<strong>Not to throw fuel on anything but keep in mind IBM has a roadmap for the current G3 and it is fairly reasonable. There is no real need to replace it with something from Motorola unless it is really worth it.</strong><hr></blockquote>
<strong>Not to throw fuel on anything but keep in mind IBM has a roadmap for the current G3 and it is fairly reasonable. There is no real need to replace it with something from Motorola unless it is really worth it.</strong><hr></blockquote>
I've heard some things about the IBM G3 (still having legs) too but what I have a problem with is the fact that Apple has moved all but one line away from the G3.
A G3 made with a new faster bus and altivec would be nice and I'm willing to bet would/could be faster than todays G4's would have to be given a new name. It would be marketspeak-hell to try and explain why a G3+++ powered PowerBook or iMac or eMac would be an 'upgrade' to todays 'G4' powered versions.
Yea in the end it's really just a problem for the marketing folks to deal with.
<strong>Yea in the end it's really just a problem for the marketing folks to deal with.</strong><hr></blockquote>
Not really. The only thing that makes the G4 is the velocity engine and a SIMD unit is scheduled for addition to the G3 in the next major revision. SMP capabilities, improved double precision, <goes on with too long a list>, are all scheduled as well. It is also aimed at being a low power solution.
I really wouldn't be surprised if Apple drops Motorola completely.
Not really. The only thing that makes the G4 is the velocity engine and a SIMD unit is scheduled for addition to the G3 in the next major revision. SMP capabilities, improved double precision, <goes on with too long a list>, are all scheduled as well. It is also aimed at being a low power solution.
I really wouldn't be surprised if Apple drops Motorola completely.</strong><hr></blockquote>
It depends on what IBM considers a "SIMD unit." They have more than one half-assed SIMD implementation attached to a G3 already. They might be perfectly suited to their intended applications - or at least, they might be a compromise acceptable to all parties - but they will not replace a solution as powerful and elegant as AltiVec.
Apple probably won't be too keen on supporting no AltiVec (current and past G3s), partial AltiVec -or worse, a totally different set of instructions! (forthcoming G3s), and full AltiVec (G4s, possibly GPULs). But if IBM implements a full AltiVec unit, their chip isn't really a G3 anymore, is it?
I liked that suggestion a few messages back about a G3 with Apple's Velocity Engine, which could be the same VMX SIMD engine that IBM will use in the GPUL. For low end Macs, there would be no need for having dual processors or imporved FPU. It could be called a G4 of sorts, with a couple letters to distinguish it from a standard G4.
The next gen G3 is meant to be switching to RapidIO. Add a SIMD unit that is fully compatible with the altivec instructions and you basically have enough to call it a G4 from a marketing point of view.
How likely that is to happen and in what sort of timeframe is another matter.
It is possible Apple could be intending to phase out Motorola altogether. I imagine whatever is released next will probably tell somewhat.
I remember a while ago that the G5 really would be called the X1 . . . . I dunno if that's right or not, but hear me out.
Just call the G3+altivec+RIO+SMP, etc. a X1, introduced along side the X2 - the GPUL.
Apple suddenly has entirely refreshed, new product names, and skirts around the issue that, well, the G4 was faster and better than the G3, but now the G3 is better.
<strong>The only big changes in the 7400 was MPX and AltiVec.
Would a 750+VMX be a G4, or would it be a G3.5?
Would it need to have an improved system bus (like RIO) to be a G4?
Barto</strong><hr></blockquote>
Yes, but the 7455 is a G4 as well and it has longer pipelines, a bunch of internal optimizations, an on-chip L2 cache, and an L3 cache controller. The original G4 also had an improved FPU over the G3.
Personally I don't think it matters -- if an IBM "G3+VMX" chip arrives they'd just call it a G4 and it would only see use in the iBook where its low cost / low power advantage is needed. The TiBook would probably continue with the Motorola 74xx family. Either chip maker (or both) could introduce RapidIO versions of their chips. Only time will tell what Apple's naming scheme will be, and who really cares since that just marketing noise anyhow? The actual name of the thing is irrelevant -- what matters is when we're going to get the hardware features that we have typically been ascribing to the "G5" meaning "next generation PowerMac".
Considering GPUL will be neutered in comparison to the POWER4, especially in the area of L3 cache, I doubt it will be able to keep up with Pentiums at twice the clockspeed even if it is in a dual core configuration like the POWER4.
Also consider that AMD's Athlon are rated as a 12-way superscalar design while GPUL is only an 8-way design. Folks, the GPUL is nothing to drool over.
<strong>Back on the GPUL train... here is a quote from MacNN. Anyone else see this there? Moki care to explain?
</strong><hr></blockquote>
What would you like him to explain? IBM will be talking about their new PPC design (referred to as the GPUL) at MDF next week and this will be used in future Apple machines. I don't think this requires much more explanation. I think Moki was just telling Codename that his negativity is unwarranted and that IBM is capable of building industry-leading processors (unlike Motorola).
