If the new powermacs are anything with which Steve can run a fake-off in something, anything, then they'll be front and center during an Expo, if not, they'll be a month later.
<strong>If the new powermacs are anything with which Steve can run a fake-off in something, anything, then they'll be front and center during an Expo, if not, they'll be a month later.</strong><hr></blockquote>
That's reasonable, assuming that Steve even wants to do bake offs any longer. I think he'd hold that distinction for the 970 tho, but it doesn't look like they will be ready until after MWNY.
There is no "the latest unix foundation". There are many different varieties of Unix -- Solaris, HP/X, AIX, SCO, MacOS X, FreeBSD, and Linux on top of that. Apple has their own flavour called Darwin and that is based on a Mach 3 microkernel and a recent version of FreeBSD. From what I know of the later versions of Mach (which is very little, BTW) I doubt Apple will ever use them as they are academic research projects. Apple will push their current Mach 3 kernel forward. They may also update to newer versions of FreeBSD when they are considered robust and stable.
[quote]<strong>
Can IBM do any worse?
</strong><hr></blockquote>
I'm really not sure what you mean by this. You seem to be implying that Apple's Unix implementation is the worst out there, which is just not true. MacOS X is a pretty darn good implementation and Apple is working hard to make it better.
An alternative reading of your statement is that IBM is somehow the worst... but this is also not true. AIX is extremely robust, mature, and stable. IBM is also leveraging the Linux community's efforts, giving their customers a choice of platform.
[quote]<strong>
Will we notice or care when the 970 hits town. It will seem like lightspeed compared to 800 mhz machines with no L3 cache... (Suck on Painter 7...)
</strong><hr></blockquote>
After all the complaining about the lack of a faster processor in the PowerMac and you're wondering if we'll notice when we finally have one?! The 970 is going to be smoking fast, don't let anybody tell you differently.
I'm really not sure what you mean by this. You seem to be implying that Apple's Unix implementation is the worst out there, which is just not true. MacOS X is a pretty darn good implementation and Apple is working hard to make it better.
An alternative reading of your statement is that IBM is somehow the worst... but this is also not true. AIX is extremely robust, mature, and stable. IBM is also leveraging the Linux community's efforts, giving their customers a choice of platform.</strong><hr></blockquote>
I think he's actually trying to ask if IBM could do worse than Motorola. And the answer is: no.
Does anyone here know anything? I swear ... geeze, I can't get through the whole thread cause people who obviously have no clue about the facts are debating them between each other ...
AltiVec - TM by Motorola
Velocity Engine - TM by Apple
Both for the same thing, the AltiVec vector subprocessor on the G4 74xx family of processors that was primarily developed by Motorola but done so in accordance with the AIM alliance. If IBM is using the name now and wasn't before that should tip someone off ...
Does anyone here know anything? I swear ... geeze, I can't get through the whole thread cause people who obviously have no clue about the facts are debating them between each other ...
AltiVec - TM by Motorola
Velocity Engine - TM by Apple
Both for the same thing, the AltiVec vector subprocessor on the G4 74xx family of processors that was primarily developed by Motorola but done so in accordance with the AIM alliance. If IBM is using the name now and wasn't before that should tip someone off ...
</strong><hr></blockquote>
Regarding AltiVec, to quote from my posting on page one:
"Someone looked up the patents, and all three are on the originals, IBM, Apple and Motorola. Any of the three can use this instruction set. The name AltiVec may be a trademark of Motorola. I was surprised to see IBM use it. It may mean that Motorola sees an advantage to let IBM and Apple use the name, since {IBM can] use the technology in any case."
Someone later pointed out that IBM may have seen an advantage to using Motorola's AltiVec name, and made an agreement with Motorola about it.
Regarding the Velocity Engine, to quote from another of my postings on page one:
"I believe Apple coined the term Velocity Engine, and was the first to use it. Regarding AltiVec, you notice that IBM refrained from using this term until now. At the microprocessor forum, IBM may have said their SIMD engine is compatible with AltiVec. They did not call it AltiVec back then."
