And now you know why I asked :)

by Anand Lal Shimpi on 4/4/2005 2:54 PM EST
Comments Locked

26 Comments

Back to Article

  • blackjack - Wednesday, June 8, 2005 - link

    Please visit the pages about blackjack http://www.casino7-online.com/blackjack.html
    - Tons of interesdting stuff!!!
  • party poker - Wednesday, June 8, 2005 - link

    You are invited to check some helpful info dedicated to online poker http://www.poker-4-u.com/online-poker.html
    ...
  • poker games - Tuesday, June 7, 2005 - link

    Please check the pages about poker games http://www.poker-7.com/poker-games.html
    party poker http://www.poker-7.com/party-poker.html
    poker rooms http://www.poker-7.com/poker-rooms.html
    texas holdem http://www.poker-7.com/texas-holdem.html
    empire poker http://www.poker-7.com/empire-poker.html
    - Tons of interesdting stuff!!!
  • online casino - Monday, June 6, 2005 - link

    You can also check some relevant pages about online casino http://www.djsdesigns.net/online-casino.html
    online casinos site http://www.djsdesigns.net/online-casinos.html
    slot machines site http://www.djsdesigns.net/slot-machines.html
    casino on net site http://www.djsdesigns.net/casino-on-net.html
    casino games site http://www.djsdesigns.net/casino-games.html
    casino gambling site http://www.djsdesigns.net/casino-gambling.html
    internet casino site http://www.djsdesigns.net/internet-casino.html
    online casino gambling site http://www.djsdesigns.net/online-casino-gambling.h...
    casino online site http://www.djsdesigns.net/casino-online.html
    roulette site http://www.djsdesigns.net/roulette.html
    - Tons of interesdting stuff!!!
  • Heron Kusanagi - Wednesday, April 6, 2005 - link

    Thanks Anand. Always nice to know that AMD can bring out good products.

    Also, after seeing this dual core, what do you think Yonah can bring to the mobile segment?
  • HammerFan - Tuesday, April 5, 2005 - link

    I'd be willing to bet that Dual Core with HT will sometimes lose to dual core non-HT because of the amount of resources available. Maybe four logical cores sharing the same FSB bandwidth/HDD/memory etc. is less efficient than two single-threaded cores using the same sized bus etc. In cases where the FSB isn't nearly saturated on two cores w/o HT, there would probably be a benefit from dual core with HT
  • Anand Lal Shimpi - Tuesday, April 5, 2005 - link

    Heron Kusanagi

    1) Dual Core with HT won't necessarily always lose to dual core with HT. Remember that a single core with HT usually gains performance with HT enabled, so there's no reason that two cores can't also get a boost from HT being enabled - assuming that the physical cores are scheduled first as the logical HT cores are scheduled next. This may take a bit of OS tuning to get right.

    2) AMD has already outlined for me their dual core desktop strategy, unfortunately it will be some months before I can talk about it. Performance wise, I think AMD's dual core offerings will wipe the floor with Intel's in performance. AMD's traditionally weaker encoding performance will get a healthy boost from dual core, but I think Intel will still be ahead in that respect. Obviously gaming and all other performance aspects will be stronger on the AMD platform. The major problem will be pricing. Intel is going to play the role of making sure that the masses have dual core chips, because honestly they have the ability to manufacture these massive chips pretty well.

    There's a very good reason for AMD to go after the workstation market first with dual core - you have to make fewer chips for the workstation market than you do the desktop market, so if you can't make that many to begin with... You get the point :) Intel has done very similar things in the past, often times they would debut a new manufacturing process on mobile chips simply because the benefits of a cooler process were better realized in mobile chips and, they didn't have to make as many chips.

    In the long run, this pricing argument won't matter because eventually all dual core chips will be available at mainstream price points. If you're buying in 2005 I'd say that Intel will be the more affordable dual core option; if you're buying in '06 however, AMD is much more competitive from a pricing standpoint.

    Take care,
    Anand
  • Heron Kusanagi - Tuesday, April 5, 2005 - link

    Thanks #18 for answering my 1st question...
  • Michael2k - Tuesday, April 5, 2005 - link

    #17: Actually, let me take a stab at this:

    1) Hyperthreading helps hide stalls in the pipelines. Video encoding doesn't benefit from HT because it is a continuous stream of data being processed. All additional threads would do is interrupt the encoding process. Here's an example:

    Imagine you're trying to translate a book (analogy to encoding a video) from English to German. In a non HT system, you translate from front to back in one more or less continuous chunk. You get it done in 4 days.

