[CubeBoy]

Originally posted by DrBoar:
Perhaps "Programmer" can tell us about the additonal work to go to 4 CPU CPU support in OS and applications (from dual support) compared the work involved going from one to dual CPU? The point of this is that if it is feasable Apple can compensate even further for slower CPUs by adding more.

I do not care if I have a fast computer at a decent price if that speed is archevied by one two or four CPUs. If it works well it works well

Multiple processors are great for multitasking and software that takes advantage of multple processors but they don't really provide a increase in performance in most of the applications we use. So unless if your using specialized software, you won't get any extra performance over single processor of the same specifications.

[Eug]

IBM and AMD to co-develop chips.

[MindFad]

Originally posted by Eug:
IBM and AMD to co-develop chips.

Interesting. What does everyone think it can be? Besides mega bad ass processors.

Edit:

I really type like ass sometimes. Sorry.

[Metzen]

Originally posted by MindFad:
Interesting. What does everyone think it can been? Besides mega bad ass processors.

1) It beens err... means a mega bad ass "lithography process" to come by 2005 Not necessairily processors per se, but close

[Metzen]

Originally posted by DrBoar:
Perhaps "Programmer" can tell us about the additonal work to go to 4 CPU CPU support in OS and applications (from dual support) compared the work involved going from one to dual CPU? The point of this is that if it is feasable Apple can compensate even further for slower CPUs by adding more.

Depends on what you do. If you do lots of gaming and word processing, you'll want a single fast processor. If you do lots of video/audio (audio is the bigger of the two IMHO), and say, run lots of power hungry apps at the same time, I'd prefer a 1.0GHzx4 CPU system vs a single 4.0GHz system. Reason being that a single processor can really only do so much work on a task then has to switch to a different task whereas multiple tasks can don't have to suffer a "switch" penalty. It's kind of the thing hyperthreading is susposed to help alleviate, multiple processors truly eliminate that problem. And for my workflow, a multiprocessor machine would be more ideal than a single fast CPU.

[voodoo]

PPC970 => iPower64

[CubeBoy]

Were the SPECmarks for Opteron and PPC970 base or peak?

[Souljah]

Apple's new "iQ Power" or "iQ Power9"
"Q" also because it comes after the "P" in P4 as in
a generation ahead.
Hmmm better register these for copyright.LOL

[voodoo]

Tha next gen Mac processor : M.T.R.L-5UX aka PPC 970

Apple has some issues w/Moto

[OreoCookie]

Originally posted by CubeBoy:
Multiple processors are great for multitasking and software that takes advantage of multple processors but they don't really provide a increase in performance in most of the applications we use. So unless if your using specialized software, you won't get any extra performance over single processor of the same specifications.

Not quite true. The other CPU takes care of other system operation (OS X being a multitasking OS).

[Superchicken]

Originally posted by voodoo:
Tha next gen Mac processor : M.T.R.L-5UX aka PPC 970

Apple has some issues w/Moto

I would bet my iMac Rev D on Steve in the key note will be making fun of motorola... a bit more than he recently did Quark.

[CubeBoy]

Originally posted by OreoCookie:
Not quite true. The other CPU takes care of other system operation (OS X being a multitasking OS).

So one of the CPUs takes care of the program while the other takes care of system operations? Is this true with all operating systems or is it only true for OS X?

Anyways, my system operations never took much performance in the first place. I do disable as many services and processes as possible though.

[Superchicken]

Dual processors mean you run at the same speed but can do twice as much in that space of time. One opperation will take as long as it would on a single proccessor, but you can do two of them at the same time. You work 1-2 times as fast
Or so I hear... sniffle... I want Dual proccessors!
Oh well some day I'll get that dream PPC 970 desktop with Dual proccessors, OS X.5 or something and I'll have a flying monkey as my pet finnally! And girls will like me for me! OK scratch that last one

[Eug]

My guess is that the current Xserve, PowerMac, and PowerBook design is so that Apple can make use of the 7457.

ie. 7457 is a 2003 part, while PPC970 is a 2004 part.

I trust this doesn't sound too stupid, or too obvious.

[clarkgoble]

The other CPU takes care of other system operation (OS X being a multitasking OS).

