Slashdot Mirror

← Back to Stories (view on slashdot.org)

AMD Finally Unveils Barcelona Chip

Posted by ryuzaki0 on from the well-isn't-that-special dept.

Justin Oblehelm writes "AMD has finally unveiled its first set of quad-core processors, three months after its original launch date due to its "complicated" design. Barcelona comes in three categories: high-performance, standard-performance and energy-efficient server models, but only the standard (up to 2.0 GHz) and energy-efficient (up to 1.9 GHz) categories will be available at launch. The high-performance Opterons, together with higher frequencies of the standard and energy-efficient chips, are expected in the out in the fourth quarter of this year. But it's far from clear that this is the product that will help right AMD's ship."

118 comments

  1. how well will it overclock? by Anonymous Coward · · Score: 2, Interesting

    Since it's essentially the same tech since their X2 design?

    I get 2.7ghz out of a 2.0ghz rated X2 (on air).

    Once again they have beaten Intel's prices by at least $100 so we all win.

    1. Re:how well will it overclock? by ZachPruckowski · · Score: 3, Informative

      Barcelona is a different architecture from K8 (the architecture of the current X2s). It's overclocking performance is currently unknown. Just as Intel's overclocking potential improved as it went from Pentium -> Core 2 Duo, Barcelona may increase or decrease AMD's overclocking potential.

    2. Re:how well will it overclock? by Dr.+Eggman · · Score: 1

      I'm not sure they beat their price, the article lists the 2.0Ghz model at US$389; I got my Core 2 Quad at $300, and they're cheaper than that too now. Unless your thinking of some bulk sale to retailer's price I don't know of.

      --
      Demented But Determined.
    3. Re:how well will it overclock? by jimstapleton · · Score: 2, Insightful

      Core2 Quad = Desktop

      They are talking about server chips, which typically are more expensive than desktop chips.

      --
      34486853790
      Connection too slow for X forwarding? Try "ssh -CX user@host"
    4. Re:how well will it overclock? by h4rm0ny · · Score: 1


      Why? What's the difference? Honest question - isn't speed just speed?

      --

      Aide-toi, le Ciel t'aidera - Jeanne D'Arc.
    5. Re:how well will it overclock? by cHiphead · · Score: 2, Funny

      No cuz when you call it a 'server chip' its magically more super and fancy.

      Kinda like an Associates Degree vs. a High School Diploma.

      Cheers. ;)

      --

      This is my sig. There are many like it, but this one is mine.
    6. Re:how well will it overclock? by Anonymous Coward · · Score: 0

      The desktop chips aren't SMP capable.

    7. Re:how well will it overclock? by Fulcrum+of+Evil · · Score: 1

      Perhaps you were looking at the HE versions - same speed, less power = lower power bills.

      --
      "We returned the General to El Salvador, or maybe Guatemala, it's difficult to tell from 10,000 feet"
    8. Re:how well will it overclock? by jimstapleton · · Score: 1

      A lot more QA mostly. Also, if I remember correctly, server chips tend to be tweaked more for multithreading and integer performance.

      --
      34486853790
      Connection too slow for X forwarding? Try "ssh -CX user@host"
  2. Try the Anandtech review? by Anonymous Coward · · Score: 0

    But it's far from clear that this is the product that will help right AMD's ship."


    OK, it's not the revolution Hector suggests, but if you look at the review on Anandtech, it looks like AMD is catching up, and leading in a few areas. It'll probably look quite nice when the 2.5Ghz CPUs come out in Q4, and AMD will be a real competator again.

    If only they hadn't bought ATi, I might consider waiting and getting an AMD instead...
  3. Benchmarks by eebra82 · · Score: 5, Informative

    Here's some benchmarking done by Anandtech.

    And a performance preview for Barcelona desktop as well.

    --
    Full Tilt
    1. Re:Benchmarks by ceeam · · Score: 3, Insightful

      Most of benchmarks are on 32-bit code. Can we at least start considering that as "legacy" and use AMD64 when performance really matters?

    2. Re:Benchmarks by 644bd346996 · · Score: 4, Insightful

      Since Barcelona is one of the bigger architectural changes from AMD in the past few years, the 32-bit benchmarks are relevant because they are good predictors of what's to come for the entire product line, including the desktop processors, where 32-bit code dominates. Also, if they used exclusively 64-bit code, they would be accused of using unrealistic benchmarks to highlight the fact that AMD has better 64-bit performance than Intel.

    3. Re:Benchmarks by White+Flame · · Score: 1
      Anandtech:

      However, this will only occur if AMD's newest quad-core is able to outperform the Intel alternatives clock for clock by a decent margin.

      I really dislike this whole "tuner" mentality from most reviewers. This is a server chip, so not just clock for clock, but also dollar for dollar, and watt for watt will be big issues. Plus, Intel still generally releases larger caches, so that weighs in.

    4. Re:Benchmarks by evilbessie · · Score: 4, Insightful

      It could be argued, however, that these are server and workstation chips and so would be expected to perform mainly 64bit tasks to get the full use out of the performance. So 64bit benchmarks would make more sense. Now when the Phenom chips are out then 32 and 64 bits would be both useful as over the next few years most software will convert to 64bit and drop 32bit.

    5. Re:Benchmarks by AuMatar · · Score: 1

      Not anytime soon. Most code is still compiled with optimizations for P2 or P3. People won't start assuming 64 bit (unless the app needs it to run) for most of a decade.

      --
      I still have more fans than freaks. WTF is wrong with you people?
    6. Re:Benchmarks by testerus · · Score: 1

      Heise benchmarked with 64bit-Linux.

    7. Re:Benchmarks by testerus · · Score: 1

      Heise benchmarked with 64bit-Linux.

    8. Re:Benchmarks by Anonymous Coward · · Score: 0

      How is it fair to use exclusively 32-bit code, when the AMD CPUs perform that much better in 64-bit mode? Surely using 32-bit code only gives Intel an unfair advantage?

      I'd think that running both 64- and 32-bit benchmarks, and using the best result from the two, would be the best way.

    9. Re:Benchmarks by clayne · · Score: 0

      Larger caches to combat the lacklustre memory architecture, right? Circular.

  4. Re:question by everphilski · · Score: 0, Offtopic

    Why can't you have the fish part on top, and the lady part on the bottom? -Fry

  5. "Full generation behind"? by Stentapp · · Score: 2, Informative

    "The delay puts the chip maker a full generation behind its archrival in terms of chip manufacturing processes. Intel's quad-core processor, which was launched in November last year, melds two of its duo-core processors into a single package."

    Heh, shouldn't that be "full generation ahead" since AMD manages to put four cores on a single die?

    1. Re:"Full generation behind"? by Anonymous Coward · · Score: 1, Informative

      "The delay puts the chip maker a full generation behind its archrival in terms of chip manufacturing processes.

      Emphasis mine. Reading comprehension 101: Read the whole sentence. AMD is at 65nm, Intel is at 45nm, just as when AMD was at 90nm, Intel hit 65nm. This qualifies them as being "a generation behind" in chip making processes.