    In a HT system, you split the book up into 4 sections of 4 chapters. Every three hours you switch chapters, so that the first day you work on all four sections, but you never finish any of them. Because you are now task switching, you actually take 5 days to complete.
  • Heron Kusanagi - Tuesday, April 5, 2005 - link

    A very good review there is. Very informative. It's also good to know that Intel's Dual Core can live up to the hype. But I got a couple of questions...

    1) In your reasoning, in most Video Encoding tests, Dual Core with HT will lose to Dual Core without HT? Since video encoding is multithreaded, more threads should benefit the times.

    2) Granted that AMD has not come out with their dual cores, but as you stated in the review, AMD will benefit more from Intel going Dual Core. So what exactly can a consumer expect from AMD and how you think will size up against Intel's offerings?
  • Anand Lal Shimpi - Tuesday, April 5, 2005 - link

    I've talked about ethics quite a bit in the past, but it usually goes much further than basic rules and extends quite a bit into my personality. I will take your suggestion to heart however and see if I can put something together that's useful. I do agree that transparency is the best policy, which is why I have no problems answering questions of this nature if they're brought to my attention.

    Take care,
    Anand
  • Ara - Tuesday, April 5, 2005 - link

    I forgot, I meant to mention this before:

    You may consider, to put any future questions of ethics to rest, publishing your policies when it comes to dealing with companies (perhaps a link from the About page).

    There's just so many shady business going on with media both onlne and off (for example getting mentioned favorably on just about any fashion magazine is either dumb luck or a direct result of how good a PR firm you can hire), that I for one would really like to see an offical policy on every media site.

    For example: "We do not attend company-paid junkets." "We do keep review units of any kind." And so on as it applies to the site.

    When it comes to things like this, it seems to me anyway, that transparency is the best policy.

    Regards,
    Ara
  • Anand Lal Shimpi - Tuesday, April 5, 2005 - link

    That's like apologizing to a garbage man for having a full trash can; it's my job :)

    That being said, I've got a long night ahead of me. Of course Intel had to spring this dual core preview on us the one weekend out of the year when we lose an hour :)

    Take care,
    Anand
  • Ara - Tuesday, April 5, 2005 - link

    No. Thanks for the quick turnaround! Didn't mean to make your day any busier than I'm sure it already is.

    Regards,
    Ara
  • Anand Lal Shimpi - Tuesday, April 5, 2005 - link

    77.0 fps at 10x7, high quality, with an ATI Radeon X850 XT. Looks to me that it's inline our previous numbers.

    Note that a Pentium 4 530 (3.0GHz) with a X800 XT gets 76.2 fps, so the numbers appear to jive.

    Is there anything else you'd like me to look into?

    Take care,
    Anand
  • Anand Lal Shimpi - Tuesday, April 5, 2005 - link

    Ara,

    I see no reason why that would be true. I'm about 10 minutes away from running our Doom 3 test on the Pentium D 2.8GHz for Part II, I'll post the result here for you to look at as soon as I've got it.

    Take care,
    Anand
  • Ara - Tuesday, April 5, 2005 - link

    I didn't mean to accuse as much as to inquire.

    I've been hearing that there's a possibility that these chips could actually be slightly slower than single-core chips of the same frequency. It would have been nice to see one or two tests just to see (early on) if this were true or not. I'm sure you'll cover this in the full review.
  • Michael2k - Tuesday, April 5, 2005 - link

    #6 Anand has it right. Until you can separate sound, graphics, AI, monsters, and interaction into separate threads (you can't, any more than you can separate your eyes, your ears, your hands, and your brain), you won't see much benefit from dual core systems in games.

    You can find SOME advantages, such as rendering every other frame on both CPUs, sometimes, or by offloading some complicated processing to the second CPU (treating it as a GPU or advanced compute unit) while the first CPU does other things, but that is strictly dual core, and the game has to be written for such things.

