The G5 vs the AMD Hammer

Posted:
in Future Apple Hardware edited January 2014
From what I heard, the G5 is a powerful chip, but how will it match up to the Hammer?



I heard the Hammer will have 8Mb of L2 cache, and stuff like that, but as the G5 is a closely guarded secret, I have heard nothing about its specs. Will the G5 just be a failure?

<img src="confused.gif" border="0">
«1

Comments

  • Reply 1 of 37
    philbotphilbot Posts: 240member
    Do you consider the iMac a failure? I can give you 6,000,000 reasons why it was successful and it only had a G3.
  • Reply 2 of 37
    mattyjmattyj Posts: 898member
    The iMac isn't a failure, but I'm not talking about the consumer side of things, I'm talking about performance, as this is cruical for Apple's pro range.



    The iMac is great, I have the original revision a machine, but I wan't to know if the G5 can rival the new chip from AMD.
  • Reply 3 of 37
    philbotphilbot Posts: 240member
    I don't know about the Hammer, but from what I've seen a single 800Mhz G5 is quicker than a Dual Athlon (I think it was a dual 1.9 Ghz but I can't be sure).



    The new 'Generation 5' mac will come with amazing graphics hardware that will change your mind about Mhz.



    This is true (only the facts have been changed).



    [ 02-02-2002: Message edited by: philbot ]</p>
  • Reply 4 of 37
    bartobarto Posts: 2,246member
    The problem is that the G5 appears to be seeded, but even if it is the Apple appears to have put the fear of god into testers.



    Hammer is still in development, and won't be out for another 1-2 years, and CERTAINLY hasn't started to be seeded yet. And AMD, like Apple, is a company which is very secretive.



    The only thing we can compare NOW is the Apollo vs Palamino vs Northwood.



    &lt;rant&gt;Apple has done well as a company in terms of keeping secrets, but I think they have gone too far. IMHO, they should use the complete lack of information as a starting point and at lease release some information on what to expect. For instance, with Macworld SF, they could have said 'See the new Flat Panel iMac, updated iBook and new 'i' application!'. And for Macworld NY, they could have said 'See the updated G4s, iDVD 2 and updated iMacs' so you don't get massive let downs, and hype is generated when deserving.



    Also, they could release a document at the start of the year to the public detailing product release dates throughout the year. For instance 'PowerMac G5, Mid 2002. 2 new Digital Devices 2nd Half 2002'. Just something to go by without releasing any trade secrets.



    Last of all, with the iPod announcement they could have said 'See the first of Apple's digital devices'. That would have attracted press, without generating the 'new Apple device is a personal teleporter' hype.&lt;/rant&gt;



    Barto



    [Edit: Motorola has released as much information on the G5 than AMD has on the Hammer.]



    [ 02-02-2002: Message edited by: Barto ]</p>
  • Reply 5 of 37
    matsumatsu Posts: 6,558member
    x86 (non-existant) vs PPC (non-existant)



    this is an interesting debate? OK, sarcasm aside, this might actually be an interesting debate for the following reason.



    Intel may support AMD. That's right, Intel may support AMD in the next generation of 32/64 bit chip.



    X86 has always had a tenacious grasp on software development. I recently read an article that put forward the following ideas.



    Intel's IA64 would almost certainly perform poorly in 'classic' x86 apps. AMD, by comparison, would not. AMD's scheme would also allow 32 and 64 product to happily co-exist for quite a while. Even Intel has never seen a move from 32 to 64 as a whole-scale platform shift. Pro's only, for at least the first 2 years. That ain't good. A whole world of software exists in ready for x86 form. AMD would be able to take advantage of that, Intel wouldn't.



    AMD has potentially provided enough of a reason to keep developers and customers off of Intel's IA64 scheme, and Intel knows it.



    But Intel has a enough marketing and engineering talent to squeeze P4 for at least another 18 months. In the meantime they'll continue to develop an already rumored 64bit x86 part that follows AMD's lead. IA64 (VLIW) will instead remain an exclusively server part. And 64bit X86 will move into competition with AMD for the high-end workstation/desktop and eventually down the line untill 32bit X86 is replaced.