      Whether or not their architecture or their core design is better is completely irrelevant to that sentence (but relevant to the next, which is why it's so odd they'd put those two sentences together in the first place; even AMD admits if they could go back in time they'd do an MCM).

    2. Re:"Full generation behind"? by Mr+Z · · Score: 4, Insightful

      This is a direct reference to 65nm vs. 45nm geometry. If AMD brings their quad core to a 45nm process, that should help yield, power and performance. If nothing else, it puts them on a level playing field with Intel (who already have product at 45nm) so that it's down to "design vs. design." Being stuck one silicon technology generation back, they need to resort to other tricks to "keep up."

      In other words, to be at overall performance parity with Intel, they have to have a more advanced design in 65nm to keep up with Intel's 45nm work.

      Another thing worth noting: By being 1 generation back, the quad core setup is a double whammy. The die area of a given chip roughly halves with each technology node. Not only is AMD putting twice as much on one chip, it's also making chips that are twice the size per transistor. (Remember, to double square area, you only increase your linear feature size by sqrt(2). 65/45 = 1.444... which is about sqrt(2).) Each additional sq mm of die area causes greater yield loss than the one before it (driven by defect density in the source silicon). Doubling die size has a huge impact on yield. So, AMD will potentially suffer significantly higher yield loss, and correspondingly higher costs. Even if it can keep its ASP (average selling price) up, the profit margins will suck.

      It'll be interesting to see if AMD can quickly shrink this design to 45nm and get closer to parity. The benefits of the quad core design probably become much more apparent at 45nm.

      --Joe
      --
      Program Intellivision!
    3. Re:"Full generation behind"? by Anonymous Coward · · Score: 0

      K10 is only an optimisation rework of K8, but the next generation of their chips, designed from scratch for power efficiency and to be integrated with GPU's, might shake Intel's pillars (as they don't actually have any existing 3D technology yet that is comparable to ATI chips).

      We'll see, maybe Intel will be forced to even buy (or merge with) Nvidia, but then they'd be at least 1 year late in the the GPGPU war (unless of course they ARE preparing something powerful, although it usually takes several iterations to produce efficient 3D chip, same for drivers).

    4. Re:"Full generation behind"? by struberg · · Score: 3, Informative

      Intel and AMD are using different production technologies for their dies. For what i know, AMD is using IBMs SOI (Silicon On Insulator) which has much less drain current and therefor is much better at the same size. But it seems also more complicated to shrink this technology to 45nm.

    5. Re:"Full generation behind"? by Mr+Z · · Score: 3, Insightful

      That could help with leakage power, but that doesn't address the yield and cost issues at all.

      --
      Program Intellivision!
    6. Re:"Full generation behind"? by Gr8Apes · · Score: 0

      AMD is at 65nm, Intel is at 45nm, just as when AMD was at 90nm, Intel hit 65nm. This qualifies them as being "a generation behind" in chip making processes. Please show me an Intel 45nm chip.

      Thank you.
      --
      The cesspool just got a check and balance.
    7. Re:"Full generation behind"? by asliarun · · Score: 2, Interesting

      The die area of a given chip roughly halves with each technology node. This is not entirely true. Although I agree overall with what you're saying, core logic transistors scale much worse than cache as the manufacturing process decreases in size. I'm not sure if AMD factors this process disadvantage into their chip design, but it is an interesting design choice that they choose to stuff their chip real estate with logic transistors instead of cache. I'm sure that I'm oversimplifying, but I have a gut feeling that they possibly might be choosing to use less cache and more logic precisely because they know they will always be a process node behind Intel, and at least this way, their process disadvantage is somewhat compensated.

      Interestingly enough, Intel has traditionally adopted the exact opposite chip design strategy. IMHO, Intel's design ethos is first driven by manufacturing, and only secondly by pure design. Of course, they have every right to do so as they've consistently led the industry in process and manufacturing technology. However, this sometimes teeters into arrogance, and they have tended to fix a shoddy design by throwing cache at it, and/or relying on a die shrink (which also fabulously shrinks cache!).

      This process/cache luxury, compounded by bureaucracy can tend to make Intel come up with conservative designs. I'm not even going to talk about P4 or Prescott as it has been beaten to death and beyond. However, Justin Rattner recently hinted the same thing as well when he encouraged his research teams to come up with bolder and even impractical designs, and not start thinking about commercial viability so early on in the research/design stage. AMD, OTOH, simply cannot afford this luxure as they're usually getting whipped by Intel manufacturing AND by Intel marketing muscle, and are usually in a "do or die" mode. This usually makes them come up with riskier or bolder design.

      Having said that, Core2 is a superb architecture, and in my opinion, will be neck-to-neck with even Barcelona (win some benchmarks, lose some benchmarks). Its only in the server space that AMD will have two distinct advantages: Hypertransport for scalability, and DDR2 instead of FBD for power consumption. Sigh... if only Intel had not scrapped Whitefield. I guess it would have released by now.. and that too with a native quadcore design and CSI. Look at the Tigerton hack-job for a contrast... pathetic (but of course, easier to manufacture).
    8. Re:"Full generation behind"? by fitten · · Score: 1

      Heh, shouldn't that be "full generation ahead" since AMD manages to put four cores on a single die?


      No... AMD's arrogance costs them dearly. Intel has superior fab/process technology and could build monolithic quad-core but it is more expensive than MCM because of decreased yield in monolithic quad-core per wafer.

      AMD already has a decent infrastructure to support MCM quad-core very well but refuse to use it to increase their yields. Only arrogance and pride keep AMD from releasing MCM parts, which would significantly increase their yield (and therefore money) and only compromise performance a small amount. As they say... pride goes before a fall.
    9. Re:"Full generation behind"? by Wavicle · · Score: 1

      unless of course they ARE preparing something powerful

      Yes, they are.

      --
      Education is a better safeguard of liberty than a standing army.
      Edward Everett (1794 - 1865)
    10. Re:"Full generation behind"? by Frumious+Wombat · · Score: 2, Interesting

      Don't knock "easier to manufacture". The Cray3 and many other interesting designs failed because yields of some critical part never reached commercial viability. My first opteron servers (right out of the gate from a major vendor) had several failures, all due to the onboard memory controller frying. A little slower but fewer defects results in fewer recalls and less bad press.

      --
      the more accurate the calculations became, the more the concepts tended to vanish into thin air. R. S. Mulliken
    11. Re:"Full generation behind"? by Sj0 · · Score: 1

      You're missing a key point: AMD has to have something to differentiate themselves from Intel. If they release an MCM quad-core chip, they're just following Intel. On the other hand, the press is eating up the monolithic quad-core chip, precisely because of the perception that the monolithic design is special and different from the existing quad-core chips.

      --
      It's been a long time.
    12. Re:"Full generation behind"? by Mr+Z · · Score: 1

      This is not entirely true. Although I agree overall with what you're saying, core logic transistors scale much worse than cache as the manufacturing process decreases in size.

      Fair enough. That said, it's not the transistors so much as it is the wires that don't scale well. I'll warn you: I'm not a physical designer, I'm just an architect. The one and only cookie I directly designed and baked was in 2 micron. That said, I'm aware of the trends.