"Apple uses the name Velocity Engine so that when (I'm sure that at the time it was an if) IBM came out with their own subset of the collaborative specs for SIMD for the PowerPC came out they would not have to change any of their marketing strategy..."
Do you see incorrect statements here? If so, point out how these statements are wrong. Regarding your own statements, I believe Motorola was not the primary developer of this technology. The SIMD Engine was defined along with the instruction set by all three, Apple, IBM and Motorola. Motorola was, however, the first to implement it in a processor, and they called it AltiVec. IBM has always been free to do their own implementation of this SIMD Engine in a processor.
"it would have helped if he'd mentioned Motorola somewhere in his message!"
They've had plenty of 'breath' wasted on them.
I figured the Potent Programmer could figure who the 'unmentionable' one was.
970. It's the one every Mac Lightwave O' holic (well, I'll count myself in there...) is waiting for Programmer. I won't buy a Mac with a G4 in it. If I'm laying out Apple-Muller prices then I expect 970 style (not outdated G4) performance and architecture to go with it.
With the 'Unix' thing. I'm not a general trivia kinda of guy, so I don't remember bullet point lists of facts. However, the point was that Apple has a flavour of Unix (BSD?)and aren't there later versions to move to as they get debugged, improved, re-compiled with better compilers?
I was fishing for a comment Because people like yer self have more software tech' knowledge than moi self.
I was wondering what level of optimisation, percentage wise, will come from moving to more recent levels of Apple's choice of Unix under the Aqua and from the re-compiling as the 'coding' tools improve. Case in point. Significant difference between say, Jag' and 10.04 (or whatever it was... )
I wouldn't expect a big speed boost to Panther..10.3 on its own, but 10.4 with 970s should surely be a leap again from 10.2? Noticeable due to both software and hardware improvements?
I was just wondering out loud if IBM, actually being more interested in performance PPC (swipe to the Moto...pow) and with heavy iron experience and selling to customers who demand 'no less' might actually provide significant software improvements due to their high level of performance in their current 'high end' markets.
I'm just wondering if they can bring any more to the table? And can they get more out of their 970 than Moto' did with the G4? I'd have thought IBM's compilers/software would take some beating? However, who cares? Even minus Altivec, the 970 looks stunning.
The law of diminishing returns? Maybe, but not until we've had 10.4 and the 970s?
Then it's all about refinement and adding features?
To those who've complained about the 970's entry Ghz being too low...
I just wanted to quickly bring up SPEC2000 for a sec. IBM *ESTIMATES* that the 1.8Ghz 970 will clock around 935 on the SPEC benchmark. Currently, the P4 2.5 Ghz clocks in around 950 or so. So, if we're to use SPEC as any kind of benchmark of performance, a 1.8Ghz 970 is (theoretically) roughly equivalent to a P4 2.5 Ghz, according to whatever it is that SPEC measures.
The thing that CAN'T be predicted is how well the 970 will run on a Mac OS X system, but considering that IBM likely worked very closely with Apple on it's design, we should be able to assume that the answer is: quite well.
Consider also that the 970 will be running on a faster bus, with faster memory and a faster video card, AND be able to speed up certain kinds of computations that use 64bit data chunks, and I think it's fair to assume that the next Gen Macs are going to haul some serious ass. Add in one more factor - that the 970 has been designed from the outset to work in tangent with 2 or more processors, and i think Mhz and speed bottlenecks will be a thing of the past on the Mac.
I'm looking VERY forward to upgrading my ancient G3400 tower to a dual or quad 970 with a cinema display. MMMMM. Now tell me 6 months is too long a wait... uh-uh.
In 6 months, the G4 will, for me, be a forgotten chapter. Prodigal Lemon will 'return' to his chosen platform.
I think the secret of the 970 will be working with more than one chip. Arrays of 2 or 4 or 8?
Everything from the humble Lightwave user (Hi Mum!) to the workstation uberlord on his 8 cpu/Maya rig (Hi Pixar?) Add in X-Grid clusters and 970s X-serve. The cooling you get going to .9. Apple are going to be all over the multimedia market like a rash.