<strong>I think Moki was just telling Codename that his negativity is unwarranted and that IBM is capable of building industry-leading processors (unlike Motorola).</strong><hr></blockquote>
Actually, to read it, it sounds like Moki is telling people to not let their imaginations get the better of them. GPUL might be great for us, but it may not be great for all.
When Apple releases PowerMacs using the GPUL I think they will most likely have single processor versions and market it as dual processors... or dual core. And rightfully so. Despite being on the same die, they will be 2 distinct processor cores.
[quote] IBM Corp. may give a peek into Apple Computer Inc.'s 64-bit future when it details a new version of its Power4 microprocessor next week. Aimed for use in desktops and low-end servers, the 64-bit Power4 could be IBM's first PowerPC-compatible chip to support the Altivec multimedia instruction extensions defined by Apple and Motorola Inc.
The IBM device is one of about 18 new processors that will be described at the Microprocessor Forum 2002, to be held Oct. 14-17. "I expect there will be a fair amount of discussion about this part," said Peter Glaskowsky, editor-in-chief of The Microprocessor Report, which hosts the annual gathering.
Unlike IBM's original Power4, the device to be described next week will use one, not two, internal processor core and will support extensions that make the Power4 compatible with the PowerPC architecture. "Because it supports a full 32-bit environment, this chip should be able to boot the Mac OS just fine," Glaskowsky said.
The same article also says single core, not dual core. And Moki's statement (and all his previous postings on the subject) have been rather optimistic about the new IBM processor's capabilities, and the likelyhood of Apple adopting it.
If a single 'GPUL' core is twice the performance of a G4 at the same clock speed eg 1 gig...then at 2 gig, this 'GPUL' baby is going to blow a dual 1.25 Moto gigger out the water. (And my pet fav'...it'll have an extra fpu unit...)
I'll take that on the bandwidth numbers they're talking about.
Lemon Bon Bon
I've already started saving: lodged my first 'big check' in the 'savings account' for this one. Amen.
I prefer a single core architecture with a good altivec unit totally independant, than a dual core architecture with bad tricks in order to save transistors.
A single core means less transistors and less heat. With 50 millions of transistors you can make a very sophisticated single core chip, that can be use in The I mac and powerbook line some months later.
If the chip is a dual core, we will have to wait a long time before seing it in a I mac or a powerbooK . Apple sell more I mac than powermac, it's very important for him to have good processors for all his product line.
Too bad about it being single core. If it were dual core, then Apple would have no choice but to offer "dual" CPU macs. But now they will probably intro it with only single CPU towers, and wait until the performance gap widens again to go dual.
Comments
<strong>Not to throw fuel on anything but keep in mind IBM has a roadmap for the current G3 and it is fairly reasonable. There is no real need to replace it with something from Motorola unless it is really worth it.</strong><hr></blockquote>
If it has AltiVec, it might be worth it.
<strong>Not to throw fuel on anything but keep in mind IBM has a roadmap for the current G3 and it is fairly reasonable. There is no real need to replace it with something from Motorola unless it is really worth it.</strong><hr></blockquote>
I've heard some things about the IBM G3 (still having legs) too but what I have a problem with is the fact that Apple has moved all but one line away from the G3.
A G3 made with a new faster bus and altivec would be nice and I'm willing to bet would/could be faster than todays G4's would have to be given a new name. It would be marketspeak-hell to try and explain why a G3+++ powered PowerBook or iMac or eMac would be an 'upgrade' to todays 'G4' powered versions.
Yea in the end it's really just a problem for the marketing folks to deal with.
Dave
<strong>Yea in the end it's really just a problem for the marketing folks to deal with.</strong><hr></blockquote>
Not really. The only thing that makes the G4 is the velocity engine and a SIMD unit is scheduled for addition to the G3 in the next major revision. SMP capabilities, improved double precision, <goes on with too long a list>, are all scheduled as well. It is also aimed at being a low power solution.
I really wouldn't be surprised if Apple drops Motorola completely.
<strong>
Not really. The only thing that makes the G4 is the velocity engine and a SIMD unit is scheduled for addition to the G3 in the next major revision. SMP capabilities, improved double precision, <goes on with too long a list>, are all scheduled as well. It is also aimed at being a low power solution.
I really wouldn't be surprised if Apple drops Motorola completely.</strong><hr></blockquote>
It depends on what IBM considers a "SIMD unit." They have more than one half-assed SIMD implementation attached to a G3 already. They might be perfectly suited to their intended applications - or at least, they might be a compromise acceptable to all parties - but they will not replace a solution as powerful and elegant as AltiVec.
Apple probably won't be too keen on supporting no AltiVec (current and past G3s), partial AltiVec -or worse, a totally different set of instructions! (forthcoming G3s), and full AltiVec (G4s, possibly GPULs). But if IBM implements a full AltiVec unit, their chip isn't really a G3 anymore, is it?
The name of IBM's new chip is the PowerPC 640.
The 640's front-end is ApplePI running at 6.4GB/s.
ApplePI is based on HyperTransport, but with some features dropped and features from MPX (like intercept) added.
The 640's execution unit is the same dual core+L1 and shared L2 as Power4.
