"Moreover, Book E defines ways that application-specific processing units (APUs) can be linked to a PowerPC processing core. The AltiVec PowerPC vector processing unit and instructions would be considered an APU under the Book E definition. Asked whether IBM will develop a PowerPC that includes an AltiVec coprocessor, Elliott Newcombe, PowerPC product marketing manager at IBM's Research Triangle Park facility, said IBM is considering adding an AltiVec APU to one of IBM's designs. "AltiVec could be an example of one of these coprocessors that could be plugged in [to a Book E-compliant PowerPC]," Newcombe said. "Nothing precludes IBM from doing that, but I cannot comment on whether a design is in progress. The market will decide whether we do that, and I can just say 'stay tuned.' "</strong><hr></blockquote>
Well, I've just reread the book E specs, and it ain't so. An APU cannot define a new set of registers, a SIMD engine can be defined as an APU, but it has to reuse an existing set of registers.
"Well, it sounds like this CPU is not in Apple's future -- the "over 160" vector instructions are not AltiVec (even though AltiVec has 162 instructions), and there are technical issues that would prevent AltiVec from ever marrying with Power4 or its successors. Furthermore, the guy came right out and said that they have pitched the desktop Power4 to Apple, but Apple was not interested."
So IBM pitched the Power4 to Apple even though it will never work with AltiVec aka Velocity Engine? None of what this guy said makes sense. Sorry, I think this whole rumor is just a bunch of BS.
Hasn't anyone read the information at the links that I posted or any of m previous reasoning?
There is absolutely no reason why the vector (SIMD) unit on this new chip isn't actually AltiVec (Velocity Engine) under their own name (VMX?) Again, Why on earth would IBM waste the time and effort to design a completely new SIMD unit that's only "compatible" with AltiVec when they are free to actually use AltiVec as it is?
I've talked with an ex PPC designer from Somerset and he's been delving into this more than I have and going over the docs and comparing them to what the word on the net is. He says that he'd be VERY surprised if this vector unit was NOT AltiVec. In fact he believes that everything points to it bieng exactly that. Still, it raises a bunch of questions. I listed them earlier, so you might wan't to revisit them and ponder them for a while.
Powerdoc's post on the top of this page is moving in the right direction, and he is, uh... talking to the right person. Also, the answer to Powerdoc's final question to me is "yes." That's all I can say without someone getting a pink slip.
[[[Apparently Apple will have two options next year with regard to CPU. Moto is working on a "next generation desktop proc for Apple" (direct quote), and although my cousin, who is my Moto guy, is not in on the project, his friends are, and that is all they will ever say. Time frame = 6-9 months. I don't know if that's shipping in quantity to Apple, first one rolls out, or what. ]]]
Well, I have a few trusted sources as well. I wonder if the guys you know worked with the person I speak with. The guy I speak with worked in the PPC design department over at the Somerset facility. He's been keeping in contact and reading up on the info being discussed on the web as well as whatever he has access to from Moto and IBM. I'm not sure what he still has access to and I'm not sure if any of his NDAs still apply. In all, he stated that He'd be VERY surprised if the vector unit wasn't AltiVec with IBM's one name tacked on.
[[[Time frame = 6-10 months to "completion". This may be what is going to be discussed in October. Also, I have no idea what "completion" means in this context.]]]
hmmmm. If I'm not mistaken I think IBM is prohibited from making any sort of "vaporware" announcements. Something about it being in one of their antitrust settlements from long ago. That said, I'd have to assume that this chip is already done and will be announced in October. Can anyone verify the above as an IBM settlement stipulation?
There's alot of irony in that statement. If there isn't that much to do, then why haven't they done it already?
More annoying is that clearly you see that 'Mot doesn't really need to do much' and you're not the only one. That's what's been so annoying about the G4. If an extra FP unit had been added then it would have held its own more. The Athlon destroys the G4 in Lightwave...and the Pentium 4's optimisations in Lightwave destroy the Athlon. FP, my area of interest I guess as a Lightwave 7.5 owner.
"Take a 7455, add an FP unit, a memory controller and RIO support, keep it MP friendly, and die-shrink it to .09. There's your screaming 32-bit chip. The biggest advances would be the additional FP and (especially) the RapidIO, since the current 7455 is hobbled by low bandwidth."
