Dual 1.8gigs by end '03?

245

Comments

  • Reply 21 of 85
    xypexype Posts: 672member
    [quote]Originally posted by Programmer:

    <strong>Everybody is going to have to come up with new performance measurements, including Intel.</strong><hr></blockquote>



    Ok, now I can't resist. Watts. Just like the audio freaks. Intel will go for Watt rating, I'm sure. "Now running at 100 Watt, 2.5 x more than Apple CPUs, that is"



    I thinkthe percycle rating that engineers would like to see would be a bit demanding for the users to grasp. If Apple will keep the 6month update cycles they might as well define a measure standard as in "970 1.5 ghz = 1" and go on from there.



    So they can start with 1 and go on with 1.2, 1.5, 2.0, etc. Much better than having 2800+ ratings (what happend once 10000 is passed? would it go 11400?). I think a "x-times faster than a UberMac (for example a 1.5ghz 970)" would be a rating customers would like. If the UberMac proves faster than x86 machines the opinion will stick - else people will assume Macs are slower than x86 anyway.
  • Reply 22 of 85
    kecksykecksy Posts: 1,002member
    I just realized something. Once the PowerPC 970 is released, PowerMacs might be faster than Intel, but what about Apple's other lines?



    How long will it take for the PPC970 to trickle down to the PowerBook, iMac, and eMac lines? I'm sure the PowerBook will get it as soon as IBM moves to 0.09-microns, but I don't think it's unreasonable to expect the iMac to still be using the G4 18 months down the road. Now Motorola could increase clock speeds, but I know it won't be enough to make the iMac compare favorably to similarly priced Windows machines, unless Apple decides to give the iMac dual processors.



    Of course we all know the iBook won't be getting a PPC970 until the PPC1070 is out, so let's not even go there. What do you think, will Apple rise to the occasion and make it a point to bring all their lines up to date, or will they just brag about the speed of machines which the average consumer can't afford and let their other lines stagnate?



    To be fair, there is a good reason why the iMac hasn't been updated in over a year. There is just no room to bump up the megahertz right now with the low end PowerMac being just 67MHz faster. I know the PowerMac has two processors, but I don't blame Apple for holding back a little. Still, I don't think announcing 1GHz iMacs at MacWorld New York would have hurt PowerMac sales if both updates were announced side by side. Faced with the choice between 1GHz and Dual 867MHz, most professionals wouldn't hesitate to by the PowerMac, so I don't know what Apple is so worried about here?



    Let's all hope the company is smart enough to make the the transition from Motorola to IBM as quickly as possible. Stick dual 1GHz G4s in an iBook, not an iMac.
  • Reply 23 of 85
    Once again you guys are comparing 32-bit CPU speeds to 64-bit CPU speeds in terms of MHz/GHz. The fastest 64-bit CPU out there is barely over 1GHz right now. Come July, a PPC970 @ 1.8GHz should fly. You can't compare a 3.06GHz 32-bit x86 CPU to a 1.8GHz 64-bit PPC CPU by GHz alone.
  • Reply 24 of 85
    ast3r3xast3r3x Posts: 5,012member
    what does it really matter about the other product lines other then Power*. Power* is what is important to argue speed, but it doesn't matter. If your computer does everything you need it to and doesn't give u lag then your good.



    You could have a 350GHz processor and it won't open iTunes faster then a 125GHz processor. What applications need speed in a consumer machine...Power* is for professional i* is for consumers who use email, internet and iTunes...iMovie and iDVD shouldn't need more then a 1 or 1.25GHz G4 in the future..specially with increased ram and bus speed
  • Reply 24 of 85
    kecksykecksy Posts: 1,002member
    [quote]Originally posted by Crouton:

    <strong>Once again you guys are comparing 32-bit CPU speeds to 64-bit CPU speeds in terms of MHz/GHz. The fastest 64-bit CPU out there is barely over 1GHz right now. Come July, a PPC970 @ 1.8GHz should fly. You can't compare a 3.06GHz 32-bit x86 CPU to a 1.8GHz 64-bit PPC CPU by GHz alone.</strong><hr></blockquote>