    Who knows, maybe someone will be the next Carmack and create an awesome dual processor game, like he did with OpenGL and the first Quakeworld to use hardware accelerated graphics. Maybe someone will figure out how to make CPU accelerate games.
  • bersl2 - Tuesday, April 5, 2005 - link

    #6: This is the price to be paid when a majority of the readership doesn't care about what's better for them but only want bigger and faster. Then later they complain when things aren't delivered as promised, until a company releases its plans for its next generation of hardware which will fix all the problems with the current hardware, double their performance, etc.; then they go back to being numbers sheep.

    Secrecy, from a user's perspective, is nothing but trouble. Secrecy means that the product cannot stand up on its own accord; or that there's some super-secret technology involved which has to be licensed, driving up the price for the retail product; or some other similarly foolish reason.

    This isn't the only area where secrecy from hardware companies is making users' lives hell. For instance, have you ever tried asking chipset makers for the technical documentation for their chips? And I'm not talking about schematics; I'm talking about a text that will tell a developer what values to put in what memory-mapped I/O registers to make the damn thing work. Will they let you have the minimum documentation available to make your own driver? In most cases, "Of course not! That's secret information!" at best, silence at worst.

    It's not your fight, but it's a perfect example of why the hardware industry is an industry, with all the shady politics that accompanies such a label. Every company is battling both every other company and its patrons/followers for economic supremacy, at the direct expense of fact, truth, and quality. A good value never causes the company coffers to overflow, so it is up to the consumer to constantly force them to offer better products.

    (rant off)
  • Anand Lal Shimpi - Tuesday, April 5, 2005 - link

    The lack of gaming benchmarks was addressed in the review:

    "Given the short lead time on hardware for this review, we left out all of our single threaded benchmarks given that we can already tell you what performance is like under those tests - so if you're looking for performance under PC WorldBench or any of our Game tests, take a look at our older reviews and look at the performance of the Pentium 4 530 to get an idea of where these dual core CPUs will perform in single threaded apps.  There are no surprises here; you could have a 128 core CPU and it would still perform the same in a single threaded application.  Closer to its launch, we will have a full review including all of our single and multithreaded benchmarks so that you may have all of the information that will help determine your buying decision in one place."

    There was no exclusive associated with the dual core article, Intel seeded sites at their own discretion, asked nothing of us and sent a stack of benchmarks. I cast the stack aside and asked the readership what they wanted to see (see my previous blog).

    It's a matter of having a limited amount of time and trying to get as much information into an article as possible. Why spend time re-running tests that we already have done, especially when the number of requests we've had for gaming performance on dual core processors is minimal at best. But, as promised in Part I, there will be follow-on articles with all the benchmarks you could want. I've been working on putting together multitasking gaming tests, none of them made the first article but there will definitely be some in the second one.

    If there are any questions about ethics or integrity I can go into further detail, just ask away, I've got nothing to hide :)

    Take care,
    Anand
  • Ara - Tuesday, April 5, 2005 - link

    This:
    http://www.theinquirer.net/?article=22332

    seems directed (at least in part) at your review.

    Funny how there was not a single gaming test in sight. Seems a rather high journalist price to pay for an exclusive.
  • red and black - Monday, April 4, 2005 - link

    Great article for Windows users. I have to say, though, that my main reaction to the multitasking tests was that Windows failed them.

    I guess it's good that Windows users can finally have systems that don't lock up horribly under load. Less aggregate suffering in the world, and all that.

    I'm looking forward to the dual-care AMD chips!
  • jak - Monday, April 4, 2005 - link

    Just wanted to say that that was a fantastic article. This is the exact kind of data I was looking to find; the performance of dual-core in multitasking situations; the effect of adding an extra processor on day-to-day tasks. The other reviews I read re: dual-core this morning simply pitted dual-core against AMD or Intel's single-core finest in games(!), of all things, and claim that dual-core 'is not there yet' because an 840 can't beat an FX55 in doom3...

    Well done, and thank you for an informative read.
  • ChineseDemocracyGNR - Monday, April 4, 2005 - link

    Talking about AMD... when will we see socket 939 PCI-E motherboards reviews?
  • Ghandiinstinct - Monday, April 4, 2005 - link

    Meh, lets wait for AMD's rival to 840 then let the testing begin = )
  • Michael2k - Monday, April 4, 2005 - link

    Can you run one of those office benchmarks while DVDshrink is running in the backgroun/foreground?

    That would be interesting.

Log in

Don't have an account? Sign up now