CONFIRMED: G5 enters volume production! - Page 6

[quote]Originally posted by DaveLee:
<strong>Re; the Reg article about Power4 low end.

This is interesting especially coupled to what has been discussed on this thread. But one thing confuses me - why on earth would IBM develop a chip (that Apple may use), only to have them compete for the desktop/workstation space?

IBM still have workstations based on the 604 and Power3 (yet at an almighty cost) so what makes people think that Apple will be able to offer a more cost effective solution?

It would be remarkable if they did though...</strong><hr></blockquote>

I can't imagine IBM sell 50,000 POWER4 boxes per year, but Apple could potentially sell 250,000+. Economies of scale might push the price down, but whether it could be brought down to prosumer levels is anybodys guess - the cheapest pSeries 690 @ 1.1 Ghz x 8 is $450,000!

POWER4Mac, just $100,000 with a free iPod

You can get the top p670 with a Power4 for a smidge over $100k I think.

But what I meant was that IBM is still making workstations well above the price of a G4 PowerMac, and these are still based on the 604e!

Their Power3 based offerings are evn more expensive, and yet there are these rumours flying around that they will be able to offer a dressed-down Power4 based PowerPC at a pricepoint where Apple could just slip it into their desktops?

I don't buy it. IBM would use this new chip first surely?

How does the POWER4 handle memory access? Does each core have a controller or is it shared?

The POWER4s The Register is talking about aren't meant for Macs they are just meant for IBM workstations.

A dummied down POWER4 may or may not be in Apple's future but this release was always planned for IBMs use so it isn't anything too special. Unless of course you are in the market for an IBM workstation.

Hmm, let's see, Apple takes 1 or 2 G4 PPC chips that cost 100-200 dollars a piece, sticks them in one of their towers and sells them for at least 500 dollars more than a comparable P4 system, when the highest speed grade P4s cost upwards of 600 dollars.

I wonder how cheap this cut-down Power-4 would have to be for Apple to use it?

Never gonna happen!

What's gonna happen is, eventually this Power-4 technology will trickle down in some mutilated fashion, and all too late, mind you, into IBMs embedded PPC products such as the PPC 7xx series and that's when Apple will use it.

[ 06-14-2002: Message edited by: timortis ]</p>

The high cost of IBM's machines isn't because of the processor, they put a lot of other high end equipment into these machines and sell it with expensive software and an IBM support contract.

If they did design a POWER4 for Apple then economies of scale would bring the processor price down to reasonable levels, and Apple would put it in a machine designed for the desktop market. Intel is selling chips of about the same order of complexity (but single core) in the desktop and notebook markets, so why do you think IBM couldn't? The only reason they haven't is because they don't have a consumer OS and application base to build machines for... they are the original PC company and their shot at defining a new platform died with PreP/CHRP in the mid-'90s. If Apple has asked them to build the processor, however, I'm sure they will.

Programmer could you define this news and what does it meant for Moto and IBM who are in Rapid I/O when Apple is with HyperTransport (ie. AMD).

"RapidIO advances
PLX's top rival, Tundra Semiconductor (Ottawa), is taking a different course, with plans to roll out in the next three months four-port switches and bridges for the RapidIO standard initiated by Motorola Inc. Other devices, including parts for serial RapidIO, are in the works, said Ian McGill, vice president of marketing for Tundra. Like Advanced Switching, RapidIO targets routers, switches and other comms systems including storage systems, and has backing from Nortel and Ericsson, among others. McGill said systems using RapidIO will probably not ship until 2004, about the same time PCs using Express hit the market."

it's from there: <a href="http://www.siliconstrategies.com/story/OEG20020607S0056" target="_blank">http://www.siliconstrategies.com/story/OEG20020607S0056</a>

Does this means that we will never see a G5 with RIO before 2004 or mid-2003. If this is true, I really think that Apple would paid big $$$ to have a new FSB, better FPU units, 256 bits AltiVec, etc. for the G4 coupled with HT... and Apple has the right to ask cuz of this: Motorola reports that its top ten end-customers, which accounted for about half its revenue in 2001, were Apple, Bosch, Delphi, Hewlett Packard, Lucent, Motorola, Qualcomm, Siemens, Sony, and Visteon.

