New Set of Next-Gen Power Mac Rumors

123468

Comments

  • Reply 101 of 152
    [quote]Originally posted by Programmer:

    But where's the beef, man? You're posting about your source, but we really don't care about that... we want content![/QB]<hr></blockquote>



    Alright Programmer, here's some "beef".



    Deep Mac on the G5 : "Keep waiting, and no the G5 that may be revealed in the next eight months will not be the G5 you think it is."



    Deep Mac on the Workstation Concept and Specs:



    1) CPU Speeds on the first batch of machines were 400/450/500 that came into his lab a few months ago.



    2) Last month the CPU speeds of the machines were 800 mHz. (I think I mentioned this earlier)



    3) He just said flat out if you want to know about the bus, bandwidth, and memory capabilty and compacity he said take Apple's custom chip, marry it to the existing

    RS/6000 44p model 170 and you've pretty much got it.



    Do keep in mind this hardware sounds to me like it is a testbed of some kind, by "testbed" I mean this in the SQA sense of the word. The systems are cobbled together from off the shelf, or at least heavily modified off the shelf components used to test specific custom parts or technologies as a prelude to integration testing.



    Once upon a time I worked at IBM doing OS/2 functional testing for the Personal Software Products group in Charlotte, NC and RTP. One thing I can tell you is that the Apple/IBM relationship has existed since the early 1990's and I do remember the IBM PowerStation PC 1000. IBM punched out a few hundred of these units, and what made them special? They were esentially RS/6000's running a duel 604e processor setup on a CHRP mobo capable of running OS/2 for PowerPC, Windows NT 3.51, AIX, and MacOS 7.x

    I remember as late as 1998 hearing about dev level RS/6000's running the MacOS, I wonder if things haven't changed much since then and Apple uses the RS/6000 (or machines very much like it) as a development platform and these are the machines Deep Mac is getting.

    It could explain all this Apple IBM speculation about the POWER-4. Come to think of it did anyone know that IBM was modifying RS/6000's to run the MacOS a few years ago?

    Anyway wanted to throw that in, since from my perspective Deep Mac isn't telling me anything I hadn't heard to varying degrees before.
  • Reply 102 of 152
    keyboardf12keyboardf12 Posts: 1,379member
    well, the beef was delivered rather quickly. you got to give the guy that.





  • Reply 103 of 152
    keyboardf12keyboardf12 Posts: 1,379member
    anyone want to give a quick wrap up on what the powerpc 4000 can do (mem,bus,speed etc)



    I think this:



    [quote]Deep Mac on the G5 : "Keep waiting, and no the G5 that may be revealed in the next eight months will not be the G5 you think it is." <hr></blockquote>



    means the g5 is going to be an all ibm affair.
  • Reply 104 of 152
    programmerprogrammer Posts: 3,467member
    That machine has a 128-bit 100 MHz 60x bus. 1.6 GB/sec of bandwidth (theoretical). I hope that they at least use a 128-bit 166 MHz MPX bus (which is an extension of the 60x bus so its not that big a change). But big/wide busses just don't make a lot of sense for Apple, IMHO.



    And I hope they get those clock rates up.
  • Reply 105 of 152
    rickagrickag Posts: 1,626member
    Commander Max



    "Deep Mac on the G5 : "Keep waiting, and no the G5 that may be revealed in the next eight months will not be the G5 you think it is.""



    By, "not the G5 you think it is" is that refering to speculation that the G5, for Apple, would contain



    a new or modified G4 core replacing the e500 core,

    with Rapid I/O,

    Ocean Switched fabric,

    PCI X,

    333 MHz DDR FSB



    ala the MPC 8540/MPC8560, or one of the numerous other designs speculated on these and other boards?????
  • Reply 106 of 152
    eupfhoriaeupfhoria Posts: 257member
    [quote]Originally posted by rickag:

    <strong>Commander Max



    "Deep Mac on the G5 : "Keep waiting, and no the G5 that may be revealed in the next eight months will not be the G5 you think it is.""



    By, "not the G5 you think it is" is that refering to speculation that the G5, for Apple, would contain



    a new or modified G4 core replacing the e500 core,

    with Rapid I/O,

    Ocean Switched fabric,

    PCI X,

    333 MHz DDR FSB



    ala the MPC 8540/MPC8560, or one of the numerous other designs speculated on these and other boards?????</strong><hr></blockquote>



    I don't think he knows, I think he posted all he knows about the subject, he is a middle man after all.