The 640 has AltiVec/VMX/VE, but it's a "soft" implementation.
The 640 is a small, single-chip package.
Starting production mid-2003.
Barto
[ 10-07-2002: Message edited by: Barto ]</p>
Would a 750+VMX be a G4, or would it be a G3.5?
Would it need to have an improved system bus (like RIO) to be a G4?
Barto
How likely that is to happen and in what sort of timeframe is another matter.
It is possible Apple could be intending to phase out Motorola altogether. I imagine whatever is released next will probably tell somewhat.
I remember a while ago that the G5 really would be called the X1 . . . . I dunno if that's right or not, but hear me out.
Just call the G3+altivec+RIO+SMP, etc. a X1, introduced along side the X2 - the GPUL.
Apple suddenly has entirely refreshed, new product names, and skirts around the issue that, well, the G4 was faster and better than the G3, but now the G3 is better.
It's too late. Time for bed
<strong>The only big changes in the 7400 was MPX and AltiVec.
Would a 750+VMX be a G4, or would it be a G3.5?
Would it need to have an improved system bus (like RIO) to be a G4?
Barto</strong><hr></blockquote>
Yes, but the 7455 is a G4 as well and it has longer pipelines, a bunch of internal optimizations, an on-chip L2 cache, and an L3 cache controller. The original G4 also had an improved FPU over the G3.
Personally I don't think it matters -- if an IBM "G3+VMX" chip arrives they'd just call it a G4 and it would only see use in the iBook where its low cost / low power advantage is needed. The TiBook would probably continue with the Motorola 74xx family. Either chip maker (or both) could introduce RapidIO versions of their chips. Only time will tell what Apple's naming scheme will be, and who really cares since that just marketing noise anyhow? The actual name of the thing is irrelevant -- what matters is when we're going to get the hardware features that we have typically been ascribing to the "G5" meaning "next generation PowerMac".
moki
Addicted to MacNN
Posts: 1491
Location: Rochester, NY
Registered: Sep 2000
Status: Offline
Posted on : 10-09-2002 03:32 AM
--------------------------------------------------------------------------------
Originally posted by Codename:
quote:
--------------------------------------------------------------------------------
Considering GPUL will be neutered in comparison to the POWER4, especially in the area of L3 cache, I doubt it will be able to keep up with Pentiums at twice the clockspeed even if it is in a dual core configuration like the POWER4.
Also consider that AMD's Athlon are rated as a 12-way superscalar design while GPUL is only an 8-way design. Folks, the GPUL is nothing to drool over.
--------------------------------------------------------------------------------
Spoken by someone who has no idea what the GP-UL is actually like.
October 15th will be fun, folks, hang in there.
__________________
Andrew Welch / el Presidente / Ambrosia Software, Inc.
Carpe Aqua -- Snapz Pro 1.0.6 for OS X is released!
<strong>Back on the GPUL train... here is a quote from MacNN. Anyone else see this there? Moki care to explain?
</strong><hr></blockquote>
What would you like him to explain? IBM will be talking about their new PPC design (referred to as the GPUL) at MDF next week and this will be used in future Apple machines. I don't think this requires much more explanation. I think Moki was just telling Codename that his negativity is unwarranted and that IBM is capable of building industry-leading processors (unlike Motorola).
<strong>I think Moki was just telling Codename that his negativity is unwarranted and that IBM is capable of building industry-leading processors (unlike Motorola).</strong><hr></blockquote>
Actually, to read it, it sounds like Moki is telling people to not let their imaginations get the better of them. GPUL might be great for us, but it may not be great for all.
We all know that October 15, Apple will release quad GPUL PowerMacs and dual PowerBooks!
the quote
[quote] IBM Corp. may give a peek into Apple Computer Inc.'s 64-bit future when it details a new version of its Power4 microprocessor next week. Aimed for use in desktops and low-end servers, the 64-bit Power4 could be IBM's first PowerPC-compatible chip to support the Altivec multimedia instruction extensions defined by Apple and Motorola Inc.
The IBM device is one of about 18 new processors that will be described at the Microprocessor Forum 2002, to be held Oct. 14-17. "I expect there will be a fair amount of discussion about this part," said Peter Glaskowsky, editor-in-chief of The Microprocessor Report, which hosts the annual gathering.
Unlike IBM's original Power4, the device to be described next week will use one, not two, internal processor core and will support extensions that make the Power4 compatible with the PowerPC architecture. "Because it supports a full 32-bit environment, this chip should be able to boot the Mac OS just fine," Glaskowsky said.
<hr></blockquote>
Sounds pretty cool to me...
I'll take that on the bandwidth numbers they're talking about.
Lemon Bon Bon
I've already started saving: lodged my first 'big check' in the 'savings account' for this one. Amen.
A single core means less transistors and less heat. With 50 millions of transistors you can make a very sophisticated single core chip, that can be use in The I mac and powerbook line some months later.
If the chip is a dual core, we will have to wait a long time before seing it in a I mac or a powerbooK . Apple sell more I mac than powermac, it's very important for him to have good processors for all his product line.