    Yes you can, because 64-bits won't mean jack, as most software will remain 32-bits for some time. The only place where it would matter is in enterprise and scientific research. Of course, you can tell consumers they need 64-bits, but that's blatant lying, which I wouldn't put it past Apple or AMD by the way. They're not going to fool me though.
  • Reply 26 of 85
    kecksykecksy Posts: 1,002member
    [quote]Originally posted by ast3r3x:

    <strong>what does it really matter about the other product lines other then Power*. Power* is what is important to argue speed, but it doesn't matter. If your computer does everything you need it to and doesn't give u lag then your good.



    You could have a 350GHz processor and it won't open iTunes faster then a 125GHz processor. What applications need speed in a consumer machine...Power* is for professional i* is for consumers who use email, internet and iTunes...iMovie and iDVD shouldn't need more then a 1 or 1.25GHz G4 in the future..specially with increased ram and bus speed</strong><hr></blockquote>



    I would agree with you, but most people are tricked into thinking they need 2.4GHz whenever they go out to buy a PC, and I don't know, I would think a fast processor would help when exporting movies and encoding DVDs. For email and the Internet, a single processor G4 is fine, but the digital hub would be much nicer on faster machines.
  • Reply 27 of 85
    jlljll Posts: 2,713member
    [quote]Originally posted by ast3r3x:

    <strong>what does it really matter about the other product lines other then Power*. Power* is what is important to argue speed, but it doesn't matter. If your computer does everything you need it to and doesn't give u lag then your good.



    You could have a 350GHz processor and it won't open iTunes faster then a 125GHz processor. What applications need speed in a consumer machine...Power* is for professional i* is for consumers who use email, internet and iTunes...iMovie and iDVD shouldn't need more then a 1 or 1.25GHz G4 in the future..specially with increased ram and bus speed</strong><hr></blockquote>





    And noone ever makes new apps that require more power
  • Reply 28 of 85
    xypexype Posts: 672member
    [quote]Originally posted by JLL:

    <strong>And noone ever makes new apps that require more power </strong><hr></blockquote>



    The geenie effect can be turned off..
  • Reply 29 of 85
    jdbonjdbon Posts: 109member
    [quote]Originally posted by Paul:

    <strong>



    yeah, and while you are at it, why not include a 23" touch screen cinema display and price the whole package at $299 so everyone can afford one... <img src="graemlins/oyvey.gif" border="0" alt="[No]" />



    (im not saying that apple shouldn't provide more "bang for the buck, but a lot of what you depict sounds pretty impossible... )</strong><hr></blockquote>



    Well I do not think what I depcited is impossible. Front ports,digital audio,svideo out, those are things that most pcs in the powermac's price range include. They are not exotic impossible features. A quieter case is very possible, along with more expansion bays. I don't understand why people may think such features are impossible. Ram prices are falling, and superdrives are coming down as well. There seems to be a mindset that including features present on PCs is "giving in" to th dark side. Apple's design is superior to PCs in many ways, but their are features present on PCs that Macs do not have, and that should be included.
  • Reply 30 of 85
    amorphamorph Posts: 7,112member
    [quote]Originally posted by Kecksy:

    <strong>How long will it take for the PPC970 to trickle down to the PowerBook, iMac, and eMac lines? I'm sure the PowerBook will get it as soon as IBM moves to 0.09-microns, but I don't think it's unreasonable to expect the iMac to still be using the G4 18 months down the road.</strong><hr></blockquote>



    Actually, the iMac tends to follow the PowerBook. I would expect a 970-based iMac not long after the jump to 90 nm. The iBook would retain a G4, but it would hardly be a slacker. It might even be on 90nm itself at that point - tiny, cool, cheap, and fast for a sub-$1K laptop.



