A Mobile 64-bit PowerPC? [Archive]

Henriok

Nov 13, 2003, 05:10 PM

This is logical. Power consuption is becomming a real problem with high end PowerPC processors, while the low end is doing a REALLY good job ceeping consumption down to a minimum. IBM is after all tageting the 440 core to PDAs so we are looking at chips pulling less than 1 W at frequencies around 400 MHz. Motorola is doing a great job with its G4-line. less than 10 W @ 1 GHz. It's really unacceptable for a 3 GHz processor to draw ~100 W so IBM should make every effort to reduce that.

Fabrication process is one thing but it can only take one so far, reducing power consumption with approximately one third. Can they improve fabrication so that they can use a low volt core, perhaps at 0.9 V instead of 1.3 V? I think complete redesign is the way to go and stip the processor from elements that might be unnecessary. Are two FPUs and a large L2 cache really that important? Is there a way to power down ALUs when they are unused? Can they run different areas of the processor at different voltages at the same time? If they improve compilers, can they skip one of the load/store unit? If they integrate the northbridge, they they might save power when looking at the chipset design in whole? AltiVec? Is it really that important?

And so forth.
Compared to other chip designs, and theeir evolution over the years, I really cant se any reason why IBM should be able to cut power consumption in the 970 design by at least halv by just being clever. 970 is a new design and we are looking at the first implementation. IBM have a continuing evaluation process going on and are constantly looking for areas that they can improve upon. Their East Fishkill facility is a relatively young plant too so there's probably a lot that can be made there too.

I wouldn't be surprised that Microsoft is concerned with power consuption of the future Xbox 2-chip. I guess that the abyssmal record of the x86 processors was one reason for them not to choose that path, but IBM's G5 line is not that impressive either. Not compared to the flavours of G3s.

In short: This is a natural move from IBM. It's a job that must be done.

cb911

Nov 13, 2003, 05:31 PM

aahhh... it's times like this... like my sig says, i always have my grain of salt with me. ;)

it would be really good if they get it a low-power version going... :D

manitoubalck

Nov 13, 2003, 05:32 PM

Originally posted by Henriok
If they integrate the northbridge, they they might save power when looking at the chipset design in whole?

Might have copyright issuse with for using AMD's true 'Hyer transport' which allows processors to address each other without going though the north bridge. It also allows each proc to have it's own dedicated RAM, which it address' without the use of a north bridge.

dgbatchelor

Nov 13, 2003, 06:17 PM

"Faster" seems to always be better for a laptop... But it is the overall performance of the laptop which is important -- not just the raw processing power of a given processing core...

The natural progression has been from 8-bit, to 16-bit, to 32-bit, to 64-bit processing cores...

Has anyone given thoughts to a future PB architecture with several (read this as 8 or more) low-power 32-bit processing cores which would act like a small local SMP "cluster"?

Would we likely see better overall performance for the same battery life? Energy savings techniques could put one or more of the processing cores to sleep if they were not needed at the time... Would this also give lower overall heat generation?

Clearly, there are issues to solve regarding bus arbitration...

I know that I'm over simplifying, but could you imagine a time when the user could purchase extra CPU power on the equivalent form factor of a DIMM and install/replace 8 (or more) processing cores at a time?

Think of the approach to be an analogy of RAID...

...that would let the consumer decide how much processing power to install ("Did you want a 2-pack of processor modules... each with eight low power 2GHz G4 cores...?")

Another BTO option?

I know that this solution wouldn't likely result in the equivalent of a supercomputer cluster of 1100 DP G5 nodes... (But we can dream...)

Thoughts?

-- dgbatchelor --

"If you have built castles in the air, your work need not be lost; there is where they should be. Now put foundations under them." -- Henry David Thoreau

wizard

Nov 13, 2003, 09:40 PM

While you bring up some good points at the same time you seem to fall intp the abyss of regressiveness.

First where is it said that the 970 burns 100 watts of power???

I agree that the Motorola processor is good on a watt/performance basis, it is unfortunate though that Apple can't seem to get Motorola to produce faster revisions. All they really need to do is to increas the clock speed of the G4 75% and they would have a viable processor for another year. I'd like to see a show of hands as to the likely hood of Motorola accomplishing that.

As to your comments on cache and FPU units hell yeah we want them. On a laptop the larger the on die cache the better. As far as the other execution units, I want it both ways. That is better power disapation and all the features of the latest processor generation. Lets not short change ourselves at least in the case of the PowerBook.

As far as Alt-Vec goes, that is really a silly question isn't it. Every mainstream processor that has come on the market recently has had some facility to process vectors. I could see a reduced capability vector unit for chips destined for certain markets such as the IBook but not for mainstream laptops. Now if they find ways to improve it fine, but lets not castrate it for the sake of power reductions.

Thanks
Dave

Originally posted by Henriok
This is logical. Power consuption is becomming a real problem with high end PowerPC processors, while the low end is doing a REALLY good job ceeping consumption down to a minimum. IBM is after all tageting the 440 core to PDAs so we are looking at chips pulling less than 1 W at frequencies around 400 MHz. Motorola is doing a great job with its G4-line. less than 10 W @ 1 GHz. It's really unacceptable for a 3 GHz processor to draw ~100 W so IBM should make every effort to reduce that.

Fabrication process is one thing but it can only take one so far, reducing power consumption with approximately one third. Can they improve fabrication so that they can use a low volt core, perhaps at 0.9 V instead of 1.3 V? I think complete redesign is the way to go and stip the processor from elements that might be unnecessary. Are two FPUs and a large L2 cache really that important? Is there a way to power down ALUs when they are unused? Can they run different areas of the processor at different voltages at the same time? If they improve compilers, can they skip one of the load/store unit? If they integrate the northbridge, they they might save power when looking at the chipset design in whole? AltiVec? Is it really that important?

And so forth.
Compared to other chip designs, and theeir evolution over the years, I really cant se any reason why IBM should be able to cut power consumption in the 970 design by at least halv by just being clever. 970 is a new design and we are looking at the first implementation. IBM have a continuing evaluation process going on and are constantly looking for areas that they can improve upon. Their East Fishkill facility is a relatively young plant too so there's probably a lot that can be made there too.

I wouldn't be surprised that Microsoft is concerned with power consuption of the future Xbox 2-chip. I guess that the abyssmal record of the x86 processors was one reason for them not to choose that path, but IBM's G5 line is not that impressive either. Not compared to the flavours of G3s.

In short: This is a natural move from IBM. It's a job that must be done.

jade

Nov 14, 2003, 01:06 AM

well i hope pbg5 have a much faster frontside bus. The current crop of centrinos have 400mhz, and version 2 will have 533mhz.

MacRAND

Nov 14, 2003, 01:28 AM

dgbatchelor :) I agree

While I drool at the thought of a G5 dual 2.5 GHz box, and I look forward to multi-chip G5 xServers
I really question the rational behind putting a currently HOT G5 chip into a PowerBook, until things cool off a bit. Why bend yourself out of shape with extreme radiators, just wait till the size of the chip and current needed to run it match for a laptop.

In the meantime, I'd love to see a G4 dual 1GHz chip configeration in a 15" or 17" PowerBook. Smokin'! Or, would it also be too HOT!
Yes, G4 multi-chip PB, even before a slow G5 1.5 GHz PowerBook, and I'll bet that this is the G5 PB direction, and NOT 1.8 or 2.0 GHz. Problem is, two G4 chips may run hotter than one G5, can't do the math, but Apple is.

The move to G4 in the iBook was long overdue, I won't miss G3 (which I have), unless they pep it up with radical cache or something. G3 Multi-chips, why not? How about some of these new low-power IBM chips (not G5)?