      There could be a couple reasons AMD throws more logic at the problem than RAM:

      • At an older technology node, the optimal RAM/logic split may be different.
      • Their RAM technology may be more susceptible to defects (e.g. less redundancy).
      • They favor shorter latencies over higher overall capacity.
      • Related to the previous bullet: They may have higher latency for the same capacity (although their large L1s tend to suggest otherwise).

      Indeed, the latency argument is probably the most interesting one. It's seemed to me that the biggest advantage AMD has held over Intel in the previous generation was due to keeping latencies down. The generous L1s, the exclusive cache protocol (where L2 is essentially a ginormous victim buffer and misses allocate directly in L1), and the integrated memory controller are all focused on latency. Lower latencies == more instructions issuing together == more instruction level parallelism. And in branchy code, it means less time branch-to-branch.

      Pentium 4 was focused on throughput and bandwidth, and lost on pointer chasing and branching. Huge L2 line sizes, and the high-bandwidth, high-latency RDRAM interface... all that focused on streaming, not on more general workloads. It killed them on the bread and butter generic stuff. Intel finally woke up to that when it shifted over to Pentium M from Pentium 4, and you can see how well they learned by the time you get to Core 2.

      Bust out the popcorn, this'll be fun to watch. :-)

      --Joe
      --
      Program Intellivision!
    13. Re:"Full generation behind"? by fitten · · Score: 1

      Yeah, that's a good reason to drive the company to bankruptcy.

  6. Techreport by Eukariote · · Score: 5, Informative

    The Techreport also has a review up: http://techreport.com/articles.x/13176/1. Barcelona is similar to Core2, clock for clock. It has better energy efficiency and SMP scaling. But the clock frequencies will need to come up in order to beat Intel's highest clocking chips in absolute performance.

    1. Re:Techreport by tietokone-olmi · · Score: 3, Insightful

      AMD's had I/O performance and memory latency advantages on Intel even before Barcelona though. I suppose Intel will be in even more serious trouble than before in the server space, until it can get its next-generation bus thingy (CSI they called it?) up and running in a year or three. Until then, Intel's stuck in a SMP scaling black hole... and I don't really see Intel coming out with integrated memory controllers and native NUMA like AMD did with their whiz-bang DEC Alpha engineers.

      Once Barcelona ramps up, Intel's going to be hard pressed to come up with an advantage besides clock speed for the C2 microarchitecture, given that Barcelona finally ups the SSE units to proper 128-bit wide computation; i.e. none of that splitting of SSE operations into pieces that are executed 2 pairs of operands at a time.

      Remember, high-performance floating point is not the mainstream workload that determines the success or failure of a microarchitecture. (Though it is one of the sexier ones.) So no yammering about "absolute performance" there; AMD's previous-gen offerings were crazy fast before the C2D and aren't half bad even after C2D.

    2. Re:Techreport by sabre86 · · Score: 1

      Good link. Glad you RTFA. However, since this is a server chip, why don't they do benchmarking with something other than Windows? Does anyone know of decent benchmarking apps or tech sites that use them for, say, Linux?

      --sabre86

    3. Re:Techreport by Zoxed · · Score: 1

      As someone who is interested in quiet desktops this article has an interest comment on CPU power specs in that AMD seems to have gotten tired of direct wattage comparisons between different methods of measurements: ie Intels "typical" rating versus AMDs "maximum" rating. So AMD are now introducing their ACP rating "average CPU power".

  7. I'm curious by ircmaxell · · Score: 1

    I wonder how much these things will go for... I know they won't be cheap (in traditional terms), but since AMD has a history of comperable performance for less, I'm really curious how affordable these things will be. If the price is right, one of these may be in my near future...

    One thing of note is that Motherboards already exist for this processor in fair number. The Barcelona uses a socket F (1207) which the current dual core Operons already use. That should give this processor a decent jumpstart in terms of people not needing to buy a new board to run it.

    --
    If a man isn't willing to take some risk for his opinions, either his opinions are no good or he's no good
    1. Re:I'm curious by llirik · · Score: 3, Informative

      * 2347 - 1,9 GHz, $316
      * 2350 - 2,0 GHz, $389
      * 8347 - 1,9 GHz, $786
      * 8350 - 2,0 GHz, $1019
      * 2344 HE - 1,7 GHz, $209
      * 2346 HE - 1,8 GHz, $255
      * 2347 HE - 1,9 GHz, $377
      * 8346 HE - 1,8 GHz, $698
      * 8347 HE - 1,9 GHz, $873

  8. Finally, (more) fair wattage numbers! by pla · · Score: 1

    AMD says it won't use the ACP number to compare the power consumption of its processors against Intel's.

    Before everyone slams them for coming up with yet another cheesy marketing gimmick, I would point out that Intel has done this ever since the first of the power-sucking P4 line. They did it a bit less up-front, however, choosing to redefine "TDP" in their specs rather than give their numbers a new term (such as "ACP").

    This still won't make for a completely fair direct comparison, because Intel's TDP refers to a sort of third quartile case rather than the mean. But better, anyway.

    Perhaps more usefully, since AMD has published the ACP spec, even though they may refuse to run similar tests on Intel chips, you can bet the farm that 3rd-party results will come out over the next few days.

  9. "right AMD's Ship" ? by gearloos · · Score: 1

    I don't understand why everyone always talks about AMD's problems. So what if they are not the forunners on every front. They still lead in other areas. Example from July's Inquirer http://www.theinquirer.net/default.aspx?article=41 138&rfp=dta/

    --
    "Computers are a lot like Air Conditioners" "They both work great until you start opening Windows"
    1. Re:"right AMD's Ship" ? by Duffy13 · · Score: 1

      Not to mention quite often in my experiences when AMD's version of something is usually better. Which may be why they are a tad slow on releasing products.

      --
      "Now you know, and knowing is half the battle!"
    2. Re:"right AMD's Ship" ? by Gr8Apes · · Score: 1

      Why rush when your current system is within 10% of your rival's best efforts? Not to mention that they still lead in the server space on a number of metrics.

      --
      The cesspool just got a check and balance.
    3. Re:"right AMD's Ship" ? by Wavicle · · Score: 4, Insightful

      I don't understand why everyone always talks about AMD's problems.

      Because it doesn't matter how many fronts you are leading on, if you run out of money and can't borrow any more, you lose.

      AMD has been running out of money, fortunately they can still borrow. If they don't stop losing money their credit rating will tank and then they will not be able to borrow any more.

      THAT is what righting the ship means.

      --
      Education is a better safeguard of liberty than a standing army.
      Edward Everett (1794 - 1865)
    4. Re:"right AMD's Ship" ? by Sj0 · · Score: 1

      Because AMD lost over half a billion dollars this quarter and they need to start making money?

      --
      It's been a long time.
    5. Re:"right AMD's Ship" ? by Courageous · · Score: 1

      Really what they need to do is pull off a big merger with just about any other very large company. Samsung would be a good one. Sun wouldn't be half bad, considering Sun's vested interest in Opteron. Although I'd say Sun isn't big enough, even. IBM?