    Then again He could prod Deep Mac some more...



    Keep posting! We want to know more!
  • Reply 107 of 152
    lemon bon bonlemon bon bon Posts: 2,383member
    "But where's the beef, man?"



    Y'know, Programmer...that sounded distinctly un-Programmer like...







    We've heard about the 8540 and ..60...the cores on these chips seem for embedded only use and no advance over the G4. no fpu? Gawd.



    So. The 'G5' is an 'altered' G4.



    More mhz, stretch the Apollo pipeline...and you've got the 7500 as opposed to the 7470(?) due for release this August?



    This 'stretch' will be for the mhz. I'm guessing 1.6 - 2 gig? Or if we 'only' get 1.2 gig now...then Jan' San Fran? 1.2 - 1.6 gig. Hmmm.



    But the big thing...surely...will be the Rio Mobo?



    So...the G4 core 'will' (maybe, I mean... ) be modified to work with Rio.



    This is what the Register has been saying.



    My main bone of contention...is that there'll be no modification to the fpu.



    It's the one area of the G4 that needs beefing up. And there doesn't seem to be a rumour pointing to this. Just more mhz and better mobo.



    A mixed bag. But far better than what we've got right now!!!



    The 'G5' 'G5' will be multicore? Well, it better have something interesting considering it's infamous stature...



    Lemon Bon Bon
  • Reply 108 of 152
    matsumatsu Posts: 6,558member
    I'm not so sure that the FPU needs so much beefing up. When you look at performance comparos between the G4 and AMD/Intel it seems that the FPU performance is right where it should be relative to clock speed as the macs tend to be about as much slower as their clock speed difference (a little better sometimes). If they could only up the clock-speed, they'd be pretty good. However, If they're going to continue to languish in the Hz dept, then perhaps multiple FPU's on a single G4 are the way to go. Instead of using the impending die shrink to make the chip a lot faster/deeper piped, it might be made only a little faster, but with a lot more FPU's on it. Maybe 4 Fpu's instead of 1 ??? ANd a FAST FSB to keep them all fed?



    Again I know nothing about this so if I've got the concept/terminology wrong just move on. But it seems like Apple could either just pump a chip up, or take advantage of smaller process to make a chip more parallel ??? I dunno
  • Reply 109 of 152
    lemon bon bonlemon bon bon Posts: 2,383member
    "I'm not so sure that the FPU needs so much beefing up. When you look at performance comparos between the G4 and AMD/Intel it seems that the FPU performance is right where it should be relative to clock speed as the macs tend to be about as much slower as their clock speed difference (a little better sometimes). If they could only up the clock-speed, they'd be pretty good."



    I recall you said this in another thread when I brought this up. (With the proviso that it was an 'interim' measure.) Certainly not something they could rest on for another 9 months!!!



    Your scenario might be about to come true IF we get DDR 333, BUS and mhz at 1.4 Gig style improvements!!! That 'would do'...for 'now'.





    "However, If they're going to continue to languish in the Hz dept, then perhaps multiple FPU's on a single G4 are the way to go."



    I don't see 'us' catching x86 by early next year. I just don't. The reason I bring the fpu up...isn't necessarily for August...but to counter the 'threat' of the Sledgehammer! The Athlon already has 3 fpu AND...I wonder if the Hammer might have more. Least ways...the Hammer is reputed to be able to do 30 % more per clock!



    So...if the 7500 chip goes on Rio...I hope it beefs up the fpu...because despite 'Rio' we'll slip further behind. To get a pipeline stretch...surely they must offer some more efficiencies per clock? eg extra fpu/integer...otherwise we'll be doing less work per cycle? ala early days of Pentium 4!?



    "Instead of using the impending die shrink to make the chip a lot faster/deeper piped, it might be made only a little faster, but with a lot more FPU's on it. Maybe 4 Fpu's instead of 1 ??? ANd a FAST FSB to keep them all fed?

    Again I know nothing about this so if I've got the concept/terminology wrong just move on. But it seems like Apple could either just pump a chip up, or take advantage of smaller process to make a chip more parallel ??? I dunno"



    Yes you do. I agree. 4 fpus it is then. Deal is done. I'm in heaven...imagine if...4 fpus...gurgle. (You said what I wanted to hear...Matsu... )



    ie if apple could 'pump' the g4 to this extent...I wouldn't worry about the G5 at all!