CoreForce

Nov 14, 2003, 02:54 AM

Originally posted by mrsebastian
drool = )..... i want my 17"pb g5!

Me too.

Apple should set up a site for taking pre-orders for PB G5, 15'' and 17'' and advertise this to some selected places only.

Just to see what the demand really is.

Henriok

Nov 14, 2003, 06:23 AM

Originally posted by wizard
First where is it said that the 970 burns 100 watts of power???
I really can't vouch for the accuracy of AppleInsider's sources, they do state (http://forums.appleinsider.com/showthread.php?s=&threadid=33845) that a 130 nm 970 @ 2.5 GHz consumes 97 W. FAR too much.

Lets not short change ourselves at least in the case of the PowerBook.
No I certainly hope they can avoid that.

As far as Alt-Vec goes, that is really a silly question isn't it.No I don't think so. Assume for a moment that we can get a "PowerPC 960" that is a 970 that run at 3 GHz but does not include AltiVec.. Wouldn't you prefer that one to a 1 GHz G4 in a iBook?

when looking at the die of 970 (http://www.theinquirer.net/images/articles/ppc970.jpg) it seems to me that they could've used the space more cleverly. When looking at what Intel did for Centrino and what they will do for Tejas, rearranging and revising what stuff goes where in the processor can do wonders for performance and power consumption. IBM should look real hard on how they can optimize the use of real estate. A large chip will always draw more power than a smaller.

On the other hand.. IBM seems to be doing some of the things we brought up here. Stuff that will go into the 90 nm version of 970 will be extensive power saving feeatures. At least from the little blurb I can read from the ISSCC'04 agenda.

Snowy_River

Nov 14, 2003, 01:04 PM

Originally posted by jade
well i hope pbg5 have a much faster frontside bus. The current crop of centrinos have 400mhz, and version 2 will have 533mhz.

Well, I think this a pretty much a given. Even a 1.2GHz G5 PowerBook would likely have a 600MHz FSB.

Snowy_River

Nov 14, 2003, 01:08 PM

Originally posted by Henriok
No I certainly hope they can avoid that.

No I don't think so. Assume for a moment that we can get a "PowerPC 960" that is a 970 that run at 3 GHz but does not include AltiVec.. Wouldn't you prefer that one to a 1 GHz G4 in a iBook?

I think any discussion of AltiVec is relatively pointless. Apple has far too much invested in the "Velocity Engine" to turn away from it. There is no doubt in my mind that any new chip that Apple uses, for the foreseeable future, will have to have AltiVec.

Spock

Nov 14, 2003, 01:54 PM

I am Still waiting for my powerbook G6 with 5ghZ FSB and 128-bit
runnig at 7.34GhZ with a GeForce 1gb V-card and a 13 Hour battery and Mac OS 12

MacRAND

Nov 14, 2003, 02:14 PM

Originally posted by Snowy_River
I think any discussion of AltiVec is relatively pointless. Apple has far too much invested in the "Velocity Engine" to turn away from it. There is no doubt in my mind that any new chip that Apple uses, for the foreseeable future, will have to have AltiVec.
While I agree with the importance Apple places in its Velocity Engine, and its commitment to persuading IBM to design AltiVec technology into the chips Apple wants to buy, I really do not think that Apple's position is that absolute. :(

Don't you think Apple might be willing to sacrifice AltiVec (hopefully, only temporarily) in order to currently kick up the clock speed to 3GHz and at least partially solve their serious problem with heat and power?

Why stay with a G5 130nm when Apple can readily boost performance with an off the shelf (because IBM is going to make them, regardless of what Apple says or does) IBM chip at 90nm and 3GHz?

The real question will be answered by Apple R&D comparing performance and cost of a 3GHz without AltiVec with the new 2.5GHz with AltiVec, while continuing to lobby IBM for incorporation of AltiVec into chips in the future.

What if Apple finds that it can put small, cool, fast (3GHz) non-AltiVec IBM 90nm chips into PowerBooks?

The larger, slower, hotter G5 with AltiVec won't work in a PB, unless Apple comes up with a motorized circulating radiator and external trap door to add dry ice every 3 hours, along with a 2 pound snap on battery to handle the power demand.

Besides Photoshop, what other Mac software is
(1) written for and
(2) really depends on AltiVec technology?

FinalCut Pro 4?

How does the speed assist of AltiVec in a G5 compare with NEWER smaller 64-bit IBM chips having raw clock speed?

Frankly, for Macs to avoid succumbing to Intel chips, Apple will have to follow where IBM technology leads.
How many computers does Apple sell a month? 100,000?
Dell sells 100,000 computers per DAY.
The number of chips IBM sells to Apple is almost meaningless compared to other computer manufacturers. The market dictates.

Rocketman

Nov 14, 2003, 02:45 PM

Originally posted by Macrumors

Apple and IBM have laid the foundation for an architectural revamp of the PowerPC to produce an ultra low power 64-bit mobile processor.

If true this is the final nail in the Motorola coffin. The main benefit of the G4 (and the IBM G3) is power efficiency and low temperature in tight spaces.

Rocketman

MacRAND

Nov 14, 2003, 03:18 PM

Originally posted by Rocketman
If true this is the final nail in the Motorola coffin. The main benefit of the G4 (and the IBM G3) is power efficiency and low temperature in tight spaces.
Rocketman
MOTOROLA has reportedly had their computer chip division up for sale since early this year, if not before. For Apple, Motorola is no longer a serious source consideration.
Things change...:(

stingerman

Nov 14, 2003, 05:17 PM

Motorola was able to reduce power and heat in the same 130 NM process with their G4 7447 by 40%. So die shrinkage is not the only way to accomplish this.

IBM can still get a lot out of the 130 NM process by optimizing their design. The 970 really is only their first generation take at a desktop 64-bit processor, and they did an awesome job. But, we can expect version 2 at the higher clock rates to be a more refined design.

As far a 90NM process, IBM has done more than just shrink the chip, they are using newer materials, better layout, and they are adding a new power management system called PowerTune. Maybe they will even throw in a memory controller (my hope, nothing to support that in the 970.) it seems the 90NM 970 is being built for the Powerbook and iMac market. I believe the 130NM processor has a lot more life to it and within the 130NM process can go to 3+GHz. The G5 case can certainly accommodate a fan right on the processor.

MacRAND

Nov 14, 2003, 05:35 PM

stingerman, good to hear about their being lots of life left in the current IBM G5 chip even at 130 nm.

So, the Big Mac G5 SuperComputer at Virginia Tech is made out of a cluster of 2200 IBM proto chips?
Next generation is the real thing?

WHY ROUND?
I've seen images of the manufacturing process at IBM where a ROUND disk containing hundreds of G5 chips is dipped, dried and ...processed by a robotic system, and then cut up to make the individual G5 chips.
Since there is so much waste on the edges of the circular DISK containing rectangular chips,
why is it ROUND?
:confused:

applekid

Nov 14, 2003, 06:10 PM

Well, my dad guaranteed me a G5 PB by the time in college, so any news that a G5 PB that won't burn my pants sounds good to me. :)

dgbatchelor

Nov 14, 2003, 06:52 PM

Originally posted by MacRAND
stingerman, good to hear about their being lots of life left in the current IBM G5 chip even at 130 nm.

So, the Big Mac G5 SuperComputer at Virginia Tech is made out of a cluster of 2200 IBM proto chips?
Next generation is the real thing?