[ 06-14-2002: Message edited by: jeromba ]</p>

oh ! :eek: this is GOOD news !

(06/06/02 14:52 p.m. EST)
BEIJING -- Motorola Inc. here today outlined its new strategy in China, announcing plans to spend an additional $1 billion in R&D and hire 4,000 more engineers in that nation.

And what if the next processor for the PowerMac would be a Dual Core 8540 with AltiVec mixed with the POWER4 ? Is it possible ? When it come to price, it will be certainly in the same ball park that the current G4, no?

Here's my new idea for a shared processor scheme that scales with flexability.

Take your master processor, with all the modern bells and whistles. Long pipeline, big L2 cache, on die memory controller, etc. It connects via the motherboard with RapidIO. In fact this processor has 2 16bit RIO ports. Because one would be for slave processors. These eunich processors would be processors with their nads cut off. They would have identical processing units, but would only have one RIO port that would connect to a central RIO hub that in turn connects to the Master processor. The RIO would have about 60Gbps bandwidth (7.5GBps). The slaves would also lack the memory controller. This would limit the pin outs to about 120 since 16bit RapidIO only requires 76 pins. A smaller package would be easier to design and make the processor card simpler to design for more than 2 processors. All processors would access memory over RIO to the memory controller in the master processor.

With a setup like this you can have 2, 3, 4, 5 processor setups with simple modifications to the processor card design. Is this not a feasible idea?

Hmmm. I suppose I could just be obtuse this morning. But why is everybody assuming that "apple pie" is really Apple PI?
Is it possible that "apple pie" really information sharing of some form (sharing the pie, so to speak) a la Amiga?
I seem to recall that Amiga remained quite viable for a long time with this method, even with dirt slow CPUs. It even remained as a hidden unit inside video-editing equipment well after it was available as a desktop.
I would think that could also explain "Trinity" in it's most basic usage - Amiga used 3 proc's - each for a different function, didn't it?

Anybody think this feasible?

My apologies, but there simply wasn't enough rampant speculation going on in this thread.

At it again: PowerPC chips with RapidIO to ship "later this year."

<a href="http://theregister.co.uk/content/61/25728.html" target="_blank">http://theregister.co.uk/content/61/25728.html</a>

[ 06-14-2002: Message edited by: Harald ]</p>

[quote]Originally posted by jeromba:
<strong>oh ! :eek: this is GOOD news !

(06/06/02 14:52 p.m. EST)
BEIJING -- Motorola Inc. here today outlined its new strategy in China, announcing plans to spend an additional $1 billion in R&D and hire 4,000 more engineers in that nation.</strong><hr></blockquote>

It's not so good when you look at it from the standpoint that they'll probably lay off another 4,000 engineers in the U.S. and shut down more fabs here.

As for the Register article, I really don't see why anyone is getting axcited about it. The article does not have any info on a workstation version of the Power4 (let alone really any different version than what's been out for awhile. It's just talking about a &lt;b&gt;workgroup&lt;/b&gt; server version. It seems to strongly allude to the chips in these just being reject silicon from the high-end line of servers, where one core doesn't work and has been cut off or disabled. While a smart move on IBM's part, as a way to make use of bad chips, it's hardly a viable approach for Apple chips. At least, I should hope IBM isn't producing enough partial reject chips to meet all of Apple's PowerMac requirements. Of course, a 1/250 success rate might help explain the cost of the Power4!

Just some more fuel to the fire. Ok, it's SpyMac, but this stuff is fairly obvious. Unfortunately they have the same line ThinkSecret was carrying about the Jobs vs. Sculley announcements, and we all know how that one turned out.

[quote]Apple plans on making big waves in the graphics industry, and at Macworld New York, CEO Steve Jobs will shed some light on a long-planned project that will showcase Apple's commitment to the high-end graphics market.