      C//

    6. Re:"right AMD's Ship" ? by Gr8Apes · · Score: 1

      Better to lose some 'x' amount of money vs losing all credibility, mindshare, etc, by releasing a flawed product in haste. We'll see if their strategy pays off in the near future. I noted from the benchmarks that these 1.7-2GHz chips keep up with Intels in many benchmarks, but Intel's best are still slightly ahead.

      It'll be good to see what comes up in the next 2-3 months as production ramps up.

      --
      The cesspool just got a check and balance.
  10. Cool by afidel · · Score: 2, Interesting

    Literally. I can't wait to get in our first DL585 G2 with 4 of these beasties and 64GB of ram. The only regret I have is that we probably won't use em for DB servers because of Oracle's asinine policy of charging per core, sometimes I wish we had gone SQL2005 for more stuff as it is going to scale better with improving hardware. Then again maybe the proliferation of quad core (and above) server cpu's will make Oracle rethink their pricing policy again. I hope they go to what the rest of the industry is doing and license per socket.

    --
    There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
    1. Re:Cool by SpecTheIntro · · Score: 1, Troll

      The only regret I have is that we probably won't use em for DB servers because of Oracle's asinine policy of charging per core, sometimes I wish we had gone SQL2005 for more stuff as it is going to scale better with improving hardware.

      That is the most draconian pricing policy I have ever heard. You actually have to pay Oracle for increasing your processing power?

      And an honest question: was there a reason why you didn't look at MySQL or PostgreSQL? I'm not a database expert but my work with them has made me believe they are robust solutions--I certainly prefer them to MSSQL, which is about as pleasant to use as a suppository.

    2. Re:Cool by jlarocco · · Score: 1

      That is the most draconian pricing policy I have ever heard. You actually have to pay Oracle for increasing your processing power?

      Um, yeah. Charging per processor (or machine) is par for the course for large "enterprise" software packages. Oracle, Rational, all the hardcore rendering software, etc. they all do it. Welcome to real life.

      And an honest question: was there a reason why you didn't look at MySQL or PostgreSQL? I'm not a database expert but my work with them has made me believe they are robust solutions--I certainly prefer them to MSSQL, which is about as pleasant to use as a suppository.

      When it comes to being robust, MySQL isn't even in the same league as Oracle. PostgreSQL is a lot better, but I still don't think it compares very well to Oracle.

      --
      Maybe not
    3. Re:Cool by Rogerborg · · Score: 1, Interesting

      Charging per processor (or machine) is par for the course for large "enterprise" software packages.

      I think what the grandparent is distressed about is that they charge per core, rather than per physically discrete processor.

      It is an interesting issue; what if you promised, honest Injun, to only use 3 cores of a quad-core CPU, for example?

      --
      If you were blocking sigs, you wouldn't have to read this.
    4. Re:Cool by SpecTheIntro · · Score: 1

      Charging per processor or machine

      There's a large gap between charging per processor or per machine. A machine could feasibly be used to independently run the software alongside other instances of it on the network, so charging for another license isn't unreasonable. But charging per core doesn't make any sense to me: unless each core is running a separate and independent instance of Oracle (can it be programmed to do so? Does virtualization play a role here?) then it just seems like you're being penalized for attempting to increase your efficiency.

      When it comes to being robust, MySQL isn't even in the same league as Oracle. PostgreSQL is a lot better, but I still don't think it compares very well to Oracle.

      Fair enough; I'm not particularly well-versed in database solutions. But judging from the pricing scheme Oracle employs, I wonder how many people find ways to work within Postgre's limitations?

    5. Re:Cool by Anarke_Incarnate · · Score: 1

      How about a Sunfire x4600 M2 ? Then you can get 8 of them in a 4U :)

    6. Re:Cool by afidel · · Score: 2, Informative

      You are charged per core and can only go below the number of physical cores in the machine if the architecture has hard partitioning of resources, for instance a zone with hard resource limits is acceptable but a container with soft limits is not (well, it is but you need licenses for the max possible resources the container has access to).

      --
      There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
    7. Re:Cool by TeknoHog · · Score: 1

      I think what the grandparent is distressed about is that they charge per core, rather than per physically discrete processor.

      I don't get it. A 'core' is simply a CPU that happens to share a piece of silicon with other CPUs. Charging per a piece of silicon doesn't really reflect the computing resources available for the application.

      --
      Escher was the first MC and Giger invented the HR department.
    8. Re:Cool by jagdish · · Score: 1

      .. with 4 of these beasties and 64GB of ram. The only regret I have is that we probably won't use em for DB servers.
      Atleast you'll be able to run Vista.
    9. Re:Cool by WuphonsReach · · Score: 1

      There's a large gap between charging per processor or per machine. A machine could feasibly be used to independently run the software alongside other instances of it on the network, so charging for another license isn't unreasonable. But charging per core doesn't make any sense to me: unless each core is running a separate and independent instance of Oracle (can it be programmed to do so? Does virtualization play a role here?) then it just seems like you're being penalized for attempting to increase your efficiency.

      It's a carry-over from the old mainframe days.

      Basically it boils down to "charge the customer as much as possible".

      Their argument is (probably) that more cores allows you to process more data simultaneously, therefore you should fork over more money for a more expensive license. Supposedly so that you can get the more expensive version of Oracle (if such beast exists) that will take advantage of all of those cores.

      Look at it from their side. In the past, you may have needed 16 individual single-core boxes to handle the workload. So you paid them for 16 licenses. Now, you can do the job with a single quad-core quad-CPU beast. Which would cut Oracle revenues to 1/16 if they only licensed per-box.

      Personally, I feel that as of PostgreSQL 8.1, it's more then a match for Oracle in most situations. But nobody gets fired for buying Oracle.

      --
      Wolde you bothe eate your cake, and have your cake?
  11. The penalty of moving too fast... by tjstork · · Score: 1

    Is that sometimes, you wind up losing the race. I think AMD tried to take a risky approach of putting all four cores on one die and they shot themselves in the foot. Now Intel got some sort of quad core out there. Even if it wasn't as good, it was still better than two completely separate 2 way chips, and now, Intel is circling the wagons to do its own native quad core implementation.

    I fear that Barcelona might well wind up as the Great Eastern of chip making - an impressive technological first, but, too much at once. It's a shame too, as I really, really, love my Opteron line, and I know that I will yet again ruin my career writing some sort of multithreaded thing for Barcelona because it is so novel, even if AMD gets battered by a less than spectacular launch.

    --
    This is my sig.
    1. Re:The penalty of moving too fast... by Rogerborg · · Score: 1

      Never mind Great Eastern, Barcelona might be the Itanium of chip making.

      --
      If you were blocking sigs, you wouldn't have to read this.
    2. Re:The penalty of moving too fast... by Chris+Burke · · Score: 1

      Well, developing MCM capabilities isn't exactly risk-free either. Since in any event the MCM version would be a temporary hold over, the question is does it make sense to spend the R&D time to develop that technology for a product that would, ideally, only exist in the market for a short period?

      Remember, AMD isn't Intel. They don't have the same resources as Intel. Spending the time on an MCM design would necessarily mean having fewer resources devoted to the native quad core version, and for whatever follow-on products are certainly in the pipeline now. So the question is -- delay the native quad core by working on an MCM, a technology that AMD has never done before and thus carries its own risks of delays, or focus on the quad core knowing you'll sacrifice some time to market, but hopefully being in a better position once you are done?