A part of me says that is where we should be now...if Apple wants to 'innovate its way out of this recession'. Innvotion should be more than skin deep.
Still, it that's what they put under the G5 nom da plum in 6 months time then I'll take it on the chin. I've waited years for a decent 'power'Mac. I won't leave it longet than this next year.
I appreciate Gamblor's 'soothing'. When I first heard the 'Nasa' rumour, I almost blew a gasket. I thought, Motorola won't deliver, Apple are now turning down a great Chip from IBM...what now?
In calmer morning light I guess Apple will probably create a healthy competition from both IBM/Moto' and leave their options open ala Nvidia and Ati. Having only one supplier for graphics cards for years was very damaging...and there's no doubt being roped into Moto' only for cpus has done Apple immense damage. The sales of 'power'Macs speak for themselves. The comment on a 'core' will modular elements and Rio adoption being virtually mandatory between IBM/Moto then I guess there's cause to be optimistic. Apple can 'choose' what's right for them in consumer and workstations.
Choice of supplier is the watchword. We can see what competition has done for graphics cards and cpus on the x86 side. There's alot to be said for competition.
Part of me, inside feels like JD's post. It scary...despite all the great things Apple are doing...all the publicity they're getting for their machines, stores and os, they're still barely able to stand still! Dell are giving them the beating of a lifetime in education. If you extrapolate that to publishing then Apple are in serious trouble.
They MUST address the performance issue in workstations.
They MUST address the price issue in education.
The only way you can make things cheaper is to sell more of them. 10% is the holy grail for Apple.
They've got great OS, great software, beyond the box thinking, adopted open standards, have more games and developers than they've ever had...yet...why do I get the feeling that they're still sinking?
Your rabid pessimism is unwarranted. Apple's global market share has grown and their software has done a 180. Although the hardware is lagging and their educational sales have dropped, they are doing well and are poised to succeed.
Apple is a thriving company with a niche market, they will not likely overturn MS, and so they will never overtake Dell. They could succeed in making MS operating systems non-ubiquitous, however. And for every Powermac sale they are losing due to hardware, they are probably gaining another plus some due to the OS and iApps. I'm sure the Powermac drought is temporary, albeit lengthy, but you will survive
PS. I don't think it's a price issue in education (since I'm a sys admin at a University). Most administration staff in school systems run Windows so they buy it for the kids, most sys admins in educational facilities know Windows, so they buy it for the schools.
Dell support is well above Apple for people like me:
Me: "My hard drive is busted, and I did this, this and this just to make sure."
Dell: " OK, where do you want the replacement sent?"
Next morning there is a replacement hard drive at my office and a prepaid return sticker ready to send the broken drive back. All free of additional charges. That is why we buy Dell over any PC vendor or Apple if Apple doesn't provide a reason to use them over a PC.
b8rtm8nn, I think pessimism is well warranted, but you sure have a strong point talking about Apple's poor support! When my old Studio Display 21 blew up I had to send it all the way across Europe to the Netherlands, at my expenses. Fortunately it was cheaper than I thought... But I had to stay for one whole month without it, so basically Apple forced me to buy a new, chep, all-beige 17 as a temporary replacement...
(Yes, the display was still covered by its one-year--BTW, isn't 3-year standard industry-wide?--warranty!)
I can understand very well now why Apple's .edu share is less than half than Dell's... <img src="graemlins/oyvey.gif" border="0" alt="[No]" />
But I can only forgive lack in raw performance up to a certain point.
Apple better have caught up with Intel/AMD in a year or two, or my next computer ain't gonna be a Mac.
If you search my post history (which you probably won't bother to do), you will see that I have belittled JYD, Lemon Bon Bon etc for being so anti-Apple due to the lack in raw performance in the Power Mac range.
Mac OS X makes up for some of that loss, but it's getting to the point where it doesn't make up for the G4's inability to compete with Intel/AMD. The G4 is great for consumer/portable products (cheap, low temp), but not for the Power Mac.
The Power Mac demo I saw at the Apple Road Show a few weeks ago didn't floor me like the latest PCs I've seen.
