is bullshit; it's simply easier to use threads to handle each object in the physics engine and each AI, and that's immensely helpful when you don't want to worry about scheduling making your game choppy due to too many threads on one CPU)

    • Re: (Score:2)

      by mikael ( 484 )
      Perhaps the CPU manufacturers are desperately competing against the GPU manufacturers for developers of scientific applications? Nvidia just announced their 8800 series GPU's with support for BLAS [netlib.org], a foundation library for intensive engineering calculations.

      All the engineering, digital content creation, and gaming use similar algorithms to model/visualize water, fire and smoke. However, engineering does require high precision (64-bit floats) while animation and gaming can get away with lower precision (16-b
    • "more for computational-heavy usage, including digital content creation, engineering analysis, such as CAD"
      that's a polite way to say "it's for stuff that does a lot of work on a little data, because those 4 processors are clogged on the same data pipes", in other words "not most games"

    • Re: (Score:2)

      by mspohr ( 589790 )
      I think the GPU (graphics process units) people have a big advantage for gamers (and all digital content creation tasks). Most GPUs are 10-20 times as fast as CPUs for their specialized functions so a "58%" speed improvement with the quad core CPU doesn't mean anything.
    • Why downplay it? Because it costs a grand. No matter how much benefit it might be to the "average user," the price of the chip and the systems that use it will drive people away.

      Enthusiasts want Kentsfield and are willing to pay for it. Average people may want it, but are only willing to pay for Celeron.

  • Office Apps (Score:4, Funny)

    by Ginnungagap42 ( 817075 ) on Tuesday November 14, 2006 @09:52AM (#16836924)
    So how does Minesweeper run on it?
  • All I want to know is if QuadCore will make my World of Warcraft Elite battle load and display 4-times faster?

  • Will it be supported... (Score:3, Insightful)

    by AppreticeGuru ( 1024775 ) on Tuesday November 14, 2006 @09:55AM (#16836952)
    4 cores is great and all, but I know they are still working on support for games such as many Steam offerings with only 2 cores in terms of multi-threading, so I'd have to imagine that game support to really take advantage of a 4-core system would be a long way away. I was still psyched about the low voltage powerhouses for laptops, and I'm wondering how much extra heat 4 cores are going to put out as well. How many apps are really geared to take advantage of 4 cores atm, really?
    • I have a number-crunching program which I rewrote multi-threaded. It is simple enough -- I need 100,000 calculations of the same thing only with different parameters each time, so run them in parallel, two at a time for two threads, four at a time for four threads. I still have the bookkeeping of saving the results of each calculation, and I have streamlined the calculation enough that the bookkeeping part is a significant piece of the overall time.

      I am about 50 percent faster on two cores -- I am guess

      • Are you sure that the bookkeeping is the limiting factor in your program? It could also be that you simply run into a memory bandwidth bottleneck. When you have datasets that do not fit in the L2 cache, main memory simply cannot keep up with the amount of data that the CPUs chew through. With multiple cores that all share the same memory link, this problem is going to become only worse.

        One more argument in favor of the AMD Opteron architecture, even though for single-threaded applications it is currently sl
        • Well, the thing is written Java and it JNI's down to a C++ module for the numerical computation. I try to maintain as much locality of data for each task as possible by allocating and passing objects for the threads to work on. JNI has a lot of overhead, but the JNI call to the C++ module is done on the parallel threads.

          I could test and profile and figure out the bottlenecks but I don't have the time to play with it right now and not having a 4-core computer, I won't see much benefit. Part of the probl

  • Missing a marketing opportunity. ... now with Intel Foursome!
  • I know CPU power is a big factor in performance, but c'mon.. What about extending the rest of the motherboard? I bet things would run faster in dual/quad core mode if there were dual buses so that bottlenecks are reduced to peripherals and memory.

    • Re: (Score:2)

      by larkost ( 79011 )
      That is the whole point of having multiprocessor systems.

    • Re: (Score:2)

      by 10Ghz ( 453478 )
      the FSB's have been getting faster, the expansion-buses (PCI, PCI-E etc.) have been getting faster, memory has been getting faster, and they reside on wider buses.

      If you complaint is the FSB; the AMD has something better for you. So what are you looking for here really?