      AMD chose the latter. Was it the right choice? I can't say. I can say that the time between Intel's MCM quad core and the Barcelona launch has been a bad time for AMD. Before that, AMD was enjoying greater than 30% server market share which is absolutely fantastic given the relative size of the competitors, but now they've dropped back down to around 20% I believe. They can probably gain a lot of it back -- the inherent advantages of Opteron vs equivalent Xeon parts haven't gone away -- but nevertheless they lost a lot of momentum. Certainly the delay in Barcelona's launch hurt them badly.

      But if they went the other way? And the MCM part was also delayed, not as good as the native quad core, and the true native quad core delayed even further because of diverting resources from it? The end result may not have been any better. That's the nature of this industry -- there isn't any choice that doesn't involve a lot of risk.

      --

      The enemies of Democracy are
    3. Re:The penalty of moving too fast... by Jorgandar · · Score: 1

      Yeah, it reminds me about what happened to Sega in the home console market. They were "firsts" in a variety of changes, like the 32-bit add-on for genesis, the sega "CD", which was 32 bit, and the "dreamcast" that had the first console networking capabilities (at, ahem, modem speed). Soon after the competitors learned what worked and what didnt, and innovated around them. Sometimes being first in a new technology isnt better.

  12. question for the local geniuses... by Ralph+Spoilsport · · Score: 1
    If this chip has four cores, how much faster does this actually make something happen? Doesn't the software have to be optimised for multiprocessors? If someone could 'splain this, it would be a public service.

    Cheers.

    RS

    --
    Shoes for Industry. Shoes for the Dead.
    1. Re:question for the local geniuses... by Anonymous Coward · · Score: 0

      Doesn't the software have to be optimised for multiprocessors?
      Yes it would. Probably is. Without multithreaded programs (i.e. programs optimised for multiprocessors/cores) the only speedup is that other programs run on the other core(s), thus eating up your program's core time. Naturally, things are not this simplistic, you have issues with cache and memory contention, but that is hardly multiprocessors 101.

    2. Re:question for the local geniuses... by TheThiefMaster · · Score: 3, Informative

      It depends on three things:
      1: Whether the software CAN use multiple cores.
      2: How efficiently it uses the extra cores.
      3: Whether the program is currently limited by cpu power or by something else.

      For "1:", if the program can't use the extra cores, then you'll only see a speed improvement from the fact that the cores are 15% more efficient. i.e. A 2GHz one of these quads performs the same as a 2.3GHz (+15%) dual core from the previous generation for applications in this category.

      For "2:", if the program can use the extra cores, but not as efficiently as the first, then you'll see a speed increase equivalent to this. e.g., if the program does two tasks at once, one that takes 70 seconds and one that takes 30, then on one core it'll take 100 seconds. On two cores it would do the 70 second task on one core and the 30 second task on the other, reducing the total time to 70 seconds, a ~40% speed improvement.

      For "3:", if the application is limited by something other than the cpu, e.g. "how quickly it can pull data from the hard-disk", you will likely see no improvement whatsoever.

      In conclusion, depending on what applications you use, you will see anywhere from no improvement up to 2.3x the previous speed (x2 for double the cores and +15% from the improved efficiency).

      Note: As these cpus also have an extra instruction set extension, applications that make use of this could exceed the speed improvements I noted above.

    3. Re:question for the local geniuses... by Rogerborg · · Score: 1

      It doesn't make "some" (i.e. one) thing faster, it makes four concurrent things faster. These are server processors. Servers do lots of things at the same time. Even on your desktop, a dual core chip is useful; you can ask stupid questions on Slashdot without slowing down encoding your goat porn home movies.

      --
      If you were blocking sigs, you wouldn't have to read this.
    4. Re:question for the local geniuses... by fm6 · · Score: 3, Insightful

      Doesn't the software have to be optimised for multiprocessors?
      Well, it has to be multithreaded. Thing is, a lot of software is multithreaded already; even on a single-core system, it makes sense to distribute functionality among multiple threads so that resources are used efficiently. On server systems (which is where Opterons are mostly used) software pretty much has to be multithreaded — you don't want all your other clients hanging when one client is waiting on a resource. A web server is a classic example.

      When you move a multithreaded program to a system with more cores, than any given thread is more likely to get a core to run on when it needs it. Assuming, of course, that you have enough threads so that's an issue.

      Shameless plug: I'm the docs lead for this Opeteron-based server, which can have up to 8 CPUs, for a total of 16 cores. When the Barcelona-based CPU modules are ready, customers will be able to upgrade their systems to a maximum of 32 cores. (Don't ask me when this will happen; Marketing would have me killed.) Obviously any software running on such a system has already dealt with the multicore optimization issue.
    5. Re:question for the local geniuses... by Workaphobia · · Score: 1

      I don't understand why everyone is so concerned with multithreaded programs that take advantage of multicore CPUs. I'm interested in it simply so I can multitask better between discrete (probably single-threaded) applications. But I run Gentoo, so I guess I need all the extra simultaneous horsepower I can get.

      --
      Evidently, the key to understanding recursion is to begin by understanding recursion. The rest is easy.
    6. Re:question for the local geniuses... by MrTwist · · Score: 1

      Software has to be multi-threaded to make the best use of multi-core chips.

      http://en.wikipedia.org/wiki/Multi-threading

    7. Re:question for the local geniuses... by karthikg · · Score: 0

      Exactly.. all you need is a scheduler which can exploit multiple CPUs/core and have a bunch of processes in runnable state. You don't really need a single program to be multi-threaded to use a multi-core CPU.

      For example, a simple shell pipeline 'p1 | p2 | p3 | p4 ' can get a theoretical speed-up of 4 on a quad-core machine when none of p1..p4 is multi-threaded.

      Karthik

  13. Not another fake number AMD! by PalmKiller · · Score: 0

    But AMD customers who relied on the company's previous power metric of TDP (thermal design power) were putting too many resources into cooling and electrical supply, said Bruce Shaw, director of server and workstation marketing for AMD. That's because TDP was developed so server manufacturers would know much power the chip consumes in worst-case maximum-power situations that very rarely occur, and design their systems accordingly, he said. So now AMD will advise customers of an Opteron processor's average CPU (central processing unit) power, or ACP. "ACP is meant to be the best real-world end-user estimate of what they are likely to see from the power consumption on the processor," Shaw said. Oh Great, first they used the + speed numbers (which I think cyrix actually started but they jumped right in there). I can see that the core speed really meant nothing so I was OK with that, but TDP is a real number. Obviously their marketing folks decided it drew too much power so they opted to make up a lower power usage number. Frankly even a home user wants to know top power usage of the cpu and video card to properly size their power supply.
    1. Re:Not another fake number AMD! by afidel · · Score: 4, Insightful

      Uh, they are doing this to come closer to Intel's TDP numbers which have been average high use numbers instead of worst case for at least the last couple generation of chips. AMD is actually being much more upfront here by offering both worst case and average case numbers, I hope Intel follows their lead and offers both numbers.