Apple, I'm a customer. Treat me better. Find a way to give me more performance, or I'm leaving.
Reading this through shows that Rivina and guTS are 2 different prototypes. Appears that guTS is a precursor to Rivina and Rivina itself is not a finished processor but a precursor to something else. I think the prototype (?) before Rivina is the Pulsar. It has larger L1 caches (128KB vs 64KB) but more L1 cache latency. Rivina was designed to test an experimental CPU floorplan to improve chip frequency beyond the improvements that extending the pipeline can provide. You have to admit, on a .22 micron process (220nm) 1.15GHz is impressive albiet at 120+ watts. With die shrinks to 180nm, 130nm (should yield 1.8-2.0GHz processors), and eventually 90nm, this new family of processors should scale well.
And here is an interesting quote from the article:
If one wants to predict what a next generation of microprocessors might look like, examining the system in which they are used usually gives a good clue. Present-day designs have migrated L2 caches and multiple cores onto the chip. It is thus interesting to examine integrating DRAM on the die with the processor. IBM has demonstrated the capability of combining high-performance DRAM, the next off-chip component, with high-performance logic on the same die [38] (Figure 34). When logic transistor performance no longer suffers, the power and density advantages of DRAM make it a good candidate for an on-chip, higher-level cache. For a large enough cache size, a DRAM can even outperform an SRAM in latency because the delay to cover the distance over the array to the processor dominates the time needed to access the subarray. Promising as this technology may be, it is by and large complementary to the improvements that are expected in the core of the processor.
<strong>Why couldn't IBM have a SIMD unit that used existing registers and still decode Altivec instructions for processing in its own SIMD unit?</strong><hr></blockquote>
Code that uses the AltiVec registers doesn't expect the standard registers to get clobbered when AltiVec instructions are used, yet this is what would happen if the AltiVec instructions were mapped onto the standard ones.
Adding a SIMD instructions that use the standard registers is much simpler than adding a whole external vector unit, and it would generally be transparent to all software that didn't use the SIMD instructions -- including the OS. It doesn't increase the amount of processor state that needs to be preserved across context switches and function calls. It makes it easier to "do a little bit of SIMD" since the data is already in your standard registers. This approach makes more sense in the embedded space (hence Motorola uses it in the 85xx series), whereas a desktop or server processor can afford the full vector unit which is a relatively small amount of transistors compared to the huge caches and other parts of a high performance chip.
IBM is not foreign to the idea of multiple cores on one chip (die). Making their VMX unit a separate core with it's own set of registers and L1 cache but a shared L2 cache and external bus would be a compromise, especially if they intend this chip to service Apple's needs as well as their own workstation needs (rs6k 3d graphics workstations such as the 270 and 170 model).
Hmmmm. (Think Marge, Simpsons.) Thanks Bartman, you almost cheered me up there...
"Apple's global market share has grown and their software has done a 180. Although the hardware is lagging and their educational sales have dropped, they are doing well and are poised to succeed."
There are more Macs out there than ever before. Compared to 1996...maybe they are selling more Macs due to the smaller slice of a larger pie sydrome. But Apple with 2.5 World Wide Market share have had a shrinking share of the the growing pie. They're slipping faster than they're climbing relatively. It'll keep them in business for the forseeable future. That and the 4 billion in cash.
"Apple is a thriving company with a niche market, they will not likely overturn MS, and so they will never overtake Dell. They could succeed in making MS operating systems non-ubiquitous, however."
Maybe the 'switch' campaign will help get the 'Mac OS' message across. Apple doen't seem to show their OS, their products off in the Switch ads. Perhaps they should along with the emotive PCs make life hard tac. Getting to that 10% market share looks to be a long haul.
"And for every Powermac sale they are losing due to hardware, they are probably gaining another plus some due to the OS and iApps. I'm sure the Powermac drought is temporary,"
Perhaps. But the 'power'Mac drought has been going on for so long it's not funny. What's worse? Waiting for the G5 or sexual frustration?
"albeit lengthy, but you will survive"
I'm gagging for it. A G5 that is.