As we reported at the beginning of April -- before the announcement of Apple-branded servers, the purchase of Silicon Grail, or even articles suggesting Pixar's move to OS X -- Apple will announce a strategy that will be mutually beneficial to both Pixar and Apple.

To protect our source, we are not able to provide any further details on the new technology. However, a quote from our original story stands true:

"[Steve Jobs] could be stepping on a lot of feet... I have a feeling [Jobs] does not want anyone to know of this strategy."
<hr></blockquote>

Ok, it's entirely possible Jobs can reveal his fantastic 3D graphics strategy and only speedbump the machines to 1.1 GHz , but like others have said, this only increases the necessity and probability of vastly faster machines.

[quote]Originally posted by jeromba:
<strong>oh ! :eek: this is GOOD news !

(06/06/02 14:52 p.m. EST)
BEIJING -- Motorola Inc. here today outlined its new strategy in China, announcing plans to spend an additional $1 billion in R&D and hire 4,000 more engineers in that nation.</strong><hr></blockquote>

But this isn't-
<a href="http://apnews.excite.com/article/20020615/D7K5ABA80.html" target="_blank">Motoral credit rating lowered</a>

[ 06-15-2002: Message edited by: KidRed ]</p>

umm.... is that the link you meant to post? or...am i missing something?? <img src="confused.gif" border="0">

[quote]Originally posted by los2000:
<strong>umm.... is that the link you meant to post? or...am i missing something?? :confused: </strong><hr></blockquote>

Damn, that is weird. It was a story on Excite yesterday about Moto's credit rating going down to right above JUNK status. So Moto isn't in good shape right now. I'll try to verify the link.

Ok, the above link is fixed now.

<a href="http://apnews.excite.com/article/20020615/D7K5ABA80.html" target="_blank">http://apnews.excite.com/article/20020615/D7K5ABA80.html</a>

[ 06-15-2002: Message edited by: KidRed ]</p>

ic .... geez, motorola looks like it's in trouble. not DOOMED, just not in the best of shape.

Maybe an odd question, but how many posts until this thread is dead? It get tiresome scrolling through 6 screens. I vote for new clean thread where people can repost cleaned up condensed versions of anything relevant. (then again my posting here is just adding to the problem I'm griping about) sorry. <img src="graemlins/embarrassed.gif" border="0" alt="[Embarrassed]" />

John,

Look to the right of the posts in the listings, there are numeric indicators of how many pages any given post is comprised of... Clicking on one of the numbers will take you directly to the page you want, no need to go through all of the preceding pages to get to the latest dirt...

I usually just click on the name of the last post-er and it takes me to the last post in the thread.

[quote]Originally posted by moki:
<strong>

I'm going to have to take that back a bit -- the mobo I mentioned has apparently slipped again, due to the reasons I mentioned earlier. What will be at MacWorld will most likely NOT be the DDR implementation I mentioned, but rather, a less ambitious one.</strong><hr></blockquote>

&lt;bad taste&gt;
&lt;gansta&gt;

Hey Yo Moki

&lt;/gansta&gt;
&lt;/bad taste&gt; not working? Oh well ...

Does this triangulate?

MOSR

"A very exciting rumor that is coming out of Cupertino: expect 8X AGP, 800Mbps Firewire, AND USB2 on the next generation of PowerMac. Sounds good to us...."

Or is this good news? As in, we might see plan A with DDR after all?

[edit, pushed return like a dumb-ass before I was ready]

[ 06-17-2002: Message edited by: OverToasty ]</p>

[quote]Originally posted by OverToasty:
<strong>

&lt;bad taste&gt;
&lt;gansta&gt;

Hey Yo Moki

&lt;/gansta&gt;
&lt;/bad taste&gt; not working? Oh well ...

Does this triangulate?

MOSR

"A very exciting rumor that is coming out of Cupertino: expect 8X AGP, 800Mbps Firewire, AND USB2 on the next generation of PowerMac. Sounds good to us...."

Or is this good news? As in, we might see plan A with DDR after all?