      --
      There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
    2. Re:Not another fake number AMD! by PalmKiller · · Score: 1

      Ah, my bad, thanks for clearing this up...so that explains Intels ability to suddenly have lower power chips...so it is they that are playing with the numbers this time, interesting :)

    3. Re:Not another fake number AMD! by Chris+Burke · · Score: 3, Informative

      Ah, my bad, thanks for clearing this up...so that explains Intels ability to suddenly have lower power chips...so it is they that are playing with the numbers this time, interesting :)

      To some extent. The Pentium 4 is where this started. The Netburst architecture was very power hungry normally, but it's maximum power was insane. The graph of power consumption vs benchmark had a long "tail", which Intel sought to chop off. See, TDP is a real-life number, since it's used by OEMs and others to design thermal solutions for the parts. If the thermal solution is insufficient, then the parts fail. So it's not actually possible to fudge TDP numbers.

      What Intel decided to do was implement an on-chip thermal diode and some logic that halved the effective clock cycle* if the temperature went above a certain threshold. What this meant is that based on how they programmed this logic, they could guarantee that the chip's power consumption would never go above a certain level no matter what code you were running. They had effectively lopped off the long tail. The downside is that if your application does draw more power than the limit, then you'll see vastly reduced performance because of the clock throttling. Most of the time this is transient so it's not that noticeable, but there were benchmarks out there that showed this effect very clearly. Like a certain game benchmark would get lower scores at 640x480 than 1600x1200 because at the lower res the game was cpu bound as was crossing the thermal threshold.

      So theoretically with this feature Intel could fudge the numbers however they wanted and claim whatever TDP they desired. In practice they don't have that much flexibility because if they set the bar too low then their effective performance would suck, and their TDP numbers are set at average power + several standard deviations.

      The main reason why Intel was able to suddenly have low power chips is because they ditched the Netburst architecture and went back to a design that was more balanced between high clock speeds and high IPC.

      They kept the clock throttling logic, though, since it does still give them some benefit in reporting lower TDP numbers. AMD doesn't have this feature, so their TDP is truly the maximum power (as determined by running a "power virus") that you would ever see, even though it's unlikely. Since power has become ever more important as a marketing feature even outside of mobile, I'm not surprised that AMD would decide to start touting expected numbers vs maximum.

      * Actually a 50% duty cycle of full speed for some number of microseconds followed by completely off.

      --

      The enemies of Democracy are
  14. If you considered using MSSQL by Anonymous Coward · · Score: 1, Insightful

    If your database requirements are such that you could have use MSSQL instead of Oracle, you really should consider using PostgreSQL, as it is more featureful and costs nothing.

    1. Re:If you considered using MSSQL by afidel · · Score: 3, Insightful

      Uh, MSSQL 2005 is a serious enterprise DB, this isn't SQL 7 anymore. Also none of our enterprise software supports PostgreSQL so invalidating our 6 or 7 figure support contracts just isn't an option even if it WOULD work.

      --
      There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
    2. Re:If you considered using MSSQL by Gr8Apes · · Score: 0

      MSSQL still sucks rocks (I've actually had the misfortune to work with a MSSQL Cluster). If your system supports Oracle, it's usually not too tough to migrate to PostgreSQL. The main problem you'll have is deprogramming your DBA's so they can effectively work with PostgreSQL. And yes, I've used Oracle, DB2, PostgreSQL, MySQL, Sybase, Informix, and, yes, MSSQL. Depending upon what you're doing, these days you'd use Oracle, DB2, PostgreSQL, or MySQL. The others just don't have any benefits over these for new projects. My current project is being developed to run on Oracle, MySQL, and PostgreSQL.

      --
      The cesspool just got a check and balance.
    3. Re:If you considered using MSSQL by Anonymous Coward · · Score: 0
      Well, first, that depends on what you mean by "serious enterprise DB." MSSQL's features don't hold a candle to Oracle or DB2, whatever Microsoft has to say, if only because of the choice of platforms you can run it on. Sorry to say, but no machine that runs Windows can provide the reliability of the really enterprise hardware you can get Oracle and DB2 running on.


      Second, if you're paying that much for support contracts, then you can afford to pay the Oracle extortion. Or go with DB2, which is as good a product for less money.


      Third, you're probably stuck with Oracle anyhow. If you're running enough software on it that you've got 6-7 figure support contracts, you've got enough DBA infrastructure that a switch to any DB platform is an expensive, multi-month/year affair that's fraught with danger.


      But the condescending tone of your response was quite charming; I was trying to offer a suggestion to someone who in all likelihood was developing things in house. Here's hoping nobody ever offers you advice again, especially when you'd profit by it.

    4. Re:If you considered using MSSQL by Gothmolly · · Score: 0

      Keep repeating that, fanboi. Never use something version less then v2 or greater than v10. That translates roughly to "after years and years of shitty, tinkertoy DB, MS finally got their act together, but we've moved on since then."

      --
      I want to delete my account but Slashdot doesn't allow it.
    5. Re:If you considered using MSSQL by afidel · · Score: 1

      fanboi, hardly. I'm quite the MS critic if you bother to read my post history, but MSSQL 2005 is something they got right and I give credit where credit is due. Actually we are having our issues with Oracle at the moment and that combines with their outrageous licensing policies has me a bit peaved off. They have an admitted bug that affects all 10gR2 versions prior to 10.2.0.4 which sends random dates at times to the Oracle Enterprise Management DB causing different OEM tools like DB replication to fail yet they have refused to issue a backport patch on Windows which is going to force us to do full regression testing of all of our apps so we can put the fix in place.

      --
      There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
  15. I've been buying Intel/Nvidia . . . by WhiteWolf666 · · Score: 2, Interesting

    For a few years now, as that was the only platform that really, reliably ran Linux.

    Intel's been good to us Linux folk, and Nvidia has been easy enough to deal with.

    If AMD comes out with an end-to-end Linux solution, CPU, GPU, and a good Linux-friendly partner for chipset, I'll seriously consider switching back to AMD parts.

    --
    WhiteWolf666 an exBush supporter. All you new-school,compassionate,save the children Republicans can rot in hell
    1. Re:I've been buying Intel/Nvidia . . . by Anonymous Coward · · Score: 4, Interesting

      Interesting....I've been buying AMD/NVidia for a few years now for the exact same reason....

    2. Re:I've been buying Intel/Nvidia . . . by ynososiduts · · Score: 1

      Intel releases OSS drivers for their hardware. AMD never really made much that needed drivers up until their buying out ATI, and ATI doesn't have the best Linux track record. So Intel really is the better company to buy for Linux compatibility.

      --
      622677120
    3. Re:I've been buying Intel/Nvidia . . . by Anonymous Coward · · Score: 0

      Erm... None of AMD's stuff required drivers, and that somehow makes them less Linux-friendly than Intel? How does that work?

  16. haha, oh man, charging per core is a break by Anonymous Coward · · Score: 1, Informative

    vs. the old days. Until not too long ago, they charged based on "power units". What's a power unit you ask? 1 MHz on x86 was 1PU, 1MHz on sparc was 1.5PU, etc. (So for example your departmental e450 with four, 400 MHz cpus would be 4x400x1.5 = 2400 PUs.) How much did a PU cost in licensing? Well, you see, said the oracle salesman with a gleam in his eye, that all depends... That they've shifted to a flat rate per core is actually a big win over the old model for their customers.