    [quote]<strong>Now Motorola could increase clock speeds, but I know it won't be enough to make the iMac compare favorably to similarly priced Windows machines, unless Apple decides to give the iMac dual processors.</strong><hr></blockquote>



    I don't see iMacs going DP, even though I personally think it would be a great idea - OS X runs so sweetly on DP machines. The clockspeed won't catch up, but Intel is taking the next year off to work on power consumption issues, so Mot and IBM - who already have efficient architectures - can regain some ground. I think as long as Apple is moving they'll be fine. Apple needs people to actually sit down and play with their machines, and once they do it shouldn't take long for them not to care what the chip's clock rating is - if the performance and the user experience is there.



    [quote]<strong>What do you think, will Apple rise to the occasion and make it a point to bring all their lines up to date, or will they just brag about the speed of machines which the average consumer can't afford and let their other lines stagnate?</strong><hr></blockquote>



    Based on what I see now, Apple will be moving aggressively. iBooks are now $999, and $899 for edu customers. PowerBooks are now serious lust-inducing machines at a competitive price point, and the top-of-the-line is a good value for the first time in a good while.



    [quote]<strong>To be fair, there is a good reason why the iMac hasn't been updated in over a year. There is just no room to bump up the megahertz right now with the low end PowerMac being just 67MHz faster.</strong><hr></blockquote>



    Actually, I don't think Apple cares about that. The iMac hasn't been updated in a while because it's getting overhauled. The 15" sold very well, then dive-bombed. The 17" has been carrying the line. I think Apple's busily scrambling to rework the machine in response to this. And I think they're busy on the eMac as well, since that's proven to be a troublesome machine. I don't think the iBook will be a clearly superior budget machine to any of their desktops for very long, let's put it that way.



    [quote]<strong>Let's all hope the company is smart enough to make the the transition from Motorola to IBM as quickly as possible. Stick dual 1GHz G4s in an iBook, not an iMac.</strong><hr></blockquote>



    Let's hope Apple is smart enough to use the best chip for the job. If it's from IBM, cool. If it's from Mot, cool. Starting at 90nm, Mot will have gotten around their achilles heel - dirty fabs - by shifting high-end production to a joint Mot/Philips/STM venture. That should allow their strengths - CPU design and laboratory fab technology - to shine through, and give Apple some more options.



    I don't want Apple 100% reliant on any one supplier. That's just bad news.



    [ 12-07-2002: Message edited by: Amorph ]</p>
  • Reply 31 of 85
    progmacprogmac Posts: 1,850member
    i think something overlooked in the fantasizing about the 970 is that while it may meet the speed of a 3.x Ghz P4 at the end of the year, all it means is Apple has caught up with speed to the PC world at that particular time. Intel certainly has something in development, that, when released, will put the P4 and 970 to shame, and then the whole cycle starts over...
  • Reply 32 of 85
    ast3r3xast3r3x Posts: 5,012member
    what makes you think they have something big planned? they need to work on keeping their processor from getting hot enough to melt the solder around the chip.



    IBM has a nice efficient processor in the works and they will catch up and be on board to compete head to head.



    Of course Intel has teh P5 or whatever in the works, but so does IBM, they have the 1020 or whatever.



    "In a race you can't get first place without passing second"

    -Forget who said it haha
  • Reply 33 of 85
    [quote]Originally posted by xype:

    <strong>



    were the estimated 970 speeds actually based on .13 or .09 sizes?



    i really would not be surprised if the 970 would go over 2.0 ghz in 2003 - at least it would surprise a few people Apple-sytle and jobs would have his "hah!" effect </strong><hr></blockquote>



    [quote] from IBM's press release: <a href="http://www-916.ibm.com/press/prnews.nsf/jan/1713CBA3CD35301085256C52004E3696"; target="_blank">http://www-916.ibm.com/press/prnews.nsf/jan/1713CBA3CD35301085256C52004E3696</a>;

    . . .IBM plans to build the chip in its new state-of-the-art 300mm manufacturing facility here using leading-edge manufacturing technologies. IBM plans to pack performance and new features into the chip using ultra-thin 0.13-micron circuitry (nearly 800 times thinner than a human hair), constructed of copper wiring and about 52 million transistors based on IBM's efficient silicon-on-insulator (SOI) technology. Additional details on the PowerPC 970 are to be disclosed by IBM this week in a paper presented at Microprocessor Forum, a chip design conference organized by industry analyst firm In-Stat/MDR.