WHY ROUND?
I've seen images of the manufacturing process at IBM where a ROUND disk containing hundreds of G5 chips is dipped, dried and ...processed by a robotic system, and then cut up to make the individual G5 chips.
Since there is so much waste on the edges of the circular DISK containing rectangular chips,
why is it ROUND?
:confused:

Imagine a Gx PB with a wafer-ful of processing cores (and associated memory/interface logic)... If you really think about it, there is not very much space currently allocated to the CPU in a laptop... How many processor cores could fit on a wafer??? OK, cooling and battery life would be an issue... But probably not as much of an issue for a lot of low power "modest" cores than for a single "screamingly hot" (nitrogen cooled?) G5 core...

...there's more than one way to get to 1 teraflop!

wizard

Nov 14, 2003, 10:53 PM

What one has to realize is that the current vector units that Apple has available to it, simply are not fast enough. They aren't fast enough at 2GHz and they won't be fast enough at 3GHz so unless Apple has a viable alternative to replace the current vector units they won't be going anywhere.

Now Apple could very well extend the vector units or come up with a more effiecent way to execute vector instructions but it would be extremely foolish to drop the capability. While a low cost chip may fit into some product lines it would have to have the capability of at least executing vector instructions. Generally though you don't see many successful processors that have regressed performance wise. Even Intels Centrino system provides a performance boost over the alternatives.

With the above considerations out of the way we have to consider if we even know that the vector unit is slowing the processor itself down. At the moment we have no solid evidence that this is the case. Even if the vector unit was slowing the rest of the porcessor, the smart thing would be to simply run it at half the main unit speed not deletion fo the functionality.

As far as small, fast, no vector unit, G5's in Powerbooks - they would never sell. Just look a few years back in the developement of the laptop. Regression has never sold well. Beyond that, the machine would not support Apples latest software devlopments, so what good would it be. About the only place the chip would have any potential is in a scalled down iBook. That is only because I believe that the iBook serves a market that is narrowly defined and not sensitive to peformance. But that is garbage anyways as Apple has already played its cards with respect to vector operation on the current iBooks.

Now don't get me wrong you could potentially get performance equal to an Alt-vec based machine with a fast enough processor. But we are talking much faster that 3GHz. If we do hit 3GHz soon, one 970 based hardware, all that means is that a vectorless processor would have to be even faster.

As far as what will and won't work in a laptop, that is an open question. Apple could have a low power version that will work or they could be doing something smarter like making an Application specific processor. By that I mean a high integration processor optimized for low power. Frankly no one really knows what direction Apple is going here other than their publicly stated position that we should not expect a G5 (970) based laptop soon. This could mean a couple of things: 1 is that we are beating Motorola with dead fish and expect to have 2GHz G4's soon. 2 is that we have something up our sleeve that you don't have a clue about. 3. is that we are just waiting for a low power 970. Maybe even Steve doesn't know.

What I do know is that the Powerbook is in the same position as the G4 Macs where a couple of years ago. That is the overall user experience, based on hardware performance, is about to be eclipsed by the Intel world. Not a good thing for Apple considering the chages in the market.

Thanks
Dave

Originally posted by MacRAND
While I agree with the importance Apple places in its Velocity Engine, and its commitment to persuading IBM to design AltiVec technology into the chips Apple wants to buy, I really do not think that Apple's position is that absolute. :(

Don't you think Apple might be willing to sacrifice AltiVec (hopefully, only temporarily) in order to currently kick up the clock speed to 3GHz and at least partially solve their serious problem with heat and power?

Why stay with a G5 130nm when Apple can readily boost performance with an off the shelf (because IBM is going to make them, regardless of what Apple says or does) IBM chip at 90nm and 3GHz?

The real question will be answered by Apple R&D comparing performance and cost of a 3GHz without AltiVec with the new 2.5GHz with AltiVec, while continuing to lobby IBM for incorporation of AltiVec into chips in the future.

What if Apple finds that it can put small, cool, fast (3GHz) non-AltiVec IBM 90nm chips into PowerBooks?

The larger, slower, hotter G5 with AltiVec won't work in a PB, unless Apple comes up with a motorized circulating radiator and external trap door to add dry ice every 3 hours, along with a 2 pound snap on battery to handle the power demand.

Besides Photoshop, what other Mac software is
(1) written for and
(2) really depends on AltiVec technology?

FinalCut Pro 4?

How does the speed assist of AltiVec in a G5 compare with NEWER smaller 64-bit IBM chips having raw clock speed?

Frankly, for Macs to avoid succumbing to Intel chips, Apple will have to follow where IBM technology leads.
How many computers does Apple sell a month? 100,000?
Dell sells 100,000 computers per DAY.
The number of chips IBM sells to Apple is almost meaningless compared to other computer manufacturers. The market dictates.

Telomar

Nov 14, 2003, 11:26 PM

Originally posted by MacRAND
Don't you think Apple might be willing to sacrifice AltiVec (hopefully, only temporarily) in order to currently kick up the clock speed to 3GHz and at least partially solve their serious problem with heat and power?

Besides Photoshop, what other Mac software is
(1) written for and
(2) really depends on AltiVec technology? You're assuming altivec is the or even a major cause of the power issues and that's nonsense. Altivec is not being removed and is here to stay in all future chip designs.

As for what else uses it try many math libraries and the OS itself. The benefits of any SIMD unit, even SSE2 which isn't close to Altivec, are sizable. There's no modern desktop microprocessor without a SIMD unit and there's a very good reason for that.

MacRAND

Nov 15, 2003, 03:21 AM

Dave the Wizard,

When you talk about "vector" units and such, please explain to us non-tech Mac users, does that relate directly to AltiVec (vector)? Is the "velocity engine" a vector accelarator? Kindly enlighten us non-tech nerds.

Originally posted by Telomar
You're assuming altivec is the or even a major cause of the power issues and that's nonsense. Altivec is not being removed and is here to stay in all future chip designs.
Telomar, there was no "assuming" anything about AltiVec except the news stories over the last 3 years about the battle by Apple to get IBM to design AltiVec capability into the next (G5) chip to replace the aging G4. My memory about what I read and heard is that IBM resisted AltiVec but that Steve Jobs and crew insisted on its importance. Since I don't really understand the techie stuff, I came to believe that IBM partially capitulated by providing a type of AltiVec technology that would run the "velocity engine", enouch to placate Apple.

Can anyone expand on this?

Without assuming anything about the necessity of AltiVec, if IBM is hell bent on refining speed and efficiency on chips it makes for its own servers and for other users (Apple is only one of many, many customers), how likely is it that IBM will produce heat and current efficient chips that do not incorporate AltiVec?

This line of thought comes from the newly announced supercomputer unit (reportedly about the size of a dishwasher) Blue Gene/L (which IBM intends to be but one of 128 units in the 350Tflop supercomputer it is planning to install for Lawrence Livermore Labs in 2005). The whole supercomputer comprised of these 128 units of Blue Gene/L is reportedly 10x faster than #1 NEC Earth Simulator 35Tflops, but occupies about 1/8th the space. It has also been designed to be heat and current efficient.

Bottom line is that IBM is going to follow the tricky path of technological advancement towards greater efficiency and power. Achieving that, might AltiVec technology become irrelevant to IBM's goals despite Apple's needs?

When I compared Apple's sales to computer sales giant Dell, I mean to imply that Apple doesn't wield a very big stick with super giant IBM in the industrial scheme of things computer. And, it won't until it grabs a bigger market share and more than doubles its sales of computers (ignore the adorable iPod).

Having started out with a Mac LC in the late 1980's, I am ecstatic over the snappy performance of Panther on my G4 dual 1GHz for those things I do on a computer to earn a living as an attorney. A G4 is just plain awesome. :)

Panther is so snappy, I have to be careful what I touch or drag with my cursor - files disappear into some desktop folder accidentially because my finger twitches, so I have to use the Finder to find it. On a MacLC, I never dreamed that could ever be a problem.