    1. Re:haha, oh man, charging per core is a break by BiggestPOS · · Score: 1
      I've never used it but Oracle is either one hell of a database, or one hell of a brand for people to put up with tactics like that.

      That shouldn't even be legal.

      --
      What, me worry?
    2. Re:haha, oh man, charging per core is a break by Chandon+Seldon · · Score: 2, Insightful

      I've never used it but Oracle is either one hell of a database, or one hell of a brand for people to put up with tactics like that. That shouldn't even be legal.

      Oracle is an amazingly powerful brand and managers think that "scalability" is something you buy rather than an engineering problem for programmers and system architects to solve. That's really the whole story. Given what servers cost and the actual performance differences between different database software given appropriately written client software, purchasing Oracle licenses is largely inexcusable unless you have existing Oracle dependent software and no time to switch databases and re-address scaling related design questions.

      --
      -- The act of censorship is always worse than whatever is being censored. Always.
  17. AMD also has more energy-efficient chipsets with.. by Joe+The+Dragon · · Score: 1

    AMD also has more energy-efficient chipsets with ECC DDR2 ram not the higher power useing FB-DIMMS also you have a lot more choice in chip set's with more pci-e lanes then intel 2p+ systems.

  18. Clock for clock Barcelona is faster than Cloverton by Anonymous Coward · · Score: 1, Interesting

    The 2.0GHz Barcelona beats the 3.0GHz Xeon X5365 (Cloverton) on floating point. Barcelona specfp_rate2006 score is 73.0 to Cloverton's 66.9. Things can only get better as AMD cranks up the clock in the coming months.

    If you scale the benchmarks to the same GHz rating you will see that clock for clock Barcelona is at worst on par with Intel's best chip, and at best 80% faster on floating point. This is really quite amazing when you consider it's using the same amount of power as the previous 2 core AMD Opteron.

  19. More Barcelona by bigwophh · · Score: 2, Informative

    Specs of the entire new Barcelons line-up, more details, and pricing are available here as well:

    http://www.hothardware.com/Articles/AMD_Barcelona_Architecture_Launch_Native_QuadCore

  20. Re:AMD also has more energy-efficient chipsets wit by Anonymous Coward · · Score: 0
    InfoWorld : AMD's $389 big iron

    ... Now, sitting on the floor in front Barcelona, I am speechless. I'm running all eight cores, full-out, and watching the watt/ammeter keep a record. The evening started with an all-nigt burn in. During the systems' first all-night toughest load, with 100% for each of eight cores, I didn't expect when I saw: 2 amps, measured at the outlet, or just 300 watts. When the workload dropped to idle in came in at 1.3 amp, or 149 watts.
     
    It was not a wimpily-configured box: 8 GB Registered ECC DRAM, on-board I/O, memory and SMP node controllers on-board. Two banks of RAM for each socket, and I'll explain when I get to the result, a third-level cache turned out to be a major win.
     
    I'm barely getting warmed up, and yet I'm weary, but after everthing I learned, it's the price that got my blood boiling: $389! That's desktop chip money. So AMD is blazing trails in more ways than one.


    He seems pretty impressed .. to say the least.
    I guess he'll put out Benchmarks later.
  21. Nonsense by Anonymous Coward · · Score: 0

    First of all, separate the GPUs from CPUs in your argument.

    Nvidia's drivers are closed opaque binary modules - how's that for open?
    AMD has opened up their specs. Nvidia hasn't.

    As for the CPU, how is AMD's CPU less open than Intel's CPU? The specs for both are widely available.

    Your arguments do not hold up.

    1. Re:Nonsense by Repossessed · · Score: 1

      AMD only *just* opened up their specs though (though it's a fairly fast turnaround from buying out ATI), so it'll be a little while before ATI is well supported on Linux.

      --
      Liberte, Egalite, Fraternite (TM)
    2. Re:Nonsense by Anonymous Coward · · Score: 0

      The guy you replied to is taking about CPUs, not GPUs. There is a difference.

  22. Re:AMD also has more energy-efficient chipsets wit by Anonymous Coward · · Score: 0

    I got 300 Watts running a SuperMicro Twin with 1 Xeon 5310 per machine and 4 GB Reg, ECC RAM, 2 160 GB SATA drives for each machine - so 300 Watts for pretty full load (I ran stress for a while) on 2 Quad core systems, not bad. Waiting for the AMD version of the supermicro twin. 6015TV, I think.

  23. Hey douche... by Anonymous Coward · · Score: 0

    Intel's quad cores are two dual cores glued together. GLUED. TOGETHER!

    1. Re:Hey douche... by OrangeTide · · Score: 1

      The use glue? What. sort. of. glue?

      --
      “Common sense is not so common.” — Voltaire
  24. Re:Clock for clock Barcelona is faster than Clover by Wavicle · · Score: 3, Insightful

    specfp rate was running faster on pre-barcelona dual core Opterons than on Intel's dual core Woodcrest. The reason is no big secret: specfp is memory bandwidth limited and specfp_rate is specfp's running in parallel. Here is a good anandtech article on the subject.

    We already know that AMD has superior memory performance. If you are doing bandwidth-limited floating point, Barcelona is the clear winner.

    If you're making a general statement about floating point performance, you're wrong.

    --
    Education is a better safeguard of liberty than a standing army.
    Edward Everett (1794 - 1865)
  25. When will Nvidia open their specs? by Anonymous Coward · · Score: 0

    When will NVidia release their hardware specifications so people can make a properly supported X11 open source driver?

    The random NVidia crashes I get from their binary X11 driver - when will they be fixed?

  26. does Intel need AMD by blowdog · · Score: 1

    I have owned a PC since early 95 and primarily for financial reasons and then because the Athlon range were, in my option, the best processors available I had not owned an Intel CPU for 12 years. Now for the first time I have an Intel chip, a Dual Core 2 Duo, in my laptop as AMD just can't compete on price or performance. My new desktop will have an Intel Quad core as Barcelona can barely compete with the 65nm chips and the new 45nm chips will just blow it away. Although Intel has it's fingers in many pies not just X86 doesn't Intel need a stronger AMD to stop the government getting involved in trying to limit it's activities to maintain competition?

    1. Re:does Intel need AMD by MrTwist · · Score: 1

      I completely disagree. AMD's memory architecture is superior to Intel's and that makes for a much more pleasant computing experience -- in my opinion. I ran an Intel P100 as my first 32bit machine and was happy with it until I got an AMD K2-300. That AMD chip gave snappier performance than a PIII-500. I fell for the Intel Core2 Duo hype and bought a laptop loaded with a 2Ghz CPU. Running Vista, it underperforms the AMD Turon X2 with half the memory. AMD's memory architecture is superior. There is no doubt about it. And in a Von Neumann machine, that's a big piece of the puzzle. Running real-world applications, I don't feel like I'm fighting the machine because the OS can keep up with me running on AMD. I'll be glad to give AMD my money when I bump my desktop machine.