I can see Apple going for the jugular with this. Especially when the 970 transitions to the .9! And with that, the PPC will close the gap with the x86 ers on mhz. Especially Clawhammer/Opteron which won't have high mhz. It will merely have a roughly '3.6 xp' rating equivalent to what Intel have next year or slightly ahead. Going by that? The 970 will have an XP rating of at least(!) 3.6 'XP' for the 970! Heh. That 1.8 Gigger doesn't seem so low now, eh? A 7.2 XP rating in dual formations.
Come on Pentium 4 if you think yer hard enough!
The 'battleground' will be redrawn. Good luck Intel with their Pentium 4. Ours will be 64bit and play well other cpus. Intel cranks mhz to chernobyl busting meltdowns via 3.6 gig by end of next year? Apple slaps in a dual 970 1.8 dual top end that will eat it alive! And run Wayyyyyyy cooler....in wayyyyyy cooler Cube-esque el capitan caseo!
Twin 970 Altivecs, 2 fpus on that bandwidth? No competition. I'll be in Lightwave orgasmo ball Woody Allen heaven. Ahem.
Add in Altivec at unheard of speeds and really cut loose on massive bus/bandwidth and memory.
This baby will make the G4 seem like the Pentium 2!
Comments
<strong>If the new powermacs are anything with which Steve can run a fake-off in something, anything, then they'll be front and center during an Expo, if not, they'll be a month later.</strong><hr></blockquote>
That's reasonable, assuming that Steve even wants to do bake offs any longer. I think he'd hold that distinction for the 970 tho, but it doesn't look like they will be ready until after MWNY.
<strong>CONFIRMED! MOSR says no new PowerMacs at MWNY so there MUST be new PowerMacs at the expo! w00t!</strong><hr></blockquote>
You mean MWSF. And as noted above, they haven't released at an expo for some time now.
<strong>
Although GCC has been brought to reasonable ppc codegen, I haven't heard anyone extol it as a paragon of blazing fast ppc code.
xlc (IBM's internal compiler for POWER/ppc) on the other hand...</strong><hr></blockquote>
Since Apple compiles OS X with gcc, I wonder what additional speed improvement would be made in moving to IBM's compiler.
Can IBM do any worse?
Will we notice or care when the 970 hits town. It will seem like lightspeed compared to 800 mhz machines with no L3 cache... (Suck on Painter 7...)
Lemon Bon Bon
<strong><a href="http://www.midrangeserver.com/tfh/tfh110402-story01-fig01.html"; target="_blank">http://www.midrangeserver.com/tfh/tfh110402-story01-fig01.html</a></strong><hr></blockquote>;
Interesting link, but the article it is attached to clearly states that they are making all of this up.
<strong>Are Apple on the latest unix foundation?
</strong><hr></blockquote>
There is no "the latest unix foundation". There are many different varieties of Unix -- Solaris, HP/X, AIX, SCO, MacOS X, FreeBSD, and Linux on top of that. Apple has their own flavour called Darwin and that is based on a Mach 3 microkernel and a recent version of FreeBSD. From what I know of the later versions of Mach (which is very little, BTW) I doubt Apple will ever use them as they are academic research projects. Apple will push their current Mach 3 kernel forward. They may also update to newer versions of FreeBSD when they are considered robust and stable.
[quote]<strong>
Can IBM do any worse?
</strong><hr></blockquote>
I'm really not sure what you mean by this. You seem to be implying that Apple's Unix implementation is the worst out there, which is just not true. MacOS X is a pretty darn good implementation and Apple is working hard to make it better.
An alternative reading of your statement is that IBM is somehow the worst... but this is also not true. AIX is extremely robust, mature, and stable. IBM is also leveraging the Linux community's efforts, giving their customers a choice of platform.
[quote]<strong>
Will we notice or care when the 970 hits town. It will seem like lightspeed compared to 800 mhz machines with no L3 cache... (Suck on Painter 7...)
</strong><hr></blockquote>
After all the complaining about the lack of a faster processor in the PowerMac and you're wondering if we'll notice when we finally have one?! The 970 is going to be smoking fast, don't let anybody tell you differently.