Typically that isn't a lot, relative to your main applications. Also some applications that users consider part of the System, such as the Finder, are horribly written in terms of multithreading. That's why when one window is busy you still can't effectively use other windows well. In terms of perceived speed I think improving Finder window performance would make a bigger difference than almost any other single thing. I honestly can't figure out why they haven't done this after this many years. Compare the Finder to Windows Explorer in this. Night and Day.

Now if you are doing two things at once, the second CPU does come into play. However for most users it is rare that their background task is doing anything CPU intensive. Typically if they are doing something it is something with a bottlekneck in terms of the network.

VPC is a little different. It seems well multithreaded and I often switch it to the background while it does its slow as mollases chugging. So having two CPUs is a definite advantage while using it. (Although having the second CPU be an x86 would have sped it up even more)

As a practical matter having multiple CPUs degrades your bus bandwidth too. So dual CPUs can often be slightly slower than a single CPU, depending upon application. (Although this is a place where the second CPU dealing with system tasks does come into play)

That's why the predicted 970 systems should be so much nicer. Much, much more bandwidth for the CPUs. It will make SMP much more efficient. It may even make quad systems a practical reality for the graphics world.

[OreoCookie]

Originally posted by clarkgoble:
The other CPU takes care of other system operation (OS X being a multitasking OS).

Typically that isn't a lot, relative to your main applications. Also some applications that users consider part of the System, such as the Finder, are horribly written in terms of multithreading. That's why when one window is busy you still can't effectively use other windows well. In terms of perceived speed I think improving Finder window performance would make a bigger difference than almost any other single thing. I honestly can't figure out why they haven't done this after this many years. Compare the Finder to Windows Explorer in this. Night and Day.

...

That's why the predicted 970 systems should be so much nicer. Much, much more bandwidth for the CPUs. It will make SMP much more efficient. It may even make quad systems a practical reality for the graphics world.

With the new PPC970, the bandwidth isn't shared, it isn't a bus topology, but a point-to-point topology.

If the motherboard is equipped with twice the memory bandwidth, there is little bottleneck at all (just if both CPUs access the same part of memory).

This is different from the bus topology, where the bus is the thing all bits have to go through. That's why we haven't seen Quad G4s yet ...

[Superchicken]

Hmm the PPC 970 is set to mass ship in second half of 2003 isn't it? That'd be like around june, hence an august announcement might very well be possible would it not?

[razzy]

I hope so

I'd really like a sweet new tower from Apple.

[OreoCookie]

Originally posted by Superchic[k]en:
Hmm the PPC 970 is set to mass ship in second half of 2003 isn't it? That'd be like around june, hence an august announcement might very well be possible would it not?

Looks like Apple is going to be ready by then ...

[Ken_F2]

Hmm the PPC 970 is set to mass ship in second half of 2003 isn't it? That'd be like around june, hence an august announcement might very well be possible would it not?

Mass production is supposed to begin in 2H 2003. IBM never said anything about shipment.

From the time that production begins, it takes 10-12 weeks to completely fabricate the processor. And then you have to account for the time for IBM to actually ship the processor to Apple, and for Apple to produce some quantity of PowerMac systems for shipment; this will probably take anywhere from two to four weeks. So if mass production on the IBM 970 begins July 1 (the first day of 2H 2003), we won't see IBM 970 systems from Apple before mid-October. If mass production begins circa October 1 (more likely), we won't see systems until mid-January.

Thus, mid-October is the earliest possible date we can possibly see IBM 970 systems. Imo, January 2004 is more likely.

[Superchicken]

aww Ken you had to spoil our hopeing

[Thain Esh Kelch]

Originally posted by Superchic[k]en:
I would bet my iMac Rev D on Steve in the key note will be making fun of motorola... a bit more than he recently did Quark.

You got yourself a bet d00d!

[voodoo]

The PPC 970 will be released this June. That is when it enters mass production. The first Macs using mass produced 970s will be ready in August. They will be introduced this September for sale. This is what will be if the G4 will not scale above 1.5GHz. On the other hand if the G4 scales above 1.5 to say 1.8 GHz, then Apple could well decide to play it safe and spend some more time on 970 integration in the Power Macs. Apple suffered from the Yikes! motherboard, so I hope they won't do anything rash like when they introed the G4. We shall see.