    2. Re:does Intel need AMD by Godsmack74 · · Score: 1

      Looks like TechwareLabs had the first review to go live: http://www.techwarelabs.com/reviews/processors/barcelona

  27. Re:Clock for clock Barcelona is faster than Clover by Anonymous Coward · · Score: 0

    When only measuring single core performance, clock for clock, Barcelona is on par with Cloverton.
    But in multi-core, it's a completely different story - based on what is available on the market today Barcelona is the fastest multicore/multithreaded x86 floating point solution available - at any price or clock speed. What good is an Intel chip that has fast floating point but the bus cannot feed it data fast enough? Nothing I can buy today can fix that problem.

  28. Re:Clock for clock Barcelona is faster than Clover by Wavicle · · Score: 2, Informative

    When only measuring single core performance, clock for clock, Barcelona is on par with Cloverton.

    Unfortunately processors are not generally sold "clock for clock." If you're on par clock for clock, but the other guy is clocked more than 50% faster than you... that could be trouble.

    What good is an Intel chip that has fast floating point but the bus cannot feed it data fast enough?

    Plenty good if the data can fit in cache, in which case the unit can be fed fast enough. For instance, say you're running LinPack. But then, who uses LinPack as a benchmark?

    --
    Education is a better safeguard of liberty than a standing army.
    Edward Everett (1794 - 1865)
  29. Re:Clock for clock Barcelona is faster than Clover by Anonymous Coward · · Score: 0

    I write multi-threaded floating point intensive systems for a living. I simply want to use the chip that gives me the greatest floating point throughput I can get. Right now that chip appears to be Barcelona. I'm not interested with hypothetical arguments - only speed. I am looking forward to using Barcelona processors because they will get my mathematical computations done faster.

    By the way, check out number 2 and 3 on your top 500 supercomputer list - they're Opterons.

  30. Gotta love mod abuse. by Anonymous Coward · · Score: 0

    Yep, apparently modding is underrated.

  31. Blah by default+luser · · Score: 1

    What can I say? I'm disappointed that they stuck with a 3-issue architecture - while it is true that Intel's 4-issue setup is often data-starved, even with exceptional I/O performance AMD can only hope to match the Core platform in most situations. The lack of progress in their cache technology means AMD gets as much burden as benefit out of the L3 cache over 20ns access time!).

    In the I/O arena, AMD potentially has the edge, and for HPC there's no question Barcelona will do well: this architecture is built for scaling beyond 8 cores (potentially as high as 8 sockets, 32 cores). The question is: how big is that market, and will it really pay-off for them?

    In the meantime, small-scale servers and workstations are the majority of the market, and Intel still has a commanding lead in that space. Intel's introduction of the San Clemente Xeon chipset with DDR2 instead of FBDIMMs will not help things for AMD, because these systems can most certainly compete with AMD's low power consumption.

    --

    Man is the animal that laughs.
    And occasionally whores for Karma.

  32. Re:Clock for clock Barcelona is faster than Clover by Wavicle · · Score: 3, Informative

    I simply want to use the chip that gives me the greatest floating point throughput I can get.

    Define throughput. At some point you need to decide if you are solving equations like LinPack or equations like spec_fp. One causes lots of cache misses and benefits from memory bandwidth, the other does not.

    Right now that chip appears to be Barcelona.

    Well that's a hypothetical statement based on perception of your needs and their marketing.

    I'm not interested with hypothetical arguments

    That explains why you're making them (???)

    I am looking forward to using Barcelona processors because they will get my mathematical computations done faster.

    Hypothetically. Are you going to hypothetically switch when Intel's Penryn with SSE4 comes out? What about Intel's Nehalem?

    By the way, check out number 2 and 3 on your top 500 supercomputer list - they're Opterons.

    And?? They were designed and built before Core 2 was released. Do you think I'm going to argue they should have used Pentium 4's? Those systems also make solid use of NUMA through a custom Cray crossbar (Seastar), and Intel doesn't have that. If they made them today I see no reason for them not to use Opterons. Do you have a computer with lots of Opterons and a Cray Seastar router on order?

    The performance of those systems is measured using LinPack. As I mentioned at the beginning, declaring a 2.0 GHz Barcelona as having faster fp throughput than 3.2 GHz Core 2 depends wholly on which types of calculations you are doing. spec_fp does calculations that are memory bound, LinPack does not (at least not as much). Barcelona's faster fp throughput is not due to markedly superior fp unit (though it may be marginally better) but its onboard memory controller. If you need that sort of thing, great, go with barcelona. If you need raw speed on smaller units (under a couple of megabytes) chances are good that the higher clocked Core 2 with huge cache will win.

    --
    Education is a better safeguard of liberty than a standing army.
    Edward Everett (1794 - 1865)
  33. Oracle charges per CPU for up to 2-CPU servers by rklrkl · · Score: 1
    Seems to me you're looking at old Oracle pricing w.r.t. cores - back in February this year, Oracle revised it because they were getting hammered by competitors' per-socket licensing.


    If you have 1 or 2 CPUs (i.e. you typically run "Standard Edition One" Oracle which limits you to 2 CPU sockets anyway), you are only charged per CPU *socket*, regardless of the number of cores per CPU - here's the details from the horse's mouth.

    So a quad core single CPU server will set you back about 3,000 pounds + VAT in the UK in Oracle licencing (note: you've infuriatingly got to add 22% annual updates/support onto the base price - this is a con...you can't just buy annual updates with no support any more [used to be only 6% of purchase price...]).

    Oh and in the first year, you can go to an Oracle reseller, who will typically roughly discount the updates/support (about 20%) and that discount is carried through to later years [which slightly strangely are handled by Oracle, not the reseller...].

  34. A little more complicated by lenehey · · Score: 1

    Its a little more complicated than that because with smaller features, your are susceptible to smaller defects. That is, a defect that is likely not to affect AMD's chip at 65 nm can obliterate Intel's 45 nm chip. So the likelihood of a defect isn't a linear relationship to area as you suggest.

    1. Re:A little more complicated by Mr+Z · · Score: 1

      Fair enough, but in terms of dollars of revenue per wafer, though, the relative cost of a given defect is generally smaller on a 45nm wafer than a 65nm wafer if the 45nm design is roughly 1/2 the size of the 65nm design. You've taken out a smaller percentage of the devices on the wafer. Note that I say "relative cost." 45nm wafers are still more expensive than 65nm wafers. :-)

      Also, with a RAM-heavy design, you can build significant redundancy into your RAM arrays and perform RAM repair (remap columns or rows) to buy back some yield. Thus, you have yet another bulwark against these defects. As asliarun pointed out, Intel favors RAM heavy designs. This could be part of the reason.

      It'll be interesting to see how this all plays out. Right now Intel probably does have to eat a bit more cost on the 45nm process as it matures. The raw cost per wafer is probably higher, and the yields not as good, just as a function of immature process. But, that's a relatively fixed expense that will be amortized over the 100s of millions of x86s they'll pump out, so the actual cost per device over the lifetime of that design will be minimal. They've got the volume to support it. Developing a process isn't cheap though. Even my employer, TI, is going to outside foundries for its high speed 45nm logic process.

      --
      Program Intellivision!