Now, if I was a professional video editor (I do use FCPro3 and iDVD a bit) or into designing media with Adobe products, especially Photoshop, and my financial health depended on number crunching speed, then I'd worry about whether I should get a G5 dual 2GHz now, or maybe wait until February to see what the 2.5 GHz machines are like probably for the same $3,000 top level entry fee. But my living depends on word documents and spreadsheets and I only get "intellectually" excited over the prospective availability of 3 GHz plus machines.

Thank you, Apple, my G4 provides me with all the speed I need.

I love my iSight, my several color printers, but most of all I love the convenience of my two DYMO LabelWriter printers.
DYMO 330 prints beautiful Shipping Labels and the
DYMO 330 Turbo prints file folder labels at "turbo" speed.

I got 2 LabelWriters because I'm too lazy to change the labels on just one machine, unless it's to print something else - like VHS video spines, audio cassettes, or labels for ID tags.
Now I'm looking to spend money on a large Cinema Display, an SLR digital camera that has interchangable lenses, a better DV camcorder, or an upgrade in software.

When am I likely to feel the NEED to move to G5 (or, by then G6)? :confused:
Probably when the software I love or want is only compatible on a 64-bit Mac.

Compatibility with better software is what persuaded me to leave 68040 Motorola chips for IBM PPC 604, and eventually to G3 and G4. My G4 already has an added USB 2.0 PCI board, Bluetooth for my cellphone, and is tied to a wireless AirPort Extreme network that works well with a TCP/IP router to cable access, I just don't understand how or why. I don't need FireWire 800 or fiber optic audio...yet.
My Mac does everything I need it to do. Well, I might swap out the 2x SuperDrive for the new 8x as soon as Apple comes up with the driver for it.

When my G4 no longer works well, I'll happily buy another Apple.
:D

legion

Nov 15, 2003, 05:54 AM

Originally posted by MacRAND
Dave the Wizard,

When you talk about "vector" units and such, please explain to us non-tech Mac users, does that relate directly to AltiVec (vector)? Is the "velocity engine" a vector accelarator? Kindly enlighten us non-tech nerds.

The term "vector unit" in relation to processors used by Apple goes by different names:

Apple uses "Velocity Engine"
Motorola uses "AltiVec"
IBM uses "VMX"

These are all the same thing (just different names)

As for vector units, they aren't "accelerators" per se because they don't work in cooperation with the main processor but actually offload all the vector computations to themselves (hence they are "units")

MacRAND

Nov 15, 2003, 05:54 PM

Originally posted by legion
The term "vector unit" in relation to processors used by Apple goes by different names:
Apple uses "Velocity Engine"
Motorola uses "AltiVec"
IBM uses "VMX"
These are all the same thing (just different names)
As for vector units, they aren't "accelerators" per se because they don't work in cooperation with the main processor but actually offload all the vector computations to themselves (hence they are "units")
:rolleyes:
Okay, then how and why is utilization of vector technology advantageous to a Mac? Motorola and IBM make chips incorporating "vector technology" (VMX or AltiVec) and Apple designs an operating system that utilizes "vector technology" calling it a "Velocity Engine", and Adobe writes some (or all?) of its applications to utilize the Velocity Engine in Mac OS X running on compliant Motorola or IBM chips.
So how or why is this helpful to history students, video editors, media designers, doctors and lawyers who buy a Mac?
:cool:
And what if IBM says to Apple,
"We know you love the G5 PPC 970 chip, but now we have this really great new PPC 880 chip running at 5GHz clock-speed, but...uh, IBM engineers have not made VMX part of the design? Are you interested, Apple?"
What does Apple have to think about when considering such chips for its PowerBooks or towers? Does VMX matter to Apple (and us), if so, why?
:confused:

yamabushi

Nov 15, 2003, 06:34 PM

If this rumor turns out to be true, I hope that this mobile cpu is based on the 980 instead of the 970. The 980 appears to be an overall superior design and should be much more efficient than a 970 at the same clock speed. Meanwhile, the 970 is only a little bit more efficient than the G4 chips using the gcc compiler.

A mobile G6 makes more sense to me than a mobile G5. It appears the the optimized desktop G5 will make it into Powerbooks next year some time. Therefore a mobile processor released in 2005 would more likely be a G6 derivative. I expect the mobile version of the G6 would be only modestly more efficient than a G5. This would be a rapid advancement to be sure, but would be entirely reasonable if IBM plans to produce replacements for all Motorola chips.

Motorola could only save face by quickly producing an improved G4 for iBooks and eMacs. A modified version of their planned succesor to the G4 might suit this purpose and allow Motorola to gain back a little bit of respect. Knowing Motorola this isn't very likely since they probably figure that we will be stuck with a G4 until 2006 regardless of what they do to improve it.

Telomar

Nov 15, 2003, 07:14 PM

Originally posted by MacRAND
Telomar, there was no "assuming" anything about AltiVec except the news stories over the last 3 years about the battle by Apple to get IBM to design AltiVec capability into the next (G5) chip to replace the aging G4. My memory about what I read and heard is that IBM resisted AltiVec but that Steve Jobs and crew insisted on its importance. Since I don't really understand the techie stuff, I came to believe that IBM partially capitulated by providing a type of AltiVec technology that would run the "velocity engine", enouch to placate Apple.

Without assuming anything about the necessity of AltiVec, if IBM is hell bent on refining speed and efficiency on chips it makes for its own servers and for other users (Apple is only one of many, many customers), how likely is it that IBM will produce heat and current efficient chips that do not incorporate AltiVec? There was never any massive battle to have VMX added to the PPC 970, IBM helped design it. VMX is Altivec and vice versa. It's the same technology by different names.

The VMX unit was done hastily but if you look at the design of the PPC 970 the whole chip was made hastily. Nothing yet has been optimised properly and just fixing that will enable significant improvement in the heat output and efficiency of the chip. Furthermore you'll find the next iteration of the chip also has some more advanced power management features allowing further improvements. Many of the improvements made to the POWER5 so it runs cooler will be working their way across for the 90 nm PPC 970 (note this is not a 980).

The reason IBM has been less inclined to add VMX to their server chips is it isn't as useful to those markets. There's no point to add the cost when for those markets there is very little benefit to it. IBM can just as easily produce efficient chips that include VMX however.

Originally posted by MacRAND
Okay, then how and why is utilization of vector technology advantageous to a Mac?

So how or why is this helpful to history students, video editors, media designers, doctors and lawyers who buy a Mac?

And what if IBM says to Apple,
"We know you love the G5 PPC 970 chip, but now we have this really great new PPC 880 chip running at 5GHz clock-speed, but...uh, IBM engineers have not made VMX part of the design? Are you interested, Apple?" SIMD units allow you to get far more performance from less silicon space. IBM has already publicly stated they not only intend to keep VMX but improve its design slightly to better suit themselves. There is no chance they are going into a PPC design that might be lobbied at Apple at this stage and not include the VMX unit from the outset.

As for why is it helpful to those groups... With the exception of lawyers and history students, who could live on a 2GHz PPC, video, media and doctors all use programs with code that can be heavily vectorised and thus gain substantial performance benefits.

MacRAND

Nov 16, 2003, 01:36 AM

Originally posted by Telomar
As for why is it helpful to those groups... With the exception of lawyers and history students, who could live on a 2GHz PPC, video, media and doctors all use programs with code that can be heavily vectorised and thus gain substantial performance benefits.
Are we to assume that images benefit from "vector units" and the velocity engine, or is it more specifically mathematical calculations?
Spreadsheets (Excel)?
Adobe Acrobat PDF files?
Adobe Photoshop?
:cool:
Thanks for the explanation regarding IBM and VMX. Some rumors are about that seemed to indicate IBM was grudgingly placating Apple's request for chips that take advantage of their Velocity Engine. I'm pleased to hear that this is not the case.
Would you say that VMX is tied to the Mac OS emphasis on GUI orientation, or is that a myth?
:rolleyes:

Hattig

Nov 16, 2003, 01:13 PM

Vector units are actually SIMD units.