    <hr></blockquote>



    If I recall correclty from this and other reports that I can find at the moment, the 970 will start out using the 0.13 process with speeds up to 1.8 Ghz, and rapidly move to the 0.09 process, which would logically improve the speed around 2.2-2.3 Ghz, given a 20% speed gain on the die shrink, and/or reduce power consumption to a level that it will be able to be used in the PowerBook line-up.
  • Reply 34 of 85
    I am here with a personal request aimed at all applicable persons.



    Please refrain from using the term "megahertz myth" to defend apple/ibm/motorola in the present, as if the current line of powermacs were in any way faster than current intel/x86 processors in all of the important calculations that really count. You look stupid to me.



    When you bring up that dastardly term, it sends a pang of grief through me. If you are truly intending to do a good thing for Apple, please don't misrepresent their products as superior in the respect of speed. Saying that wretched Myth is the only thing that makes people stick to PCs is an insult to Apple, because its defenders are of low intelligence and considerable ignorance when fact is concerned. And that means that Apple must be a company of brainwashers. It's not good to support brainwashers.



    PCs have more hardware and software variations. PCs have better framerates and overall graphics performance. PCs can render 3d graphics faster. PCs have hundreds of games. PCs are understood by a larger group of people.



    Macs are not faster computers overall. They are not cheaper to own. They do not outperform PCs. Please stop saying that 2 1.25ghz PowerPC G4s are actually faster than, or as fast as, a single 3.06ghz Pentium 4. The MHZ really do matter, whether you want to deem it as a falsehood or not. You know this is true, deep in the back of your mind. You know that when you comment about the "MHZ Myth," you are essentially suggesting to those people you talk to, that "It's really Only a myth, and it's not at all true that the Intel processors are faster. Apple is faster." And this makes you stupid, because you know it's wrong and you say it anyway, whether intentionally meaning to mislead and be loyal to apple (in an unconscious, backstabbing kind of way), or by sheer accident of ignorance.



    This is my plea for the day. I hope noone takes it personally; it's my opinion, and though only an opinion, I feel that for you to think about it for yourself would be a good thing for you. You may learn something new about your own views. If you throw it by the wayside, then that would say something about your being less open-minded than perhaps you ought to be, for optimal interpretation of your world, and in consequence, for your own good. But it's not just because I say it's that way that it really is that way; I'm only interpreting the world that I see.
  • Reply 35 of 85
    ast3r3xast3r3x Posts: 5,012member
    I am in no way saing g4's are faster then P5's, but you can't say that MHz is the only thing that matters, there is aloto take into effect, but at a certin point it doesn't matter if something of more efficient, if the thing that is less efficient can do it 2x as fast then it more then makes up for it



    but...



    Long Live Apple (apple may not be a cult but AI supporting apple is )
  • Reply 36 of 85
    jdbonjdbon Posts: 109member
    lemmingway your response may be a bit harsh, but you correct in saying that presently, the P4 is faster than the G4. I think we can all agreee that mhz for mhz, the g4 would outperform the p4 at equal clockspeeds. troubl is, the P4 never ran at 1.25ghz, I think the slowest clockspeed for the p4 was 1.3ghz, and that two years ago.



    Focusing on the short term, i think there is little to no chance that 970 machines will be realeased this january though hope springs eternal). Hopefully we'll get dual 1.4ghz, maybe 1.6 if we are very lucky!





    The G4 as a Pro processor is on its way out. We all knwo this. However Apple can;t snap its fingers and have the ship of its dreams ready tomorrow.