[edit, pushed return like a dumb-ass before I was ready]

[ 06-17-2002: Message edited by: OverToasty ]</strong><hr></blockquote>

The only thing Moki really mentioned was the chipset DSPs, so the MOSR rumour doesn't counter what he said. Of course if he wants to share more I wouldn't object...

[quote]Originally posted by haderach:
<strong>Hmmmm...:

<a href="http://www.theregister.co.uk/content/61/25722.html" target="_blank">http://www.theregister.co.uk/content/61/25722.html</a></strong><hr></blockquote>

Curiouser and curiouser, no?

[quote]Originally posted by moki:
<strong>Curiouser and curiouser, no? </strong><hr></blockquote>

Well, yes and no... that particular article seems to indicate that those "low end" machines are just using failed POWER4s where one of the two cores was stillborn. I don't think a true consumer level POWER4 has materialized (yet).

I think it would be very cool if Apple had two completely different PowerPC designs to choose from, one from Moto and one from IBM (both with AltiVec & BookE compliance). Even better if they all spoke RapidIO. AIM would finally come to fruition, well after it had been declared dead.

[quote]Originally posted by Programmer:
<strong>

Well, yes and no... that particular article seems to indicate that those "low end" machines are just using failed POWER4s where one of the two cores was stillborn. I don't think a true consumer level POWER4 has materialized (yet).

I think it would be very cool if Apple had two completely different PowerPC designs to choose from, one from Moto and one from IBM (both with AltiVec & BookE compliance). Even better if they all spoke RapidIO. AIM would finally come to fruition, well after it had been declared dead.</strong><hr></blockquote>

Right.. when we talk hypothetically about "consumer POWER chips," the AltiVec unit(s) are implied, right? It would be silly of Apple to lose it after building what JYD called a "software empire" on it.

[quote]Originally posted by Programmer:
<strong>

Well, yes and no... that particular article seems to indicate that those "low end" machines are just using failed POWER4s where one of the two cores was stillborn. I don't think a true consumer level POWER4 has materialized (yet).</strong><hr></blockquote>

Of course... but there are some interesting signs being shown by IBM (in a good way), and MOT (in a bad way), no?

[quote]Originally posted by moki:
<strong>Of course... but there are some interesting signs being shown by IBM (in a good way), and MOT (in a bad way), no?</strong><hr></blockquote>

Well I for one found it kinda strange that all of a sudden the G4 upgrade market (for G4 based systems) is finally gonna happen AND at speeds pretty dern close to 'todays' PowerMac's.

Being the owner of the 1st AGP PowerMac I've been stuck at 450Mhz for years and no upgrade path to speak of... Well that's gonna change in just over a month and I'll finally be able to add some speed to the system. Sorry but I'm the type that finds it really hard to 're-buy-into' the same class CPU. At work it's a different story (not my money) but at home (with a wife) getting the same CPU is a lot harder to sell. "G4 you already have a G4... yadda yadda yadda"

Well something MUST HAVE changed to allow those upgrade card makers to finally provide support for these machines. After all you know they WOULD HAVE been selling them today (if they could) and since two different mfgs announced support so close to one another I'm pretty sure whatever problem existed wasn't a technology related one but a political one.

Just speculation here but:

Apple to MOT: "We want an exclusive on all of your &gt;500Mhz G4s"
MOT to Apple: "So long as MOT is used in top end Macs and/or if you pay an extra $x per CPU"

Well it looks like either Apple said... nahhh go ahead and sell the CPUs to whoever you like (not likely) or something else happend and the deal was broke by Apple or MOT.

While I/we LOVE the idea of a G4 based upgrade path without having to buy a new box you have to know Apple must HATE IT. Apple don't make a dime outta the deal! Well all I know is something had to change.

Dave

Moki... what do you think of the CELL architecture being developed by IBM/Sony/Toshiba (and hopefully Apple ) IBM has stated that this can be used from pda's,phones up to powerful servers. Plus since Sony is the biggest electronics maker in the world and there are more playstations than any other console talk about economies of scale... do you think this could be used for future Mac's?