    There was a code name mentioned, but I can't remember it. It all sounds reasonable to me because x86 has so many devs and customers that the winning solution may have more to do with who has the best compatibility than with who has the fastest part.
  • Reply 6 of 37
    [quote]Originally posted by philbot:

    <strong>I don't know about the Hammer, but from what I've seen a single 800Mhz G5 is quicker than a Dual Athlon (I think it was a dual 1.9 Ghz but I can't be sure).



    The new 'Generation 5' mac will come with amazing graphics hardware that will change your mind about Mhz.



    This is true (only the facts have been changed).



    [ 02-02-2002: Message edited by: philbot ]</strong><hr></blockquote>



    One question: How do you KNOW what you saw was a G5...much less what speed it is? <img src="confused.gif" border="0">
  • Reply 7 of 37
    According to AMD the K8 line of processors beginning with Clawhammer are already being sampled and they expect to ship actual processors beginning third quarter of 2002.



    [quote] In the second half of 2002, we plan to introduce processors utilizing our 130-nanometer SOI technology, including the first members of our eighth-generation Hammer family, which we believe will enable us to lead the way in delivering 64-bit performance to the desktop, while opening up new market opportunities for AMD the highest-end segments of the market. The Hammer family will deliver uncompromising performance on existing 32-bit applications while enabling users to migrate to 64-bit computing as required. Our latest plan calls for first samples to be available on 0.13-micron SOI technology by the end of this year. We have successfully produced 130-nanometer SRAMs as test vehicles in our Dresden production fab. We plan to generate revenues from our Hammer family processors in the second half of 2002. <hr></blockquote>

    <a href="http://www.amd.com/us-en/Corporate/InvestorRelations/0,,51_306_2317,00.html"; target="_blank">2001 AMD Shareholder's Annual Report</a>



    The great news is that we will be able to in near future have an opportunity to see both the G5 and the first of the K8's in action . Personally, I am anticipating the move to 0.13 process for the AthlonXP codenamed Thoroughbred. This is scheduled to occur 2nd quarter.
  • Reply 8 of 37
    According to this article over on TheInquirer.net AMD says Hammer silicon has been produced. This is more than we know about the actual state of the G5.



    <a href="http://www.theinquirer.net/31010207.htm"; target="_blank">AMD confirms first Hammer silicon out</a>



    [ 02-02-2002: Message edited by: stupidbiatch ]</p>
  • Reply 9 of 37
    amorphamorph Posts: 7,112member
    [quote]Originally posted by Derrick 61:

    <strong>



    One question: How do you KNOW what you saw was a G5...much less what speed it is? :confused: </strong><hr></blockquote>



    He answered that question in the post you quoted: "This is true. Only the facts have been changed.)"
  • Reply 10 of 37
    mattyjmattyj Posts: 898member
    I've heard enough of speculation of when the G5 will come out, but will it be worthy of the name PowerMac, compared to other PCs that will only cost one half of the price?



    Will it be worth the two years it has been in development for?



    I just hope it isn't a beefed up G4. <img src="graemlins/hmmm.gif" border="0" alt="[Hmmm]" />



    [ 02-02-2002: Message edited by: mattyj ]</p>
  • Reply 11 of 37
    telomartelomar Posts: 1,804member
    AMD is more inclined to release news about the Hammer series than Moto is about their desktop chips.



    It would be fair to assume that desktop G5s have also been produced in some quantity. Even if they hadn't been produced in some limited quantity already then they would need to be within the next quarter. I can promise we won't be getting any announcements from Motorola on that fact though.



    Just different companies and different business practices
  • Reply 12 of 37
    eskimoeskimo Posts: 474member
    [quote]Originally posted by LinuxMan:

    <strong>According to AMD the K8 line of processors beginning with Clawhammer are already being sampled and they expect to ship actual processors beginning third quarter of 2002.



    </strong><hr></blockquote>



    No, it says they will ship in the second half of 2002. That means sometime in the second half of 2002, not the beginning, itcould very well be the last week of december. Hammer in any form will not be out in Q3.
  • Reply 13 of 37
    cdhostagecdhostage Posts: 1,038member
    Funny. Arguing about processors that won't be around for years and years.