    Apple still has the opportunity to release a substantial upgrade to its powermacs in January. My ideas:



    AGP 8x with two slots (AGP 8x supports more than one slot).





    Serial ATA only (IDE adapters are cheap)



    USB2 Firewire 2



    Digital audio in and out.



    Radeon 9700



    New Case ala Next cube,really quiet! (new cases make you feel like you are getting something completely new!)



    Maybe a ridiculously expensive Quad option on the high end?





    These ideas are not impossible! They may not happen because of Apple's economic model (squeeze as much profit out of its machines as possible), but they would spur sales,IMHO.
  • Reply 37 of 85
    amorphamorph Posts: 7,112member
    The MHz Myth states nothing other than that MHz do not compare across processor lines. For proof, compare performance per clock between the Intel Pentium 4 and the Intel Itanium.



    No-one disputes that, except for tasks like RC5 where AltiVec kicks in, the G4 is currently lagging in performance. <a href="http://www.aceshardware.com/read.jsp?id=50000336"; target="_blank">Two of them</a> hold their own against a 3.06 GHz P4 in Lightwave rendering (notice: I said "hold their own against," not "beat"). The article I linked to blames the fact that the G4 is fabbed on an older process, which is fair enough. That problem gets fixed in about a month.



    However, if the G4 is not behind the P4 by a ratio equal to the ratio of their clockspeeds, the MHz Myth argument holds. And guess what? It's not.



    [ 12-07-2002: Message edited by: Amorph ]</p>
  • Reply 38 of 85
    matsumatsu Posts: 6,558member
    Intriguing, while I don't think that anyone who really knows whats coming would be foolish enough to divulge, I think there are a few reasonable voices, some with genuine info, some with their ears to the right sources. Voices you can trust to at least be close to reality, like Amorph.



    With that in mind, I find it intersesting that you say the G4 fab issue will get resolved in about 1 month. That points to .13u (.09u would be a dream, but not likely for 2003 even if it fits the embedded target market beautifully, lets them get higher yields, cheaper parts, yadda yadda, we're still talking moto here!). Elsewhere, upgrade makers have stated that they will soon supply a DP 1.2Ghz upgrade that drains less power than a DP 800. This could come from re-jigging the card layout/design, but that seems like a stretch, I dunno a lot about this, but I imagine about the only way you could put in 50% faster chips and drain less power is with a smaller process, ergo .13u 7457's, we hope. This chips is supposed to weigh in with a 200MHz FSB, which is still behind, but getting better, and it should let increased clock speeds at least continue to scale linearly (in terms of performance)



    And for me at least, it raises another intriguing possibility -- laptop updates! .13 G4's, especially if they're a drop in replacement for 7455 chips, would at least make a Powerbook update possible, possibly even the appearance of a G4 iBook -- high-end only, between the G3 iBook and the PB, something to explain the new (lower) iBook prices? Suddenly the prospect of notebook update a mere three months apart is not as outlandish as it at first may have seemed.
  • Reply 39 of 85
    A single G4 is 1.25, about 5/12 of a single P4, at 3.06. With 2 processors, fine. 10/12 of a single P4, which is almost exactly in tune with the MHZ deficit, when that lightwave 7.5 benchmark is compared. 89 vs 102. Darn close; you'd think altivec didn't hardly help. And if such a "high" speed was because of Altivec, imagine what it would be without it. In tracer with radiosity benchmarks, the gap widens a bit. The PowerPC G4 Dual 1250 gets 597 seconds to the Pentium 4 Single 3060's 395 seconds. The mac is roughly 50% slower, resulting in about 1.5X the amount of rendering time. In large projects this makes a huge difference. 10 minutes becomes 15. 2 hours becomes 3.



    Although not all of the deficits are as big (and some not as small), it is still behind in almost everything.



    The initial topic of this thread doesn't really seem like it's impressive; it's saying that, assuming the 970 is what Apple uses (there still has been no confirmation), by the end of next year Apple will only be Half as slow MHZ-wise.