    "Apple: bumping prices, not specs."



    Aint dat duh truth.



    Lemon Bon Bon



    [ 07-28-2002: Message edited by: Lemon Bon Bon ]</p>
  • Reply 110 of 152
    junkyard dawgjunkyard dawg Posts: 2,801member
    I don't think there's a middle man here....I think cmmdr max IS Deep Mac! And I think he doesn't know a whole lot, he's posted all he knows and it's nothing that couldn't be found already on the internet.



    Know what I think? I think Deep Mac is over. No more BEEF.





    Hrrmmm, now that I think of it, I'm going to sign in as a new alias and conjure up some bullsh"t for everyone. Steamin' hot and fresh outta the bull's anus! I'll just repeat some rumors from the net, give myself a cool handle and act like Jobs found me out and canned me! All in the name of spillin' the beans...
  • Reply 111 of 152
    programmerprogrammer Posts: 3,467member
    [quote]Originally posted by Matsu:

    <strong>I'm not so sure that the FPU needs so much beefing up. When you look at performance comparos between the G4 and AMD/Intel it seems that the FPU performance is right where it should be relative to clock speed as the macs tend to be about as much slower as their clock speed difference (a little better sometimes). If they could only up the clock-speed, they'd be pretty good. However, If they're going to continue to languish in the Hz dept, then perhaps multiple FPU's on a single G4 are the way to go. Instead of using the impending die shrink to make the chip a lot faster/deeper piped, it might be made only a little faster, but with a lot more FPU's on it. Maybe 4 Fpu's instead of 1 ??? ANd a FAST FSB to keep them all fed?



    Again I know nothing about this so if I've got the concept/terminology wrong just move on. But it seems like Apple could either just pump a chip up, or take advantage of smaller process to make a chip more parallel ??? I dunno</strong><hr></blockquote>





    Heh, that's funny -- I just wrote the same thing in another thread. Some more random thoughts: the POWER3/4 family has multiple FPUs and turns in terrific floating point numbers at low clock rates. If the new G4 is faster, reduces the memory bottleneck, and has two FPUs then it might be able to trounce AMD's new flashship. The Hammer may have improved the FPU programming model, I'm not sure -- that could offer some of the 30% improvement but would require a recompile.
  • Reply 112 of 152
    jerombajeromba Posts: 357member
    so in the meantime, can Mot and Apple release this in august ? G4+ : around 1.5 Ghz, 13 microns, 166 FSB (333), more FPU, maybe AGP 8X

    Is this will be competitive with PIV and AMD ?

    This will be 100% improvement from the current Dual Gig, no ? and with Jaguar on top, do you think that we can ask for a bundle free .Mac account for 1 year ?



    And for all these rumors about Power4/5, AMD, Nvidia, etc. I don't buy them. Not this year and not in 2003. But a e500 core... MMMhhhh <img src="graemlins/smokin.gif" border="0" alt="[Chilling]" />



    [ 07-28-2002: Message edited by: jeromba ]</p>
  • Reply 113 of 152
    The only person I know at apple just e-mailed me. I've been bugging him for days about those powermac pics, trying to get the specs on them, trying to get details, etc.. and hes been sending me back e-mails that say "I don't know.... NDA.. etc...."



    Until today.



    To-day I got an e-mail from a hotmail (or other free web mail) account, i recognized the last name in the mail address as his. This is what he had to say.



    "Aloha ****,



    Sorry to have been so unfriendly with my e-mails, but heads have been rolling @ infinite loop this week. I don't know about the accuracy of those case pics, but the mlb is certainly familiar. *******(dirty nickname for his supervisor, i guess) told me that there were some nice new hardware enhancements.



    I don't know the new processor speeds. I know they're over 1Ghz, but its been in the air that SJ is planning on shifting from the "hertz" terminology in favor of a less arbitrary speed measurement.



    New processors may be named and rated in terms of gigaflops. for example apple may start calling the dual gigahertz machine, the 15 gigaflop machine. I think the only thing stopping them so far has been the terminology, SJ wants to avoid using the word "flop" in relation to macs.