Single Instruction, Multiple Data

As you can see, this is ideal when doing a lot of work with audio, graphics, video, etc where the tasks will be performing the same action on all the data.

For example, you can fit 4 pixels worth of RGBA data in an Altivec/VMX register. Then you can simply do a multiply/add/etc instruction on the data to modify all 4 pixels in one go. In conventional integer or FP math you'd have to do each pixel at a time, and also because each pixel is comprised of 4 elements, you would have to break them up into their relevant parts first, and then perform logic to account for rounding, etc.

I'm no expert, especially on Altivec/VMX, but this is a general and very basic comparison:

So: VMX

4 load instructions to load the data into the register
1 VMX instruction to modify (VMX does auto bounds checking IIRC)
4 instructions to store the result back

= 9 instructions

Normal:

16 load instructions
16 modification instructions
16 bounds checking instructions
16 instructions to store the result back

= 64 instructions

MacRAND

Nov 16, 2003, 06:00 PM

Originally posted by Hattig
Vector units are actually SIMD units.
Single Instruction, Multiple Data
As you can see, this is ideal when doing a lot of work with audio, graphics, video, etc where the tasks will be performing the same action on all the data.
I'm no expert, especially on Altivec/VMX, but this is a general and very basic comparison:
VMX = 9 instructions
Normal = 64 instructions
Okay, so by using a software/hardware combination of Velocity Engine / AltiVec /VMX - the chip is only burdened with 9 instructions instead of 64, sort of a shortcut from "normal" chip operation IF the chip has a VMX/AltiVec design, and the software (both OS and Application) is designed to take advantage of special Vector instructions.
OS X + Photoshop + G5 or G4 = faster processing of vector information by using SIMD units
:) Right?
So, clock-speed is a constant, which can be aided by utilizing SIMD Units.

DharvaBinky

Nov 16, 2003, 08:12 PM

Originally posted by leet1
I couldn't see it loosing 2 hours just playing a game, thats way overexagerating lol

Not really. My old dell latitude C800 was the first line of Dell's featuring the P4m. Mine ran at 1.6GHz, with 1Gb of PC133 RAM. Plus the graphics chips (Gforce 420go) and the 15" UXGA screen. Now... the thing had DUAL 8cell LiIon Batteries, and during Hurricane Lili I was able to play Diablo2 for about an hour once the power had gone out. On one battery, it probably would have run for 30-40 minutes. <shrug>

:)

Dharvabinky

Spock

Nov 17, 2003, 04:29 PM

A 64bit G4 would be nice for Powermac upgrade companys like Sonnet etc..

manitoubalck

Nov 17, 2003, 08:59 PM

Originally posted by phrancpharmD
Maybe this will help propel the fledgeling fuel cell technology I've read about the past few months? If I recall correctly, fuel cells in a more "traditional" battery form factor should last weeks in current laptops before needing a "refill." Maybe this technology would at least allow several hours of use with a PPC 970 or later processor. . . Sorry, I don't have a very good grasp on the tecnhical aspects of this, but if anybody else does that's cool. . .

an interesting concept but the cost at this time is prohibitave as is the size of the current tech. Storeing Hydrogen is an issue and I wouldn't like to carry petrol in my laptop or pressurised methane.

JohnStrass

Nov 18, 2003, 01:42 PM

Originally posted by MacRAND

Now I'm looking to spend money on a large Cinema Display, an SLR digital camera that has interchangable lenses, a better DV camcorder, or an upgrade in software.

When am I likely to feel the NEED to move to G5 (or, by then G6)? :confused:
Probably when the software I love or want is only compatible on a 64-bit Mac.

When my G4 no longer works well, I'll happily buy another Apple.
:D

You may notice that once you really start using a nice digicam and a nice video camera for real, your computer will become VERY VERY slow. With the release of several apple apps today now optimized for 64 bit, for me the G4 is really end-of-life. Even try making a nice video based presentation in Powerpoint, and you will hit a wall with your G4. I just wish it wasn't so.

I cant justify a new replacement for my iMac/17/800/1gig now, but I need a laptop ASAP. I just cant stomache the thought of buying a G4 based laptop and missing out in 6 months when everything good is G5 optomized and I have a cripple.

MacRAND

Nov 18, 2003, 06:48 PM

Originally posted by JohnStrass
I cant justify a new replacement for my iMac/17/800/1gig now, but I need a laptop ASAP. I just cant stomach the thought of buying a G4 based laptop and missing out in 6 months when everything good is G5 optomized and I have a cripple. Seriously consider getting a NEW G4 :) iBook (unfortunately, SuperDrive NOT available) so you can have a decent laptop at a very reasonable price, speed equal to or better than your iMac, and much improved over my G3 iBook.
Save and get a new Dual :D G5 (1.8, 2.0, 2.5, whatever...) with SuperDrive (extra, grrrr!) :(
Speed for editing comes from:
1. raw clock speed
2. max RAM (2.5 GB plus)
3. dual chips (40% added)
4. Panther OS X (optimized for G5)
5. G5 64-bit CPU
6. Editing software (optimized for G5) FCP4.1, et alia
7. Add 1 Fast internal hard drive (if you add 2 internally, make identical HDs into RAID) - takes some jury rigging near Laser drive - talk to authorized indie
8. Fast firewire or scsi external hard drives
9. Endless supply of Mountain Dew, Coke, Pepsi or Dr Pepper:p

Snowy_River

Nov 19, 2003, 01:54 PM

Originally posted by manitoubalck
an interesting concept but the cost at this time is prohibitave as is the size of the current tech. Storeing Hydrogen is an issue and I wouldn't like to carry petrol in my laptop or pressurised methane.

Not to mention the fact that they'd probably never let you take one on a plane. Your fuel-cell battery could easily double as a mini bomb...

Ah, the age of paranoia that we live in...

Snowy_River

Nov 19, 2003, 01:56 PM

Originally posted by Spock
A 64bit G4 would be nice for Powermac upgrade companys like Sonnet etc..

That's not a possible upgrade path. I even have some doubts about the possibility of a G5 upgrade card for G4 towers. They might manage that one, but there's no such thing as a 64bit G4 and never will be.

Snowy_River

Nov 19, 2003, 02:30 PM

Hi MacRAND,

I thought I'd throw my hat into the frey, just to shed some light on this subject, as I understand it.

Originally posted by MacRAND
... the news stories over the last 3 years about the battle by Apple to get IBM to design AltiVec capability into the next (G5) chip to replace the aging G4....

Can anyone expand on this?

and

Some rumors are about that seemed to indicate IBM was grudgingly placating Apple's request for chips that take advantage of their Velocity Engine. I'm pleased to hear that this is not the case.

I think that you've mixed up some stories. This goes back to the development of the G4, not the G5. When the development of the G4 started, Apple wanted to incorporate an SIMD unit. IBM didn't like this idea, and so they continued development of the SIMD-less G3 line, while Moto did like this idea, and so became the original designer of AltiVec and the only supplier of the G4. So far as I've read, however, there was never any conflict with IBM about putting an AltiVec equivalent SIMD unit on the PPC970.

This line of thought comes from the newly announced supercomputer unit (reportedly about the size of a dishwasher) Blue Gene/L ...