[razzy]

voodoo,

how do you know this? are you speculating?

you're playing with my emotions here!

[Ken_F2]

June is 1H 2003. IBM was very clear that production would begin 2H 2003. Assuming IBM wasn't lying, that means production will begin somewhere between July 1 and December 31. But again, it takes 10-12 weeks (average in semiconductor industry, and at IBM) from the time production starts to the the time that a processor is finished.

[MindFad]

Originally posted by razzy:
voodoo,

how do you know this? are you speculating?

you're playing with my emotions here!

Yeah, voodoo, dude, you're playing with my emotions. Stop it.

[OreoCookie]

Originally posted by MindFad:
Yeah, voodoo, dude, you're playing with my emotions. Stop it.

But looks like he's on the right track ...

[voodoo]

Originally posted by MindFad:
Yeah, voodoo, dude, you're playing with my emotions. Stop it.

I was summing up the most likely scenario according to rumors, speculation and information from those who say they know.

I am just as eager for the 970 as you guys are

I can't wait.

[Superchicken]

oh well... I'm just 'l be getting a powerbook next.. and I don't need to put up with the pain of power mac suers

[voodoo]

I, on the other hand, have a portable and *need* a stationary machine. I *need* the PPC 970 <slobber>.

[Simon]

Originally posted by voodoo:
I ... *need* a stationary machine. I *need* the PPC 970 <slobber>.

OK, why not get a cheap one now and sell on Ebay when you see the 970 you want. Get the small 867MHz MDD PowerMac; if you're worried about noise get a 933 QS. You're going to have to wait at least another 8 months for the 970. You might as well make the wait more comfortable.

[Superchicken]

His portable's probably powerful enough for him right now, he's just got future hardware lust.

[voodoo]

Originally posted by Superchic[k]en:
His portable's probably powerful enough for him right now, he's just got future hardware lust.

That about sums it up. I can wait for the 970, but when it comes I will get it like *immediately*.

[Apple Pro Underwear]

i have future lust too, but i realize that my real upgrade cycle will coincide with Jan 2004

if anything is out there for MWSF (or equivalent event) or if i can force myself to wait for a new revision (as my machine now; tibook 500 now is still really fast and usable but that may change in 2004) i wil purchase it

i can see myself forcing and forcing myself to wait for that one last upgrade to make it a monster machine (2005? )....can you say dual 2.5 and a video card 2 generations from the ati 9700?, hell yeah!)

[Superchicken]

I would "settle" for a Dual 2Ghz PPC 970... any day of the week

[cowerd]

i can see myself forcing and forcing myself to wait for that one last upgrade to make it a monster machine (2005? )....can you say dual 2.5 and a video card 2 generations from the ati 9700?, hell yeah!)

Yeah cause God knows, once it hits that speed computers just won't get any faster. Don't really use your computer to make money do you?

[Apple Pro Underwear]

Originally posted by cowerd:
Yeah cause God knows, once it hits that speed computers just won't get any faster. Don't really use your computer to make money do you?

yes, i do...do you?

ever work as the sole graphic artist for a client? print, presentation and web at the same time?

thats a hell of a lotta apps. as of right now, i prefer to boot into 0s9 because it's faster than osX too.

what the heck do you do that you don't want a 2005 powermac to be all that it's capable of being? you like spending 3000 on a machine that uses technology thats a year behind?

[fulmer]

I've done graphics and video work, too, and having a faster Mac is a godsend. If rendering and encoding times were shaved by just 30%, that would mean hours of additional productivity per week.

[voodoo]

Nothing new on the 970 rumor front huh? If only IBM was a little bit more lose lipped. Tell us how the designing and manufacturing process is going and things.

[Eug]

The new PowerMacs are out, at 1.42 GHz.

Rumour has it that they're 7455 chips, and NOT the 7457. If so, I'm impressed they've managed to milk this 0.18 u part for so long.

Do you think this means it's gonna be G4 7455 --> PPC970 directly?

Or is it gonna still be 7455 --> 7457 (RM?) --> PPC970?