    They're trying to invent another name for it, something more impressive, but i haven't heard anything. It signals the end of an era, and I guess it means that SJ is considering a surrender in the battle of the megahertz myth, but plans on using the occasion to remind everyone that they're still winning the performance war.



    Well, it'll be interesting, but its more of a marketing ploy than the hardware specs you asked about, and still only a theoretical marketing ploy at that, but its the most interesting news I've got for you.



    See you wednesday.

    -*********************"



    Ok, i got his permission before I posted it here, but he asks you all to be very conscientious in how you read it.
  • Reply 114 of 152
    snofsnof Posts: 98member
    [quote]Originally posted by Junkyard Dawg:

    <strong>Hrrmmm, now that I think of it, I'm going to sign in as a new alias and conjure up some bullsh"t for everyone. Steamin' hot and fresh outta the bull's anus! I'll just repeat some rumors from the net, give myself a cool handle and act like Jobs found me out and canned me! All in the name of spillin' the beans...</strong><hr></blockquote>



    And the funny part is that even after telling us that, people would still fall for it.
  • Reply 115 of 152
    othelloothello Posts: 1,054member
    [quote]Originally posted by Agent Cooper:

    <strong>The only person I know at apple just e-mailed me. I've been bugging him for days about those powermac pics, trying to get the specs on them, trying to get details, etc.. and hes been sending me back e-mails that say "I don't know.... NDA.. etc...."



    Until today.



    To-day I got an e-mail from a hotmail (or other free web mail) account, i recognized the last name in the mail address as his. This is what he had to say.



    "Aloha ****,



    Sorry to have been so unfriendly with my e-mails, but heads have been rolling @ infinite loop this week. I don't know about the accuracy of those case pics, but the mlb is certainly familiar. *******(dirty nickname for his supervisor, i guess) told me that there were some nice new hardware enhancements.



    I don't know the new processor speeds. I know they're over 1Ghz, but its been in the air that SJ is planning on shifting from the "hertz" terminology in favor of a less arbitrary speed measurement.



    New processors may be named and rated in terms of gigaflops. for example apple may start calling the dual gigahertz machine, the 15 gigaflop machine. I think the only thing stopping them so far has been the terminology, SJ wants to avoid using the word "flop" in relation to macs.



    They're trying to invent another name for it, something more impressive, but i haven't heard anything. It signals the end of an era, and I guess it means that SJ is considering a surrender in the battle of the megahertz myth, but plans on using the occasion to remind everyone that they're still winning the performance war.



    Well, it'll be interesting, but its more of a marketing ploy than the hardware specs you asked about, and still only a theoretical marketing ploy at that, but its the most interesting news I've got for you.



    See you wednesday.

    -*********************"



    Ok, i got his permission before I posted it here, but he asks you all to be very conscientious in how you read it.</strong><hr></blockquote>



    the SJ distortion field is kicking in strong...



    <img src="graemlins/lol.gif" border="0" alt="[Laughing]" />
  • Reply 116 of 152
    [quote]Originally posted by Junkyard Dawg:

    <strong>I don't think there's a middle man here....I think cmmdr max IS Deep Mac! And I think he doesn't know a whole lot, he's posted all he knows and it's nothing that couldn't be found already on the internet.



    Know what I think? I think Deep Mac is over. No more BEEF.





    Hrrmmm, now that I think of it, I'm going to sign in as a new alias and conjure up some bullsh"t for everyone. Steamin' hot and fresh outta the bull's anus! I'll just repeat some rumors from the net, give myself a cool handle and act like Jobs found me out and canned me! All in the name of spillin' the beans...</strong><hr></blockquote>



    Surely you jest...



    Anyway read my posts, understand what I'm sharing with you. Nope I'm not the guy spilling the beans, I'm the guy who is posting.

    I think the real value of this thread, as I've speculated before is I believe I've stumbled on a major source of how leaks and rumors have been been consistant yet seldom have been accurate when it comes to finished products.



    So we have Deep Mac, stating he has proof of "Workstation Macs" working on no less than Proof-of-Concept hardware.

    I do not know if you have ever done any professional Quality Assurance, but in the hardware biz it is not uncommon for a computer company to seed development machines with the intention of testing Technologies rather than testing a complete product. So with this mind, if Deep Mac, Dorsal, kormac and who ever else has laid eyes on similar hardware it would explain the consistancies in the postings yet the products in the end fail to materialize. Are they lying? I don't think so. I think they may very well be working on prototype hardware but they are not clued in to the hardware's ultimate purpose. Again, not to test a complete product, but rather test components that may or may not end up in a finished product.