For that SC they're using a completely different type of processor. Though it's in the PPC family, as I understand it, each processor is notably less powerful than their current G3 line. So, they're not really developing a new processor that's so efficient, but rather putting an already existing processor into a new architecture to get more power. (Again, as I understand it...)

Having started out with a Mac LC in the late 1980's

Just to be nit-pick... :)
I don't think that you were working on an LC in the late 80s. Why don't I think so? The LC wasn't introduced until October of 1990. ;)

...an SLR digital camera ...

Personally, I'd look at the new Canon Digital Rebel. I've been using the Rebel line for years, and they are excellent cameras. The Digital Rebel is the first digital camera that I've even considered getting (and probably will when I get enough money).

Would you say that VMX is tied to the Mac OS emphasis on GUI orientation, or is that a myth?

No, not a myth. Apple makes good use of AltiVec/VMX in the GUI of OS X. Take away AltiVec/VMX and you'd notice a considerable decrease in snappiness...

Okay, so by using a software/hardware combination of Velocity Engine / AltiVec /VMX

Just to clarify something here, Velocity Engine is simply what Apple calls AltiVec/VMX. All that's needed in order to access and gain the advantages of AltiVec is for there to be appropriate hooks in the compiler that you're using. So, OS X could, in principle, ignore AltiVec altogether, and other applications could still use it. (I believe that this is essentially how things were under OS 9 to OS X 10.0).

OS X + Photoshop + G5 or G4 = faster processing of vector information by using SIMD units
:) Right?
So, clock-speed is a constant, which can be aided by utilizing SIMD Units.
Yup. That's right, though, as I mentioned above, OS X is incidental in that equation. Properly, it would be:

Photoshop + G4 or G5 = faster processing of vector information = faster image manipulation

Here's a question that no one else has offered any kind of answer to:

Originally posted by MacRAND
WHY ROUND?
I've seen images of the manufacturing process at IBM where a ROUND disk containing hundreds of G5 chips is dipped, dried and ...processed by a robotic system, and then cut up to make the individual G5 chips.
Since there is so much waste on the edges of the circular DISK containing rectangular chips,
why is it ROUND?
:confused:

The answer has to do with how the silicon wafers are produced. They are essentially slices of a bar of silicon that is drawn out in a nearly molten state. They need to make sure that the integral stresses in the bar are uniform in order to make sure that the silicon is consistent throughout the bar. If they were to make a square (for example) bar, the corners would have a lot of stress build up, and different parts of the resulting wafer would give you chips that behave differently. So, they use the uniformly consistent shape: ROUND! :cool:

jeff.macaddict

Nov 20, 2003, 03:04 AM

The G5 is a remarkable chip. It can heat the room faster than any of its' precedors! How could they put a chip that hot into an aluminum case without the case melting and becoming fused to your working surface. Have you ever burned one of those plastic dixie cups. Just imagine melting one of those Aluminum G5 notebooks. Besides-Apple should be more concerned with power management and working on getting better batteries. Or bigger batteries-I'd opt for a removable/hot swappable Superdrive that also worked as a second battery bay- kinda like the PowerBook G3. (Two batteries capable of up to 8-9 hours work time.)

MacRAND

Nov 20, 2003, 04:10 AM

Originally posted by jeff.macaddict
Aluminum G5 notebooks... Apple should be more concerned with power management and working on getting better batteries... I'd opt for a removable/hot swappable Superdrive that also worked as a second battery bay- kinda like the PowerBook G3. (Two batteries capable of up to 8-9 hours work time.) Apple, Steve Jobs, are you listening :confused:
removable/hot swappable SuperDrive that also works as a 2nd battery bay
excellent idea! ;)

Snowy_River

Nov 20, 2003, 01:17 PM

Originally posted by jeff.macaddict
The G5 is a remarkable chip. It can heat the room faster than any of its' precedors! ...

Uh... where are you getting your information?

From everything that I've read, the G5 is not much hotter than a G4. The heat of the processor is not a major impendiment to putting it into the PowerBook. As I understand it, there are other system elements, including, but not limited to, the system controller, that do run much hotter than the equivalent G4 elements, and it is these that are the the current stumbling block for the PowerBook G5.

Now, I'm happy acknowledge when I'm wrong. So, if you have information to the contrary, I'd love to see it.

jeff.macaddict

Nov 20, 2003, 07:16 PM

Quote: Snowy_River said:

Uh... where are you getting your information?

From everything that I've read, the G5 is not much hotter than a G4. The heat of the processor is not a major impendiment to putting it into the PowerBook. As I understand it, there are other system elements, including, but not limited to, the system controller, that do run much hotter than the equivalent G4 elements, and it is these that are the the current stumbling block for the PowerBook G5.

Yes, you are correct, but no you aren't. The G5, is an extremely hot running chip, as well as the rest of the machine. The reason it may seem like it is cool, and cooler than the G4 is because there are eight (8) fans inside trying to keep the blazing PC-butt-kicking monster inside the beautiful G5 case cool. I don't think that there is a way to incorporate a cooling system extravegant enough to cool the G5.

yamabushi

Nov 23, 2003, 10:54 AM

How many times are we going to see this silly fan topic pop up? Will it never go away? :rolleyes:

Snowy_River

Nov 25, 2003, 12:36 PM

Originally posted by jeff.macaddict

Yes, you are correct, but no you aren't. The G5, is an extremely hot running chip, as well as the rest of the machine. The reason it may seem like it is cool, and cooler than the G4 is because there are eight (8) fans inside trying to keep the blazing PC-butt-kicking monster inside the beautiful G5 case cool. I don't think that there is a way to incorporate a cooling system extravegant enough to cool the G5.

Again, I ask where you're getting your information from. Is it simply the number of fans?

Here are some numbers that I was able to lay my hands on at short notice. Some of these may be old but I didn't want to spend the entire morning looking for the most up to date information.

Processor Power Dissipation (i.e. heat output)

G4 1.4GHz 42 W
G5 1.8GHz 42 W

This information was gleaned from the archives at ArsTechnica. As I said, it may not be completely up to date, but it clearly shows that the G5 is not the heat monster you seem to think it is. The reason for the nine fans (not eight) is to allow them to use slow fans and still move enough air to keep the computer cool. In other words, it's a noise issue, not (so much) a heat issue.

But, just to reitterate, the G5 itself is not "an extremely hot chip". If you actually have information to the contrary, please post it.

jouster

Mar 2, 2004, 11:48 AM

I gotta call bs on this. What laptop with a p4 has a 30 minute battery?

Yeah really. I have a Dell P4. Its battery life is around 4.5 hrs.

ionas

Mar 2, 2004, 05:05 PM

I would take this with a big grain of salt.

I am sure by 2005, Apple would be able to put the 90nm G5's into the powerbooks.

is a 64bit cpu that can execute ppc code and got an altivec unit, then apple can call it g5.

(there are many variations of the g4 that differ a lot in some cases and produce different results in some cases)

Capt Underpants

Mar 2, 2004, 06:43 PM

Me too.

Apple should set up a site for taking pre-orders for PB G5, 15'' and 17'' and advertise this to some selected places only.

Just to see what the demand really is.

Why, why, why? If Apple were to do that, it would crush all Powerbook G4 sales. I bet Powermac G4 sales fell after the Powermac G5 was unveiled. Same would happen with the Powerbook.

jouster

Mar 3, 2004, 12:08 AM

If true this is the final nail in the Motorola coffin.
Rocketman

Well, from the point of view of selling *some* of the processors required by a company that has ~2% marketshare, as opposed to selling phones.