    It WOULD be good for Apple to release powerful sever-grade cihps aslong as they are seelling serveers. Y a know, right now the G4s are the exact same G4s in the prosumer line.
  • Reply 14 of 37
    spartspart Posts: 2,060member
    Prosumer line...Apple hasn't had one since the Cube and it was a flop...if you are referring to the iMac then it has a different G4 chip...
  • Reply 15 of 37
    [quote]&lt;rant&gt;Apple has done well as a company in terms of keeping secrets, but I think they have gone too far. IMHO, they should use the complete lack of information as a starting point and at lease release some information on what to expect. For instance, with Macworld SF, they could have said 'See the new Flat Panel iMac, updated iBook and new 'i' application!'. And for Macworld NY, they could have said 'See the updated G4s, iDVD 2 and updated iMacs' so you don't get massive let downs, and hype is generated when deserving.



    Also, they could release a document at the start of the year to the public detailing product release dates throughout the year. For instance 'PowerMac G5, Mid 2002. 2 new Digital Devices 2nd Half 2002'. Just something to go by without releasing any trade secrets.



    Last of all, with the iPod announcement they could have said 'See the first of Apple's digital devices'. That would have attracted press, without generating the 'new Apple device is a personal teleporter' hype.&lt;/rant&gt;

    <hr></blockquote>



    So if Apple followed your advice, can you explain how they would empty channels of old stock to make way for the new products?
  • Reply 16 of 37
    programmerprogrammer Posts: 3,409member
    If Apple issued press releases like that then they could be held accountable to them as well, which is not a good thing. Plans change sometimes for reasons beyond their control. Pre-announcing also lets other companies upstage them with their own announcements. And pre-annoucing would just focus the speculation a bit, but it wouldn't eliminate it. No, as frustrating as it is I think the secrecy is a good move for a company like Apple... as long as it doesn't get in the way of allowing 3rd party companies to prepare for new products that need 3rd party support.
  • Reply 17 of 37
    rickagrickag Posts: 1,626member
    I just saw this link posted @ Arstechnica, based on a Geek.com report.



    I don't know how accurate the G5 and IA-64 information is, but looks very promising for Apple if it is. (I'm still placing my bet that the G5 doesn't appear until very late 2002/early 2003)



    Also, note the spec's for MPC7460(still on a 0.18µm process )



    <a href="http://www.bayarea.net/~kins/AboutMe/CPUs.html"; target="_blank">http://www.bayarea.net/~kins/AboutMe/CPUs.html</a>;



    [ 02-03-2002: Message edited by: rickag ]



    [ 02-03-2002: Message edited by: rickag ]</p>
  • Reply 18 of 37
    amorphamorph Posts: 7,112member
    Good catch, rickag.



    I'm not sure where that guy got all that information on the G5 (although I notice that he's including the SPEC Scores That Ate New York - the same ones the Register claimed), but there's a wierd thing in there: Unless I'm really missing something, he's saying the G5's FSB will be 128 bits wide, and 100MHz quad-pumped (it's given as "400MHz (100x4)".



    Does that arrangement make any sense to the more hardware-inclined here? Is that compatible with DDR-II? It seems far too wide to make sense for RDRAM (unless Apple is going with 8-channel RDRAM! Talk about streaming performance, and talk about expensive...)



    [ 02-03-2002: Message edited by: Amorph ]</p>
  • Reply 19 of 37
    The guy is guessing, just like the rest of us. He has no hard info, and his guesses aren't even very informed -- e.g. it is extremely unlikely that Apple will move to a 128-bit bus. A narrow high speed bus is cheaper to build and more in line with the direction other companies are going.
  • Reply 19 of 37
    outsideroutsider Posts: 6,008member
    A quad pumped 100MHz 128 bit bus would be so expensive. What makes more sense is a Rapid IO bus running at 400MHz (800MHz effective if you factor in the DDR of the bus and full duplex) and 16 bit mode. Remember the G5 is said to have the memory bus on die so it wouldn't constrain the RIO bus. The RIO bus would be strictly used for Ethernet/PCI/Firewire/etc. , not memory.
Sign In or Register to comment.