[MindFad]

Originally posted by Eug:

Do you think this means it's gonna be G4 7455 --> PPC970 directly?

Or is it gonna still be 7455 --> 7457 (RM?) --> PPC970?

Well, I think we all hope it's the former, but I have a feeling it will be the latter. Oh well -- when they get here, they get here.

[Eug]

Yeah that's what I'm afraid of too. The 7455 has already drunk too much from the fountain of youth, and it's clear that the magic can't last forever. (I am assuming that the rumours are true that the new PowerMac still is 7455.)

I suspect that the 7457 may appear, but in line with what Motorola's roadmap said when it will appear. This will mean the DDR-capable high-GHz 7457-RM not appearing until the 2003H2 at the earliest (or possibly later, given that we still haven't seen the 0.13 u 7457 non DDR part which was already supposed to be here), with the PPC970 shipping not until well into 2004.

The current PowerMac is made for the 7457. One can only guess when we'll see the 7457 actually being used in the towers.

[Superchicken]

perhaps we'll see one rev of the newer G4s and then the PPC 970... after all the iMacs and eMacs still use G4s and the power books will still and what not.. least for a while... and once the power books are about to move to PPC 970 I could easily see the iBooks steppin up.

[Eug]

Originally posted by Superchic[k]en:
perhaps we'll see one rev of the newer G4s and then the PPC 970... after all the iMacs and eMacs still use G4s and the power books will still and what not.. least for a while... and once the power books are about to move to PPC 970 I could easily see the iBooks steppin up.

That's a good point.

I've always said the iBooks and iMacs will continue to use the G4 for some time, and likely the PowerBooks as well initially. The G4 AluBook 12" is already the new prototype 7457 iBook. And of course, the new 17" AluBook is the new prototype 7457 PowerBook. I suspect the difference between the PowerBooks and iBooks will not be G4 vs. G3 by 2004. They will be both G4 but with different speeds (and features).

And I suspect incorporating the 7457 into a PowerMac will not be a huge problem. I wonder if the 7457 is pin compatible (at least for the non-DDR version) and if the voltage settings are already incorporated into a PowerMac mobo compatible BIOS firmware.

[madsenj37]

Maybe its just me but when the 970 was announced I seem to remember an acronymous name being thrown around. I believe it was GPUL for Giga Processor Ultra Light. Does that ring a bell?

[clarkgoble]

That's even the name of the thread over at ARS.

http://arstechnica.infopop.net/OpenT...3470943335&p=1

[Superchicken]

yeah it was orignally reffered to as the GPUL, probably the name used internally before a propper name was assigned to it.
But yeah in the begining it was known as the GPUL, now as the PPC 970... eventually it'll be known by whatever name steve gives it.
But then again if IBM intends to ship these in boxes of their own... possibly even into the consumer desktop space... wouldn't it be smarter for IBM and Apple to agree on a name?

[clarkgoble]

More than an "if" for IBM selling 970 boxes. They've already announced they'll be selling Linux boxes for it. I keep hoping that Apple licenses OSX to IBM and that we have IBM branded OSX boxes. But that is probably hoping too much.

[BkueKanoodle]

Originally posted by OreoCookie:
Not quite true. The other CPU takes care of other system operation (OS X being a multitasking OS).

This isn't true unless apple has set their system to do asymmetrical multiprocessing, which I don't know for sure, but doubt. In ASMP the two cpus can be assigned to to different task, where as in Symmetrical MP the work load is divided evenly across the CPU's.

[OreoCookie]

Originally posted by BkueKanoodle:
This isn't true unless apple has set their system to do asymmetrical multiprocessing, which I don't know for sure, but doubt. In ASMP the two cpus can be assigned to to different task, where as in Symmetrical MP the work load is divided evenly across the CPU's.

Nope. ASMP means that one CPU delegates the work to the other one. Classic (up to OS 9) worked that way. MacOS X has SMP -- the OS schedules work (threads) to each CPUs. If both tasks don't depend much on each other, they interfere even less. In this case, if the OS does a background task that is independent of your app (e. g. printing, iTunes playing a song, encoding video) while you are doing something else, then you will notice an enormous boost.