Which Motorola is quite good at.......

http://www.bizjournals.com/phoenix/stories/2004/01/19/daily28.html

MarkCollette

Mar 3, 2004, 01:08 AM

Let me get the straight. We've seen rumors of:
- 2 GHz G3s
- G3s with VMX
- PPC 4xx
- PPC 3xx
- PPC 970fx
- PPC 980

And yet, Apple has yet to release anything but some newer xServes, which I don't even know if they've shipped yet...

Either something really great is coming down the pipeline, or someone at Apple dropped the ball, tripped over it, and fell off a cliff, landing on cactus.

Opteron

Mar 3, 2004, 02:03 AM

According to this report, the upcoming mobile PowerPC will be part of a 300 series of processors from IBM. This new mobile processor is not due to debut until 2005.

This is too long a wait, AMD have 2 mobile 64-bit processors on the market at the moment, and are using a very similar process to produce the chips. What's more is that AMD's chips are @ 130nm, not 90nm where the G5 currently is, What seems to be the big issue?

wizard

Mar 3, 2004, 02:24 PM

I hate to burst anybodies bubble here but there where rumors floating around about this device at least a year and a half ago. To maintain the same position it now has in the market Apple has no choice but to go with a more power frugal system for 64 bit portables. I just don't see a lot of happy customers if Apple where to throw the 970FX into a laptop and leave their customers with 30 min of operating time on a battery. So it is pretty much a given if Apple wants to play in the 64 bit portable segment, then it needs to reduce the power usage of the 970FX series in a dramatic manner.

I would suspect though that the processor would be due out earlier, with portable devices based on it due for delivery in 2005.

According to an anonymous report...

In light of the ongoing difficulties in incorporating desktop-class Processors (such as the PowerPC 970) into the PowerBook, Apple and IBM have laid the foundation for an architectural revamp of the PowerPC to produce an ultra low power 64-bit mobile processor.

According to this report, the upcoming mobile PowerPC will be part of a 300 series of processors from IBM. This new mobile processor is not due to debut until 2005.

wizard

Mar 3, 2004, 02:28 PM

The big issue is beig able to market a machine that fits into the PowerBooks niche in the market place. One defining feature of Apple portables over the last few years is the ability to run for a long time on Battery. Produce a portable that can't run for long on a battery and you will see Apples customer base run to something differrent.

This is too long a wait, AMD have 2 mobile 64-bit processors on the market at the moment, and are using a very similar process to produce the chips. What's more is that AMD's chips are @ 130nm, not 90nm where the G5 currently is, What seems to be the big issue?

daveL

Mar 3, 2004, 04:01 PM

I hate to burst anybodies bubble here but there where rumors floating around about this device at least a year and a half ago. To maintain the same position it now has in the market Apple has no choice but to go with a more power frugal system for 64 bit portables. I just don't see a lot of happy customers if Apple where to throw the 970FX into a laptop and leave their customers with 30 min of operating time on a battery. So it is pretty much a given if Apple wants to play in the 64 bit portable segment, then it needs to reduce the power usage of the 970FX series in a dramatic manner.

I would suspect though that the processor would be due out earlier, with portable devices based on it due for delivery in 2005.
And exactly where do you get the data that allows you state that a 970FX-based PB would only get 30 minutes of battery life? There's absolutely nothing, that I'm aware of, that would support your claim. The lower clocked 970FX (1.6-1.8 GHz) have power consumption numbers in the same ball park as current G4 PBs. In addition, the 970FX has PowerTune, which the current G4 processors don't have. In fact, it's very possible that the processor being referenced by this rumor is, in fact, the 970FX (remember the time frame of the original rumor). In short, I consider your assertions to be misleading, at best, and obviously uninformed.

MacRAND

Mar 3, 2004, 05:11 PM

daveL, let's cut the wizard some slack,
since no one knows any real figures,
his presenting an estimate or example at this point is illustrative and nothing more than a guess, certainly not to be taken literally.
I didn't.
And exactly where do you get the data that allows you state that a 970FX-based PB would only get 30 minutes of battery life? There's absolutely nothing, that I'm aware of, that would support your claim.
...
In short, I consider your assertions to be misleading, at best, and obviously uninformed.I think most of us felt the wizard's "30 minutes" was a hypothetical and not to be taken as factual.
wizard makes good points.
and, so do you, daveL.
This is more than a Rumors forum, it's a place to share fears and worries.
We are all worried about Apple avoiding another "lap burning Titanium" fiasco with "chipping paint" problems. The Aluminum PBs are such an improvement, and I would hope a G5 PB would continue to improve Apple's premium laptop, again setting a high mark of achievement within the whole computer industry, just like the PowerMac G5 has.

POINT: How are Apple's Engineers going to get the guts of a G5 Xserve into a 1-inch PowerBook case and at the same time maintain the chip clock-speed with manageable power consumption, and without too much of a heat problem that might require additional power consumption for more fans or a liquid cooling system.

When Apple's Engineers are finally successful in pulling off such wizardry, let us honor them as:
Apple's Wizards
I think the wizard is right, Apple could probably give us a PB right now, if we would be willing to accept a short (i.e. 30-minute) battery life
because it has to supply power to run the CPU, LCD screen, DVD Laser SuperDrive, 7200rpm 60GB HardDrive, fans or more sophisticated liquid circulation /cooling system (compressor?), FireWire and USB 2.0 peripherals, BlueTooth, and AirPort E card.

wizard

Mar 3, 2004, 11:08 PM

Well while it is always good to get a little support I have to stick by my guns, the figures being bandied abotu for power disapation of the 970FX, in relation to a PowerBook, are for slowlly clocked 970's and frankly are not outstanding relative to the G4. Add in the NorthBridge and the power equation becomes interesting to say the least. It is the total of all the componnents on the motherboard that will make or brake a G5 laptop.

People have to ask themsleve do I want a 64 bit laptop with about the same pefromance as a G4 with half the lifetime on battery or less. In essence this is the issue, how much of the good stuff associated with the current Powerbooks are you willing to give up for the ability to claim ownership of a 64 bit portable.

Then the issue of reasonable clock rate comes to mind. Bump a 970FX up to 1.8 to 2 GHz and its power usage goes up quite a bit. And one does have to take into account maximum power usage, which is still high with respect to the FX.

All I'm saying is that as things look at the moment a 970fx in a laptop will not offer huge advantages over a G$ based laptop. They very likely would be running at the same clock rate, and laptop design considerations for the 970, would mean that some of its advatages such as a 2:1 ebus/processor ratio will be gone.

Now there are alot of things that Apple could do to address the issues related to the 970FX in a laptop if they had the brass to do so. One possibility would be to intgrate the entire laptop into a "NorthBridge" chip. In effect the laptop would have two chips, memory and some support logic. A high level of integration in the Northbridge should reduce heat and costs. I sometimes wonder if this is what is being referenced by the slip up recently with that referrence to a SOC. While not a total SOC this would be one way to take advantage of IBM's 90nm technology. Of course a total SOC implementation would be nice also.

The point is with the little bit of information available now it does not look good for a 970fx in a PowerBook. That is a machine that keeps the same form factor and same battery lifespan. It is interesitng that some of the guys over on Ars have dug up some interesting information on the 970fx and its power usage. While it does appear that the 1.4 GHz fx may do better than first realized I'm not yet convinced that it is a hands down winner over the G4.

I would also suspect that Apple is currently rinning both of these proecssors through their paces in prototype PowerBooks to get a handle on real world performance. The problem is to get the performace that many expect out of the 970 high clock rates are required, at which point power usage goes through the roof, or your pants if it is sitting on your lap. 50 watts is a lot of power to push through the powerbooks housing.