[quote]Originally posted by Programmer:
<strong>AIM would finally come to fruition, well after it had been declared dead.</strong><hr></blockquote>

Assuming its not common knowledge, Ill mention that after the last MWNY (QS 867 + dual800's), Steve said that when he got to Apple, the AIM alliance had fallen to shambles. The processors at that show were supposedly the first product from a slowly revitalizing AIM. Though perhaps that was just feel good nonsense, especially considering Mot did all the work on those processors, and the ones in the next upgrade too. Though his statement(which I can't find, and could be a figment of my imagination) might of held some relevance to future developments.

[ 06-17-2002: Message edited by: nathan22t ]</p>

[quote]Originally posted by nathan22t:
<strong>Assuming its not common knowledge, Ill mention that after the last MWNY (QS 867 + dual800's), Steve said that when he got to Apple, the AIM alliance had fallen to shambles. The processors at that show were supposedly the first product from a slowly revitalizing AIM. Though perhaps that was just feel good nonsense, especially considering Mot did all the work on those processors, and the ones in the next upgrade too. Though his statement(which I can't find, and could be a figment of my imagination) might of held some relevance to future developments.
</strong><hr></blockquote>

Wasn't that soon after IBM had started fabbing G4's under license from Moto in order to get enough to satisfy Apple? I guess that would be a first step to bringing AIM back to life.

"What IBM is almost certainly getting ready to announce is what have been called the Regatta-L machines, where L stands for low end. Not much is known about these machines, but it is very likely that they will use single-core Power4 processors - most likely dud parts that can't be used in pSeries 670 and pSeries 690 servers or in the iSeries line - to create uniprocessor, two-way, four-way, and maybe even six-way or eight-way configurations. Odds are that the smallest forthcoming Regatta-L machines will be called the pSeries 630 - we guess that this will be uniprocessor and dual-capable models - and the pSeries 650 - perhaps with four-way and six-way capability like existing pSeries machines. "

Could this split the 'power'Mac market into two?

The current range with a 'sgi' level further up the pricing bracket. Especially in light of recent 'high end' software aquisitions..? Shake et al?

In light of Prog' and Moki said...these may not be the actual 'G5' but evidence of IBM's move into the 'pro-sumer' space.

The 'e500' as per Motorola 8450? (8540?) THT has doubts. Cept it's modular. For Apple devices. It may have its uses in terms of portability? Do we know its cooling requirements? Unless certain features are taken into the 7500. If the core runs at 1 gig. If moto make a 'dual' e500 core then...they'd claim an effective '2 gighz'? Bolt on enhanced altivec and fpc and you get your 'numbers' in terms of throughput via Rio?

The Power4 architecture seems to offer Apple potential on so many fronts.

Mhz wise. It seems to have room to stretch its legs. In real performance terms, it's a clear leader in integer and a killer in fpu with room to grow. The ability to add cores while expensive now seems to offer a real 'upgrade' path for the 'power'Mac line ('power' as in how it currently stands...) to head into Sgi style territory...while return to competivity in its current market.

If next generation machines are both on the 'book e' architecture...the AIM alliance, even indirectly, is back on track and seems to give Apple options. Something, for the last 18 months, it doesn't seem to have had. Seems Apple have toughed it out and that the 'G4' may be 'retired' soon.

Or...perhaps Apple will use IBM for a move into low end Sgi workstation market and Moto's next chip for current 'power'Mac markets. Or single and dual core IBMs for 'powe'Macs and Moto's chip for general purpose and low power devices. (Cept Moto's current chips still run hot in Powerbooks...)

Gut instinct tells me Apple have finally had enough of Motorola and we'll see its deaththrows from now until San Fran next year. G4 .13 Xserve at 1 to 1.4 gig. And early next year, same G4, stretched a little further but in a Rio framework.

Then next Summer for IBM in said Rio framework. Cept this time...the processor will be 64 bit.

All the stuff on these boards is vague enough to confusing. You can take much of it any way you like. Even from a non-techie viewpoint...it looks like Moto is dutifully serving its G4 on its way out of the 'power'Mac line. It'll stay for Consumer...but IBM is picking up the bat with renewed interest in the 'desktop/console' space?