"PS. I don't think it's a price issue in education"
Nope. I simpley don't agree with this. Price for people in education (and I work in a school myself where they're upgrade to Pentium 4s...) is a significant issue. It's the different between an extra few PCs...people go PCs. The mhz issue has hurt them in my school. Price also. I've tried 'bleating' the overall cost of ownership mantra. People just see the 'mainstream' view. Everybody has Windows. Mhz matters and so does the bottom line.
The 'useage' semantics of 'ease of use', iphoto automatically taking photos from yer camera, the iapps, appleworks etc. It's like you're talking a different language to x86 folks. I really think Apple have got to come out with a thin client, something below an eMac (which is simply not cheap enough in the UK I'm afraid...) Headless with onboard graphics etc. Barebones. A cheap plastic cube box. Sub: £399. I see gig x86 clients and in a space restrictive environment, 30 kids needing a computer each... Apple have got to do better with price in Education. Perception counts and mhz and price... Apple better hope they do something or maybe the move to 64 bit will save them as Intel bleats about the mhz myth.
"Me: "My hard drive is busted, and I did this, this and this just to make sure."
Dell: " OK, where do you want the replacement sent?"
The support issue is noted. Dell are very responsive to the needs of their customers. They must be doing something right to sell millions of computers a month in a saturated market. Perhaps Apple should take note. Price premium...just curious...where does the premium become unacceptable for people on these boards? For me, it happened a while back. To get those switchers...in the numbers Apple wants, I think they're going learn a couple of tricks from Dell.
"b8rtm8nn, I think pessimism is well warranted,"
Apple, to me, are 'poised' or 'hanging' on. I just don't see any forward momentum yet of any significance. Software and design, they've got the market cornered...yet... Maybe it's going to take a 'cool' olive branch to the x86 community iPod style gadget to convince x86 heads that Apple really is cool and x86..., 'Jeeze...why have I been putting up with this all my life...'
Yeesh, they need some device...SOMETHING to act as a tactical nuke to shake off that Redmond grip on the market. They're so small that they've got to keep hitting home runs! Maybe a few more cool devices like iPod may be the answer. Advertising blitz would help? They don't advertise nearly enough. I can't remember the last time I saw an iPod ad' on UK TV!
Anyhow. Feel better now I've got that off my chest...carry on...
OK, Lemon BB and Bartman... could you guys take this outside? To the "whine and moan" thread perhaps? This thread is supposed to address the issue of whether Apple's next powermac will contain an IBM-designed CPU. And actually, you may note, that apart from the distraction that you have managed to provide, after many pages the thread is still fresh and receiving notable contributions. I'm actually intrigued by what "I Have Questions" and the others have to say on the issue.
I think S. Jobs realizes that it will take a real winner of a chip and system to get any gain in market share. With the current economic state the only way people will buy anything is if it has cheap financing or big discounts. Look at the auto industry. Only BIG GM can seem to get any big sales gains, but look at the discounts. GM can do discounts longer than others because volume is their ONLY advantage, same for Dhell. S. Jobs will intro a G5 when he is sure that it will have a bigger market. Right now only the speed mongers will put orders in and then sales fall right back since everyone else is in the bargain mode. Apple is just reloading is arsenal right now and will let it all loose when the time is right.
Ford could have sold a lot of Edsel's in the 1962 but no one wanted a medium priced car in the 1958 recession. Timing is everything.
The new G5 will beat Intel to the 64 bit market. x86 is 1980?s technology and in the high end 3D Cad industry 1 ghz v.s. a 2 ghz Pentium is no great speed jump, I know this first hand. Intel keeps getting more RPM but the horsepower gains are disappearing. I suppose Intel is working on its ?ALL NEW P5? with pipelines so deep it will be able to run at 5 to 10 ghz. But 100 horsepower is 100 horsepower whether it take 6000 rpms or 15000 rpms to get the power it is still ONLY 100 horsepower.
Steve has learned his lesson over the last 2 years with the CUBE fiasco and the G4 Moto thing. Look at the new iMac flexible design, bigger display needed, no problem. With IBM and Moto both working for Apple business whom ever has the best will be the winner, but both will get business just the best will go in the Pro machines and runner up in consumer. A G4 with pumped up bus and other tweaks could really gain it real performance and would be great in consumer models. Great things come to those with patience.
Thanks, I Have Questions (even if your post is currently inscrutable ) and Ed M.. This thread is becoming interesting.