[ 12-17-2002: Message edited by: Programmer ]</p>
<strong>
I'm really not sure what you mean by this. You seem to be implying that Apple's Unix implementation is the worst out there, which is just not true. MacOS X is a pretty darn good implementation and Apple is working hard to make it better.
An alternative reading of your statement is that IBM is somehow the worst... but this is also not true. AIX is extremely robust, mature, and stable. IBM is also leveraging the Linux community's efforts, giving their customers a choice of platform.</strong><hr></blockquote>
I think he's actually trying to ask if IBM could do worse than Motorola. And the answer is: no.
<strong>
I think he's actually trying to ask if IBM could do worse than Motorola. And the answer is: no.</strong><hr></blockquote>
Heh, thanks for the clarification... it would have helped if he'd mentioned Motorola somewhere in his message!
AltiVec - TM by Motorola
Velocity Engine - TM by Apple
Both for the same thing, the AltiVec vector subprocessor on the G4 74xx family of processors that was primarily developed by Motorola but done so in accordance with the AIM alliance. If IBM is using the name now and wasn't before that should tip someone off ...
<strong>
Does anyone here know anything? I swear ... geeze, I can't get through the whole thread cause people who obviously have no clue about the facts are debating them between each other ...
AltiVec - TM by Motorola
Velocity Engine - TM by Apple
Both for the same thing, the AltiVec vector subprocessor on the G4 74xx family of processors that was primarily developed by Motorola but done so in accordance with the AIM alliance. If IBM is using the name now and wasn't before that should tip someone off ...
</strong><hr></blockquote>
Regarding AltiVec, to quote from my posting on page one:
"Someone looked up the patents, and all three are on the originals, IBM, Apple and Motorola. Any of the three can use this instruction set. The name AltiVec may be a trademark of Motorola. I was surprised to see IBM use it. It may mean that Motorola sees an advantage to let IBM and Apple use the name, since {IBM can] use the technology in any case."
Someone later pointed out that IBM may have seen an advantage to using Motorola's AltiVec name, and made an agreement with Motorola about it.
Regarding the Velocity Engine, to quote from another of my postings on page one:
"I believe Apple coined the term Velocity Engine, and was the first to use it. Regarding AltiVec, you notice that IBM refrained from using this term until now. At the microprocessor forum, IBM may have said their SIMD engine is compatible with AltiVec. They did not call it AltiVec back then."
Also to quote a posting by @homenow:
"Apple uses the name Velocity Engine so that when (I'm sure that at the time it was an if) IBM came out with their own subset of the collaborative specs for SIMD for the PowerPC came out they would not have to change any of their marketing strategy..."
Do you see incorrect statements here? If so, point out how these statements are wrong. Regarding your own statements, I believe Motorola was not the primary developer of this technology. The SIMD Engine was defined along with the instruction set by all three, Apple, IBM and Motorola. Motorola was, however, the first to implement it in a processor, and they called it AltiVec. IBM has always been free to do their own implementation of this SIMD Engine in a processor.
They've had plenty of 'breath' wasted on them.
I figured the Potent Programmer could figure who the 'unmentionable' one was.
970. It's the one every Mac Lightwave O' holic (well, I'll count myself in there...) is waiting for Programmer. I won't buy a Mac with a G4 in it. If I'm laying out Apple-Muller prices then I expect 970 style (not outdated G4) performance and architecture to go with it.
With the 'Unix' thing. I'm not a general trivia kinda of guy, so I don't remember bullet point lists of facts. However, the point was that Apple has a flavour of Unix (BSD?)and aren't there later versions to move to as they get debugged, improved, re-compiled with better compilers?
I was fishing for a comment
I was wondering what level of optimisation, percentage wise, will come from moving to more recent levels of Apple's choice of Unix under the Aqua and from the re-compiling as the 'coding' tools improve. Case in point. Significant difference between say, Jag' and 10.04 (or whatever it was...
I wouldn't expect a big speed boost to Panther..10.3 on its own, but 10.4 with 970s should surely be a leap again from 10.2? Noticeable due to both software and hardware improvements?