      • Re:What about other parts of the computer? (Score:5, Interesting)

        by s31523 ( 926314 ) on Tuesday November 14, 2006 @10:47AM (#16837526)
        Faster, good... Wider, good... But why not parallel with dual-DMA? Right now, it seems you could have 10 cores, but if all the threads running on each core have to contend for 1 bus, it doesn't matter how fast the bus is. I want each core to be able to access its own memory so it is not blocked by the other core's if it is accessing memory. I want one core to be able to access my NIC while the other accesses the hard drive and the other access the video card. All this requires some sort of parallel bus setup. It is my understanding we have not done this sort of architecture yet, but if we keep increasing the number of processor cores, this would seem to be the next step. BUT, I am not a hardware guy. I am a software guy, and expect it all to just work! :)

        • Re:What about other parts of the computer? (Score:4, Informative)

          by 10Ghz ( 453478 ) on Tuesday November 14, 2006 @11:16AM (#16837876)
          "I want each core to be able to access its own memory so it is not blocked by the other core's if it is accessing memory."

          Say hello to AMD, HyperTransport and integrated memory-controllers. Each CPU has it's own bank of RAM, and Each CPU is directly (well, 8-socket system needs one intermediate jump) connected to the other CPU's, and they can access the RAM connected to the other CPU's as well. So if you have dual-socket system, each socket has it's own RAM-bank, with 128bit bus between the CPU and the RAM, and the CPU can access the RAM attached to the other CPU as well. So as the number of CPU's goes up, the memory-bandwidth goes up as well.

          This tech has been used since 2003 in the AMD's x86-64 CPU's. In the future AMD will have systems where you can plug co-processors and vid-cards to HyperTransport-sockets, alloweing them to directly communicate with the CPU's.

          • Re: (Score:2)

            by s31523 ( 926314 )
            But... Does it come with a Hemi? :)

            Nice! I gotta get me one of those!

          • Re: (Score:3, Informative)

            by default luser ( 529332 )
            It's not quite that good in it's current incarnation. Right now, high-end (4 and 8-way) Opteron chips have only three HyperTransport links.

            Try connecting 4 of these chips together using only 3 HyperTransport links per core, with a single-hop memory latency, and allow for one link to external I/O. Can't be done. There are two hops required for the core that handles I/O, which is not a good thing when you consider how important I/O links are in a server.

            Try connecting 8 sockets using only 3 HyperTransport

            • Re: (Score:2)

              by 10Ghz ( 453478 )
              It's not quite that good in it's current incarnation. Right now, high-end (4 and 8-way) Opteron chips have only three HyperTransport links.


              Well, I DID say that 8-socket configuration need one additional hop...

            • Re: (Score:2)

              by LWATCDR ( 28044 )
              Okay four sockets with dual core cpus means an 8 way 64 bit server. That is some big iron no matter how you look it.
              I will take two please and load them up with raid of 15k rpm drives and a quad SLI please.

        • Re: (Score:2)

          by kabocox ( 199019 )
          Faster, good... Wider, good... But why not parallel with dual-DMA? Right now, it seems you could have 10 cores, but if all the threads running on each core have to contend for 1 bus, it doesn't matter how fast the bus is. I want each core to be able to access its own memory so it is not blocked by the other core's if it is accessing memory. I want one core to be able to access my NIC while the other accesses the hard drive and the other access the video card. All this requires some sort of parallel bus setu

    • Re: (Score:2)

      by sbryant ( 93075 )

      Damn right!

      I'm using a dual-core (AMD) right now, and the machine has a lot going on on it, but the bottleneck more often than not is not the CPU or even the memory - it's the disk! I have 2 cores, 4GB RAM, but only one HD (SATA), and it's definitely noticeable. I don't think an Intel Quad core would help significantly, even though I've got plenty of processes to spread across the cores.

      I'm wondering if a hardware RAID solution would make things better, but I don't have first-hand experience with that

      • RAID5 won't help. It's good at reads, but generally sucks for writes and rebuild times are often horrid. It's major claim to fame is net capacity.

        If you want increased throughput (and maybe slightly faster seek times) go with 2 or 3 disk RAID0. If you want peace of mind, go with a 4 or 6 disk RAID10 setup which will be 2x to 3x faster (throughput) then a single disk or a single RAID1 set. Net capacity is the same as RAID1, half the gross capacity.