    I remember reading about those 970 processors being potentially very expensive, and warranting Apple's raising the prices and considering computers based on these processors to be more of a Workstation class (So it will be more in competition with the Xeon than the standard P4). Accompanied by this rumor in my memory is the suggestion of a mainstream, less powerful processor being the primary competitor against things like present-day P4s and Athlons. This isn't encouraging to me; Apple doesn't even have a comparable processor to the standard, consumer-class X86 processors.
  • Reply 40 of 85
    amorphamorph Posts: 7,112member
    [quote]Originally posted by Lemmingway:

    <strong>The initial topic of this thread doesn't really seem like it's impressive; it's saying that, assuming the 970 is what Apple uses (there still has been no confirmation), by the end of next year Apple will only be Half as slow MHZ-wise.</strong><hr></blockquote>



    But the 970 isn't the G4. It's derived from a processor core that, with an approximately 30% performance penalty resulting from a more rugged construction, is right up with the fastest P4 in integer, and faster at FP. And that's running at 1 GHz, with no AltiVec.



    The 970 won't have the more rugged construction. It'll be engineered like most mass-market processors (including Intel's). It's going to have AltiVec. It's going to have three times the bandwidth to and from the processor (to each processor in MP setups) than the G4 does to all processors, and fast, direct busses between processors in MP setups so that they can talk to each other directly. By contrast, multiple G4s share one bus to RAM, and use that bus to communicate with each other. So there's every indication that a 1.8GHz 970 will smoke a 1.8GHz G4 (that MHz Myth again ) and that it will scale to SMP setups much more effectively than the G4 does (and really, the G4 does amazingly well in dual-processor setups, considering).



    Also, 1.8GHz is where it'll start. It's by no means the highest speed the design will be capable of.



    [quote]<strong>I remember reading about those 970 processors being potentially very expensive, and warranting Apple's raising the prices and considering computers based on these processors to be more of a Workstation class (So it will be more in competition with the Xeon than the standard P4).</strong><hr></blockquote>



    I do too. That was before IBM revealed it at MPF and we got so see how small it was! It's bigger than a G4, but smaller than a P4, and die size correlates directly to cost. Given that it will have a pretty decent market right off the starting line (IBM's popular RS/6000 line, at the very least) and given that it will migrate relatively quickly to a 90nm process (meaning that it will get smaller, and therefor cheaper still) it should be far less expensive than Intel's Xeon, or AMD's Opteron, and most likely cheaper than Intel's P4. For its part, the G4 will shrink to a miniscule 55 square mm on 130 nm, at which point Apple should be able to buy them for pocket change.



    [quote]<strong>Accompanied by this rumor in my memory is the suggestion of a mainstream, less powerful processor being the primary competitor against things like present-day P4s and Athlons. This isn't encouraging to me; Apple doesn't even have a comparable processor to the standard, consumer-class X86 processors.</strong><hr></blockquote>



    That's also pre-MPF. Now it looks like the 970 could be that processor, coming just as the G4 gets small, cool and cheap enough to take over for the G3. And one poster who has a few contacts here and there has intimated that someone is holding something back, so either IBM or Mot (or, outside chance, someone outside AIM) has a surprise for us. I have read that since the issues involved in migrating from 180nm to 130nm are the same as the ones involved in migrating from 180nm to 90nm, and since Mot's own 130nm plant has been having problems (something to do with the fans not running, I'm sure...) that Mot was just going to make the big leap once the plant they've partnered with goes online. Maybe so. And if so, they could unleash a processor as hot and monstrous as the Mot G5 was rumored to be on such a small process that it would actually be quite reasonable.



    Oh, and the core the 970 is derived from belongs to the POWER4, which ships with four of 'em on die. So if the design is relatively faithful to its parent's, IBM could easily roll out a dual-core variant of the 970, too. When they went to 90nm, say.



    [ 12-07-2002: Message edited by: Amorph ]</p>
Sign In or Register to comment.