    How does Apple keep their hardware secets safe?

    Simple, no one outside of Apple sees them, at least not the whole product except very late into the development cycle when Apple sends the finished products to specific vendors which have a long history of keeping their mouths shut.

    The bottom line here is, NO ONE has a clue, merely a few tidbits here and there. Aside from Worker Bee, and maybe someone else at Apple that is willing to share dirt, everything you read on the Net and elsewhere is basically bogus.

    But that's not why we're here is it?

    Relax, be entertained, and go from there.
  • Reply 117 of 152
    xypexype Posts: 672member
    [quote]Originally posted by Junkyard Dawg:

    <strong>I don't think there's a middle man here....I think cmmdr max IS Deep Mac! And I think he doesn't know a whole lot, he's posted all he knows and it's nothing that couldn't be found already on the internet. (snip) (snip) (snip) </strong><hr></blockquote>



    Geez, man! You're SO smart! Can I be your friend?
  • Reply 118 of 152
    kidredkidred Posts: 2,402member
    [quote]Originally posted by keyboardf12:

    <strong>anyone want to give a quick wrap up on what the powerpc 4000 can do (mem,bus,speed etc)



    I think this:







    means the g5 is going to be an all ibm affair.</strong><hr></blockquote>



    Well, if the successor to the G4 is called a G5, then yes, it will be an IBM POWER4-core affair.
  • Reply 119 of 152
    [quote]Originally posted by Commander Max:

    [QB]



    Surely you jest...



    &lt;intelligent stuff&gt;

    ---------------------

    No it's true, you really don't have to sacrifice a small goat to get your SCSI chain to terminate properly.



    <hr></blockquote>



    I've been using chickens! &lt;smack-on-the-forehead&gt; umm.. anyone for Buffalo wings?



    I'm of the same sentiment as yourself. There seems to be less and less leaks from Apple. This Armas guy, for example, sounds a little bit like the guy who leaked the <a href="http://www.theregister.co.uk/content/39/25421.html"; target="_blank">stuff to the Register earlier this year,</a> from the "7500" designation to the 2.4 GHz (as Amorph already mentioned).



    Now, the question is whether or not he's legit. I doubt it. The true Apple/Mot employees seem to be clammed up tighter than my first girlfriend. This guy could be some third-party supplier, or an industry "insider" without any direct Apple ties.



    Of course, this is all just my fantasy.



    Anyway, if Deep Mac can give us a Side of Fries with his info, I'd be more than entertained.
  • Reply 120 of 152
    [quote]Originally posted by Commander Max:

    <strong>



    Surely you jest...



    Anyway read my posts, understand what I'm sharing with you. Nope I'm not the guy spilling the beans, I'm the guy who is posting.

    I think the real value of this thread, as I've speculated before is I believe I've stumbled on a major source of how leaks and rumors have been been consistant yet seldom have been accurate when it comes to finished products.



    So we have Deep Mac, stating he has proof of "Workstation Macs" working on no less than Proof-of-Concept hardware.

    I do not know if you have ever done any professional Quality Assurance, but in the hardware biz it is not uncommon for a computer company to seed development machines with the intention of testing Technologies rather than testing a complete product. So with this mind, if Deep Mac, Dorsal, kormac and who ever else has laid eyes on similar hardware it would explain the consistancies in the postings yet the products in the end fail to materialize. Are they lying? I don't think so. I think they may very well be working on prototype hardware but they are not clued in to the hardware's ultimate purpose. Again, not to test a complete product, but rather test components that may or may not end up in a finished product.

    How does Apple keep their hardware secets safe?

    Simple, no one outside of Apple sees them, at least not the whole product except very late into the development cycle when Apple sends the finished products to specific vendors which have a long history of keeping their mouths shut.

    The bottom line here is, NO ONE has a clue, merely a few tidbits here and there. Aside from Worker Bee, and maybe someone else at Apple that is willing to share dirt, everything you read on the Net and elsewhere is basically bogus.

    But that's not why we're here is it?

    Relax, be entertained, and go from there.</strong><hr></blockquote>





    Cool. You sound pretty legit to me. I was just testing you.
Sign In or Register to comment.