I was just wondering out loud if IBM, actually being more interested in performance PPC (swipe to the Moto...pow) and with heavy iron experience and selling to customers who demand 'no less' might actually provide significant software improvements due to their high level of performance in their current 'high end' markets.
I'm just wondering if they can bring any more to the table? And can they get more out of their 970 than Moto' did with the G4? I'd have thought IBM's compilers/software would take some beating? However, who cares? Even minus Altivec, the 970 looks stunning.
The law of diminishing returns? Maybe, but not until we've had 10.4 and the 970s?
Then it's all about refinement and adding features?
Fishingly,
Lemon Bon Bon
[ 12-19-2002: Message edited by: Lemon Bon Bon ]</p>
<a href="http://www.eetimes.com/semi/news/OEG20021216S0030"; target="_blank">http://www.eetimes.com/semi/news/OEG20021216S0030</a>;
also makes explicit mention of some variant of PowerPC to be produced in the near future on the 90-nm process, but does not state which one.
boots
I just wanted to quickly bring up SPEC2000 for a sec. IBM *ESTIMATES* that the 1.8Ghz 970 will clock around 935 on the SPEC benchmark. Currently, the P4 2.5 Ghz clocks in around 950 or so. So, if we're to use SPEC as any kind of benchmark of performance, a 1.8Ghz 970 is (theoretically) roughly equivalent to a P4 2.5 Ghz, according to whatever it is that SPEC measures.
The thing that CAN'T be predicted is how well the 970 will run on a Mac OS X system, but considering that IBM likely worked very closely with Apple on it's design, we should be able to assume that the answer is: quite well.
Consider also that the 970 will be running on a faster bus, with faster memory and a faster video card, AND be able to speed up certain kinds of computations that use 64bit data chunks, and I think it's fair to assume that the next Gen Macs are going to haul some serious ass. Add in one more factor - that the 970 has been designed from the outset to work in tangent with 2 or more processors, and i think Mhz and speed bottlenecks will be a thing of the past on the Mac.
I'm looking VERY forward to upgrading my ancient G3400 tower to a dual or quad 970 with a cinema display. MMMMM. Now tell me 6 months is too long a wait... uh-uh.
<img src="graemlins/lol.gif" border="0" alt="[Laughing]" />
In 6 months, the G4 will, for me, be a forgotten chapter. Prodigal Lemon will 'return' to his chosen platform.
I think the secret of the 970 will be working with more than one chip. Arrays of 2 or 4 or 8?
Everything from the humble Lightwave user (Hi Mum!) to the workstation uberlord on his 8 cpu/Maya rig (Hi Pixar?) Add in X-Grid clusters and 970s X-serve. The cooling you get going to .9. Apple are going to be all over the multimedia market like a rash.
I can see Apple going for the jugular with this. Especially when the 970 transitions to the .9! And with that, the PPC will close the gap with the x86 ers on mhz. Especially Clawhammer/Opteron which won't have high mhz. It will merely have a roughly '3.6 xp' rating equivalent to what Intel have next year or slightly ahead. Going by that? The 970 will have an XP rating of at least(!) 3.6 'XP' for the 970! Heh. That 1.8 Gigger doesn't seem so low now, eh? A 7.2 XP rating in dual formations.
Come on Pentium 4 if you think yer hard enough!
The 'battleground' will be redrawn. Good luck Intel with their Pentium 4. Ours will be 64bit and play well other cpus. Intel cranks mhz to chernobyl busting meltdowns via 3.6 gig by end of next year? Apple slaps in a dual 970 1.8 dual top end that will eat it alive! And run Wayyyyyyy cooler....in wayyyyyy cooler Cube-esque el capitan caseo!
Twin 970 Altivecs, 2 fpus on that bandwidth? No competition. I'll be in Lightwave orgasmo ball Woody Allen heaven. Ahem.
Add in Altivec at unheard of speeds and really cut loose on massive bus/bandwidth and memory.
This baby will make the G4 seem like the Pentium 2!
Hotrod PPC, BAY-BEEEEE!
Lemon Bon Bon