        I have a few RAID10 arrays now and I keep wanting to

  • Not native Quad core (Score:3, Informative)

    by kid_oliva ( 899189 ) on Tuesday November 14, 2006 @10:01AM (#16836992) Homepage
    From what I've read about Intel's quad-core; it is not native like AMD's will be. They are basically are going to have two dual core's and they will communicate via FSB. That sucks compare to AMD's offering which will be native.
    http://xstremehardware.co.uk/index.php?option=com_ frontpage&Itemid=1&limit=10&limitstart=20 [xstremehardware.co.uk]

    • Re: (Score:3, Insightful)

      by k_187 ( 61692 )
      yes, but communication between the 2 cores in each of the sets will be faster than any of AMD's cores. My guess is that it'll be a wash. The other (in my opinion more important) thing, is that INtel is shipping now, while AMD is about a year away. By then I believe that INtel will have a quad-core on die chip out. Either way, more FPS!
      • I doubt intercore communication in the Intel design is any better than the Barcelona (it's impossible to say this far out in any event). What is known (I have worked with some of the quad core Intel stuff) is that the two dies produce a higher load on the FSB and require the FSB be clocked down from the equivalent dual-core model. This means that AMDs remaining advantage over Intel's offerings is made more drastic (aggregate memory performance, particularly in multi-socket configurations). I.e. an Intel

        • Re: (Score:2, Insightful)

          by dfghjk ( 711126 )
          When comparing quadcore approaches, aggragate memory performance of multisocket, multi-memory controller designs is irrelevant. No doubt the AMD approach scales better but that's not important to the argument. When AMD announces a single die processor with multiple integrated memory controllers then it matters. Offsetting AMD's memory throughput advantages are Intel's much larger caches. It's a complicated subject.

          Intel's approach gets quadcore to market far faster, and once AMD can deliver quadcore on

          • Re: (Score:2)

            by Junta ( 36770 )
            "When comparing quadcore approaches, aggragate memory performance of multisocket, multi-memory controller designs is irrelevant."

            Not necessarily, if the quadcore hinders the clock of the FSB compared to what it could be, it's highly relevant.

            "When AMD announces a single die processor with multiple integrated memory controllers then it matters."

            Barcelona is that.

            "Offsetting AMD's memory throughput advantages are Intel's much larger caches."

            True, it brings more problems into the space of operating entirely in

            • Re: (Score:2)

              by dfghjk ( 711126 )
              "if the quadcore hinders the clock of the FSB compared to what it could be, it's highly relevant."

              It could be, but this isn't hypothetical. We know how much the effect is (and it's not dramatic). Meanwhile, not all cores in the AMD design have the same pathway to memory.

              "Barcelona is that."

              When that arrives it will be part of the discussion. Until then it is not competition for current products.

              "The exception being that in the short term we see Intel clocking down FSB for quad-core."

              They also clock down

        • Re: (Score:2)

          by k_187 ( 61692 )
          It actually is, as each pair of cores share their L2 cache. So between each pair its better, but not between the 2 pairs (as this is covered by the FSB) nor between the whole and the system. WE'll see if this makes up for the FSB problem (probably not)

    • Re: (Score:2, Informative)

      by OrangePeril ( 739827 )
      True, it is not native quad core. However AMD's first venture into quad-core will not be native either. In an effort to catch up to Intel, they will also be releasing a quad-core processor thats "taped together" as Intel's is.

      I recently met with an Intel rep and they are very much pushing their new core architecture. Quad-core this year, Octo-core next.. Core count is the next clock speed. However one of it matters until the software manufacturers can take advantage of it, and very few server applica

    • Re: (Score:3, Informative)

      by dfghjk ( 711126 )
      In the timeframe Intel offers this, AMD will have no quadcore part at all. Considering that, it's clear that AMD sucks, not Intel. Later on, Intel's "native" version (Yorkfield, discussed in your link) will have cache improvements and a bump in FSB speed. All things considered, the dual die part doesn't look like it sucks at all (except for AMD).

      There are three sides to this: Intel's, AMD's, and the truth.

  • Funny thing happened on the way to IT Support... (Score:3, Interesting)

    by topham ( 32406 ) on Tuesday November 14, 2006 @10:08AM (#16837064) Homepage

    We were issued laptops at the start of the project. Typical laptop is a Thinkpad T42p. They average somewhere between 1.6Ghz and 1.8Ghz.