Thanks
Dave

daveL, let's cut the wizard some slack,
since no one knows any real figures,
his presenting an estimate or example at this point is illustrative and nothing more than a guess, certainly not to be taken literally.
I didn't.
I think most of us felt the wizard's "30 minutes" was a hypothetical and not to be taken as factual.
wizard makes good points.
and, so do you, daveL.
This is more than a Rumors forum, it's a place to share fears and worries.
We are all worried about Apple avoiding another "lap burning Titanium" fiasco with "chipping paint" problems. The Aluminum PBs are such an improvement, and I would hope a G5 PB would continue to improve Apple's premium laptop, again setting a high mark of achievement within the whole computer industry, just like the PowerMac G5 has.

POINT: How are Apple's Engineers going to get the guts of a G5 Xserve into a 1-inch PowerBook case and at the same time maintain the chip clock-speed with manageable power consumption, and without too much of a heat problem that might require additional power consumption for more fans or a liquid cooling system.

When Apple's Engineers are finally successful in pulling off such wizardry, let us honor them as:
Apple's Wizards
I think the wizard is right, Apple could probably give us a PB right now, if we would be willing to accept a short (i.e. 30-minute) battery life
because it has to supply power to run the CPU, LCD screen, DVD Laser SuperDrive, 7200rpm 60GB HardDrive, fans or more sophisticated liquid circulation /cooling system (compressor?), FireWire and USB 2.0 peripherals, BlueTooth, and AirPort E card.

MacRAND

Mar 3, 2004, 11:33 PM

Well while it is always good to get a little support I have to stick by my guns, the figures being bandied about for power disapation of the 970FX, in relation to a PowerBook, are for slowlly clocked 970's and frankly are not outstanding relative to the G4. Add in the NorthBridge and the power equation becomes interesting to say the least. It is the total of all the componnents on the motherboard that will make or brake a G5 laptop.

People have to ask themsleve do I want a 64 bit laptop with about the same pefromance as a G4 with half the lifetime on battery or less. In essence this is the issue, how much of the good stuff associated with the current Powerbooks are you willing to give up for the ability to claim ownership of a 64 bit portable.

Then the issue of reasonable clock rate comes to mind. Bump a 970FX up to 1.8 to 2 GHz and its power usage goes up quite a bit. And one does have to take into account maximum power usage, which is still high with respect to the FX.

All I'm saying is that as things look at the moment a 970fx in a laptop will not offer huge advantages over a G4 based laptop. They very likely would be running at the same clock rate, and laptop design considerations for the 970, would mean that some of its advatages such as a 2:1 ebus/processor ratio will be gone.
Thanks • DaveI agree, and that apparently is why Apple is reluctant to release a mediocre PowerBook that cannot distinguish itself well from a G4 running at 1.33 or even 1.4 GHz

Snowy_River may know if you do not, what is the standard clock-speed of the 970FX that it has to get bumped to 1.8GHz?

Everybody keeps talking about 64-bit capability of the 970FX G5, yet it is my understanding that Apple has yet to tap into 64-bit capability for anything. Right? If wrong, then what?

My preference is that Apple do the PowerBook RIGHT for G5, or wait until it can, Steve Jobs' prediction or promised release date be damned. No do overs, no false starts. Macs need to work perfectly, right out of the box, the very first time. No trial runs.

Anybody want to be an early adopter of an experimental prototype for $3,000 plus?

Hattig

Mar 4, 2004, 02:41 AM

I really can't vouch for the accuracy of AppleInsider's sources, they do state (http://forums.appleinsider.com/showthread.php?s=&threadid=33845) that a 130 nm 970 @ 2.5 GHz consumes 97 W. FAR too much.

And a 970FX has a thermal design power of 39W at 2GHz, and typical usage of 25W.

If IBM can lower power usage even more by dropping the speed to, e.g., 1.6GHz, then there shouldn't be too much problems with making a thin laptop incorporating it.

In the long run, making a dedicated mobile processor is a good thing though.

Hattig

Mar 4, 2004, 02:51 AM

:rolleyes:
how rude

So how or why is this helpful to history students, video editors, media designers, doctors and lawyers who buy a Mac?

Vector Instructions are extremely useful for media applications, such as video and image processing. They are useful for anything that performs the same instruction many times in repetition.

Anyway, the VMX unit is a small percentage of the die on the 970. I really don't see them removing it at any point, or that it would be a major source of heat issues or clockability issues.

Hattig

Mar 4, 2004, 03:00 AM

Anyway, who bumped this 4 month old thread?! I didn't notice until near the end.

Opteron

Mar 4, 2004, 07:15 AM

Anybody want to be an early adopter of an experimental prototype for $3,000 plus?

I'd be a prototyoe adopte if
1: I had the money
2: it came with 3 year unlimited warrenty. :p

wizard

Mar 4, 2004, 12:06 PM

IBM is currently advertising a clock rate range of 1.4 to 2.0 GHz. Of course there most likely is a lot of upside potential with clock rate.
[quote]

Everybody keeps talking about 64-bit capability of the 970FX G5, yet it is my understanding that Apple has yet to tap into 64-bit capability for anything. Right? If wrong, then what?

Do not underestimate the utility of having a 64 bit processor! Apple has "tapped" into some features of the processor but at the moment really hasn't delivered a 64 bit OS, nor a huge number of 64 bit applications. Full ultilization is a bit off. The really good thing at the moment is more addressable memory on desktops and possibly laptops before the end of the year. As it is now 32 bit applications get the entire 32 bit address range to use, for some applitcations that is a win right now.

My preference is that Apple do the PowerBook RIGHT for G5, or wait until it can, Steve Jobs' prediction or promised release date be damned. No do overs, no false starts. Macs need to work perfectly, right out of the box, the very first time. No trial runs.

While I would hope that that is very much the attitude the reality is that, Apple has NOT been doing real well of late with respect to hardware introductions. Does any body know what happened to the G5 XServes for example?

To be honest I'm not to sure that Apple has any intentions at all of introducing a 970 based laptop. They could very well be looking at different technology.

Anybody want to be an early adopter of an experimental prototype for $3,000 plus?

daveL

Mar 4, 2004, 02:38 PM

Everybody keeps talking about 64-bit capability of the 970FX G5, yet it is my understanding that Apple has yet to tap into 64-bit capability for anything. Right? If wrong, then what?
For one, the Photoshop G5 plug-in uses the 64-bit instructions available on the 970. OS X also has math libraries that exploit the 970 64-bit instructions.

Mac Dummy

Mar 4, 2004, 02:55 PM

That would be in the next Powerbooks!

Prom1

Mar 6, 2004, 03:35 PM

If true this is the final nail in the Motorola coffin. The main benefit of the G4 (and the IBM G3) is power efficiency and low temperature in tight spaces.

Rocketman

Rocketman you may just be right. This got me thinking though....remember the wars of whom could produce the G4 (motorola -small case because their lack of efforts don't deserve my respect & haven't raised any elation from me recently). IBM was blocked by the Apple/IBM/motorola consortium contractual agreements. Now since IBM is only providing the G5 - and has a committment as mentioned in the IBM vs. Intel thread here, what options does Apple have if IBM cannot deliver performance in their chips around 5 yrs from now, especially if motorola is non-existent (for desktop/mobile cpu chips) then?? What if a new company wants to produce "G-era" chips for Apple at that time will the current consortium agreement block such a new alliance??

Hey don't shoot me I have TOTAL faith in IBM & Apple - this is their mutual second time around kicking the wintel competition.

can't wait till August/September after owning a Dual 1.8/2.0Ghz G5 I'll also be in the market for a 12" PowerBook G4 rev.2 (DVI). Yeah moto's in their but the product suits my needs for some 2 years into the future.....let's all hope that the sales for the next year from moto's cellular phones (V300/400/500/525/600/80, Mpx100/200/MPX, A920/925, etc) will be enough to fund a new mobile cpu line.