Hmmm. All very intriguing...but I don't think I'm gonna get what I 'want' this Summer or early next year. The realist in me says, 'G5'? IBM, this time next year. (They may call that the 64 bit G5 and the G4 in the Rio the 32 bit G5?)

Lemon Bon Bon

<img src="graemlins/hmmm.gif" border="0" alt="[Hmmm]" /> <img src="graemlins/bugeye.gif" border="0" alt="[Skeptical]" /> :confused: <img src="graemlins/hmmm.gif" border="0" alt="[Hmmm]" />

[ 06-18-2002: Message edited by: Lemon Bon Bon ]</p>

BUMP.... BUMP...

Look back over this thread... search for comments from 'justafriend' (AI search is broken for me).

Some highlights but I might have missed some:

---------------[ 1st post ]-------------
Here's your inside info on IBM PPC designs:

IBM has developed Power4. For the workstation.

On the low-end is a 1 chip, dual core design. High-end is a 1 chip, quad core... Shipping 4th quarter, 2002.

No word on any relation to these proc's and Apple. However, if this chip begins shipping in Q4, that would mean Sept., which would work out well if new Power Macs utilizing this chip were intro'd during MWNY...
---------------[ 2nd post ]-------------
For clarification purposes, what's been out since last year is the SERVER version. This is a WORKSTATION version - a derivative of the server chip... this isn't out yet ;o
---------------[ 3rd post ]-------------
Unfortunately, no direct public reference. I posed a question to my acquantance in IBM, and was told that the answer to the specific question I asked was "IBM Confidential". Then, it was followed by, "but what I can tell you..." and that's what I posted here.

As an aside, Jamee Abdulhafiz is listed on this page ( <a href="http://www.research.ibm.com/journal/rd/461/luddeaut.html" target="_blank">http://www.research.ibm.com/journal/rd/461/luddeaut.html</a> ) as Senior Engineer and verification leader for a POWER4 derivative microprocessor...

So, there is some proof to the statement that Power4 derivatives are planned... Just no detail like what I was
-------------------------------------------
End of posts I could find...
-------------------------------------------


So after reading about the new IBM chip it looks like 'justafriend' was dead on.

Dave

Why is everyone unregistered?...

[quote]Originally posted by DaveGee:
<strong>So after reading about the new IBM chip it looks like 'justafriend' was dead on.

Dave</strong><hr></blockquote>

If you're referring to the POWER4 derivative chip for the desktop and lowend server, I highly doubt that it will have dual and quad cores. But it's likely justafriend got the scoop on the IBM processor to be previewed on October 15.

OTOH, maybe he's just referring to the POWER4 that is being packaged in single (dual cores) and dual (quad core) configurations on "lowend" AIX Servers and will likely be used in some RS/6000 workstaions.

[quote]Originally posted by zaz:
<strong>OK, wait a second...

Has anyone actually tested an Xserve and is anyone well informed as to its real memory performance...?

Or is all this conjecture and dubbing of the Xserve DDR system a 'hack' and 'ineffective' interpolation of Jobs' Slide show?

Until we see some hard numbers this whole 'down on Xserve DDR hackery' seems like a lot of hot air.


..but better than what?</strong><hr></blockquote>

Here we go...

<a href="http://www.barefeats.com/xserve2.html" target="_blank">http://www.barefeats.com/xserve2.html</a>

Who revived this long dead thread?

MOOOODS!

G-News

[quote]Originally posted by G-News:
<strong>Who revived this long dead thread?

MOOOODS!

G-News</strong><hr></blockquote>

The Mods are now unregistered. Anarchy! Anarchy! Anarchy! How I've dreamed of this day. Now that its here, I must admit I'm a little afraid. Forget the anarchy. Restore the Mods! Restore the Mods!

Must - find - waste paper basket - to - throw - through window... must find window...


Screed ..."We're the Hooligans" &lt;knock&gt;