    Some people were complaining about performance (java is a hog, and they were using stuff that makes java look 'light'). so they requested new machines.

    They were issued Core 2 Duo systems that are 1.8Ghz, with 2 Gigs of ram. This machines are nice. They guy from IT Support comes up to replace the system and starts saying that he doesn't know why we would upgrade to the desktops, they are the same speed as the laptops.

    Ok, I expect that from some guy off the street, but IT Support?

    (Note: For this work there is a significant speed difference, it is obvious, and almost immediate.)
    Never mind the differences between a single core from a Core 2 Duo, and the core used in a Thinkpad anyway...

  • I have one of these babies (Score:4, Funny)

    by Nichotin ( 794369 ) on Tuesday November 14, 2006 @10:12AM (#16837090)
    .. since I am a journalist for a computer rag. Anyway, I would say it is a waste of money for most people at this time. Applications can barely use two cores properly, and games are still not as SMP aware as they should. On the other hand, if you run gentoo, THIS CPU IS KILLER :)

    • Re: (Score:3, Insightful)

      by mgblst ( 80109 )
      You know, it doesn't take being a computer journalist to realise that any chip released in the last 3 years is a waste of money for most people. Most people mainly use the computer for broswing the net, and despity Intels previous claims, a faster processor won't make any difference. And despite adverts on UK tv reporting that with the new dual cores, you can read email and listen to music, you don't need a 4 core or 2 core to do any of that.

      The whole thing is a joke, for most people. Like cars that go 1000

      • Re: (Score:2)

        by laffer1 ( 701823 )
        It depends if people are using flash. Try running a flash 9 intensive site on an old PC or Mac. It will not keep up. To some degree, you do need a faster processor to handle flash and the new "High Def" video codecs coming out. It entirely depends on what you use the net for, but it is important to some people.

        I noticed a difference upgrading from a Dual Xeon 2.0Ghz to a Dual Core Pentium D 805 (2.66Ghz) with quicktime streams for instance. Aside from my poor choice in video card (Geforce 7300), my new

      • Re: (Score:2)

        by r3m0t ( 626466 )
        "adverts on UK tv reporting that with the new dual cores, you can read email and listen to music"

        I complained to the Advertising Standards Agency about one of those adverts which was made by PC World. I'm still sour about it. I quote from the letter:

        "We did not consider that the advertisement implied that dual core processors were the only type of processor that could multi-task, or that they improved internet connectivity or performance. [the sales guy said something like 'playing a game while downloading

    • Re:I have one of these babies (Score:5, Interesting)

      by pla ( 258480 ) on Tuesday November 14, 2006 @10:33AM (#16837326) Journal
      since I am a journalist for a computer rag.

      I will say "lucky bastard", but that also explains your follow-up comment:



      Applications can barely use two cores properly, and games are still not as SMP aware as they should.

      Although apps and games have started to improve their multithreading, you don't get multi-core for single-app performance. You get it so you can play a modern FPS at the same time you have DVD Shrink backing up a movie for you, with little to no slowdown to either. With a quad core, you can add in two more CPU sucking tasks, again with little to no slowdown (though currently, memory needs to catch up to task of dealing with more cores).

      Six(ish) years ago, I got my first dual CPU machine. Almost nothing except the OS itself ran multithreaded at that time. And the improved performance of the machine just blew me away - Only last year did I eventually decommission that ancient dual CPU box because modern single-CPU speeds had passed it (and I still would have held out, except for the knowledge that I could do an in-place upgrade to a dual-core CPU whenever I wanted to).

      So you may not see the point of multi-cores, when your favorite game won't run any faster on four than on one. But that doesn't even come close to meaning that "most" people won't benefit. Quite the opposite, I'd say that only hard-core gamers wouldn't benefit. Everyone else will feel the improved responsiveness the first time they touch a multi-core box.

      • Re:I have one of these babies (Score:4, Insightful)

        by MikeBabcock ( 65886 ) <mtb-slashdot@mikebabcock.ca> on Tuesday November 14, 2006 @01:17PM (#16839806) Homepage Journal
        It never ceases to amaze me how many people don't realize that the average long-running PC has a lot of background processes with sporadic activity levels that ruin the performance of your desktop. Having something like BitTorrent running in its own core makes all the difference in the world too.

        MSN, BitTorrent, an MP3 player and a web browser all running at once (on top of background services) on a single-core system leads to a lot of task switching that is entirely unnecessary in a multi-core environment. And while throughput may not increase 4x, responsiveness will be very much improved.
        • You say a silly things with such seriousness. I commend you. So each of your programs should be on their own dedicated core? It almost seems like the background tasks you vaguely suggest are mostly euphemisms for software that makes up for Window's shortcommings, like AV scans and such.

  • Soon (Score:3, Insightful)

    by Mateorabi ( 108522 ) on Tuesday November 14, 2006 @10:24AM (#16837192) Homepage
    Soon the number of cores in my desktop machine will surpass the number of blades in my shaving razor.


    But seriously, as it gets harder and harder to make larger CPUs run faster the trend is going to be more, smaller processors per die. Each core is by itself slower than a huge monolithic one, but the sum is greater thanks to non-linear scaling. The trick is getting software to efficiently utilize them all.

    • Soon the number of cores in my desktop machine will surpass the number of blades in my shaving razor.

      Mateorabi's law, anyone?

  • Hype (Score:4, Funny)

    by h2g2bob ( 948006 ) on Tuesday November 14, 2006 @10:25AM (#16837196) Homepage
    It's going like razorblades - the razorblade companies just try to outdo each other on how many blades that can be placed in a single razor. At this rate, expect as many processors as you can physically fit on, plus an extra processor for those tricky, hard to reach programs.
    • What is funny to me is that the list of reasons for this chip: "Engineering, CAD, high-end workstations" is the exact same list that Motorola had as target users for the 8 Mhz 68000, back then used in the Mentor Graphics Workstations, and which now is labelled a Micro-controller [wikipedia.org] for embedded applications.
      All that matters to me are the issues:
      • Cost per gigaflop
      • watts per gigaflop
      • availability
      • memory bus speed and width

      And I will decide the applications that I will use it for!
      --jeffk++

  • Benchmarks! (Score:3, Informative)

    by Ironsides ( 739422 ) on Tuesday November 14, 2006 @10:39AM (#16837400) Homepage Journal
    Here's one from Toms Hardware. [tomshardware.com]

    Intel's right. On games it doesn't do any better. On video though, well, lets just say I know some architecture majors who would have loved these in their lab several years ago, when 1 frame took 10 minutes to render. And they had 300 frame videos to do.

  • Yeah, Right (Score:4, Insightful)

    by Nom du Keyboard ( 633989 ) on Tuesday November 14, 2006 @12:13PM (#16838676)
    Their Quad Core offering, Intel is claiming, is up to 80% faster than the dual-core Conroe released this past July.

    Yeah, that much faster on carefully selected software. And slower on some single thread applications that rely most of all on clock-speed and uncontested memory bus access.

    Would be nice for once to have headlines read something more honest like:

    Speed improvements range from -20% for 50% of your software, up to +80% for 10% of your software.

    There could even be a nice graph of how much software is improved (or degraded) at each 5% bin of performance. Otherwise it's no more honest than saying that your new Ferrari is capable of speeds up to 220mph, without mentioning that this can only be utilized during .01% of your driving.

  • "...digital content creation systems, for video, photo editing or digital audio. In other words, Kentsfield is for high-end desktops or workstations only"

    That sounds like a Mac user to me. The Article says "New systems boasting the CPU set to be announced today", and it's a Tuesday, the traditional launch day for new Macs. Do I see a pattern emerging here?
  • Kentsfield is a processor, not a chipset as the parent states. According to other articles, the Kentsfield processor will work with Intel's 965 and 975 chipset-based boards that are running Core2Duo's, providing a super-sweet upgrade path.

  • In 2013... (Score:3, Funny)

    by dorianh49 ( 988940 ) on Tuesday November 14, 2006 @02:35PM (#16841118)
    In 2013, Intel will look back and say, "Four cores and seven years ago, our engineers brought forth on this continent a new microarchitecture, conceived in Santa Clara, and dedicated to the proposition that all men's wallets are created equal."
  • We've been waiting on these things to replace our "old" dual-core hyperthreading Xeons. They are pretty powerful, but spam is getting insane these days and when you're scanning a million messages a day through each mx server, then lots of cores will get a good workout, believe me.

Slashdot Top Deals

Get hold of portable property. -- Charles Dickens, "Great Expectations"

Close