It looks like it'll be a wait for the IBM chip, but it'll be worth it when it arrives.
One thing that might account for the confused rumors: As Programmer intimates, IBM could be taking their embrace of onboard SIMD in two directions at once: As a full-on separate unit in the big, POWER4-derived unit, and as a Book E APU for a next-gen Sahara G3. IBM has said that the Sahara will be sprouting a SIMD unit.
Comments
<strong>
So Amorph you consider this as craps ? :
"Moreover, Book E defines ways that application-specific processing units (APUs) can be linked to a PowerPC processing core. The AltiVec PowerPC vector processing unit and instructions would be considered an APU under the Book E definition. Asked whether IBM will develop a PowerPC that includes an AltiVec coprocessor, Elliott Newcombe, PowerPC product marketing manager at IBM's Research Triangle Park facility, said IBM is considering adding an AltiVec APU to one of IBM's designs. "AltiVec could be an example of one of these coprocessors that could be plugged in [to a Book E-compliant PowerPC]," Newcombe said. "Nothing precludes IBM from doing that, but I cannot comment on whether a design is in progress. The market will decide whether we do that, and I can just say 'stay tuned.' "</strong><hr></blockquote>
Well, I've just reread the book E specs, and it ain't so. An APU cannot define a new set of registers, a SIMD engine can be defined as an APU, but it has to reuse an existing set of registers.
michael
D
[ 09-08-2002: Message edited by: DaveGee ]</p>
"Well, it sounds like this CPU is not in Apple's future -- the "over 160" vector instructions are not AltiVec (even though AltiVec has 162 instructions), and there are technical issues that would prevent AltiVec from ever marrying with Power4 or its successors. Furthermore, the guy came right out and said that they have pitched the desktop Power4 to Apple, but Apple was not interested."
So IBM pitched the Power4 to Apple even though it will never work with AltiVec aka Velocity Engine? None of what this guy said makes sense. Sorry, I think this whole rumor is just a bunch of BS.
There is absolutely no reason why the vector (SIMD) unit on this new chip isn't actually AltiVec (Velocity Engine) under their own name (VMX?) Again, Why on earth would IBM waste the time and effort to design a completely new SIMD unit that's only "compatible" with AltiVec when they are free to actually use AltiVec as it is?
I've talked with an ex PPC designer from Somerset and he's been delving into this more than I have and going over the docs and comparing them to what the word on the net is. He says that he'd be VERY surprised if this vector unit was NOT AltiVec. In fact he believes that everything points to it bieng exactly that. Still, it raises a bunch of questions. I listed them earlier, so you might wan't to revisit them and ponder them for a while.
--
Ed M.
Powerdoc's post on the top of this page is moving in the right direction, and he is, uh... talking to the right person. Also, the answer to Powerdoc's final question to me is "yes." That's all I can say without someone getting a pink slip.
Well, I have a few trusted sources as well. I wonder if the guys you know worked with the person I speak with. The guy I speak with worked in the PPC design department over at the Somerset facility. He's been keeping in contact and reading up on the info being discussed on the web as well as whatever he has access to from Moto and IBM. I'm not sure what he still has access to and I'm not sure if any of his NDAs still apply. In all, he stated that He'd be VERY surprised if the vector unit wasn't AltiVec with IBM's one name tacked on.
[[[Time frame = 6-10 months to "completion". This may be what is going to be discussed in October. Also, I have no idea what "completion" means in this context.]]]
hmmmm. If I'm not mistaken I think IBM is prohibited from making any sort of "vaporware" announcements. Something about it being in one of their antitrust settlements from long ago. That said, I'd have to assume that this chip is already done and will be announced in October. Can anyone verify the above as an IBM settlement stipulation?
--
Ed M.
There's alot of irony in that statement. If there isn't that much to do, then why haven't they done it already?
More annoying is that clearly you see that 'Mot doesn't really need to do much' and you're not the only one. That's what's been so annoying about the G4. If an extra FP unit had been added then it would have held its own more. The Athlon destroys the G4 in Lightwave...and the Pentium 4's optimisations in Lightwave destroy the Athlon. FP, my area of interest I guess as a Lightwave 7.5 owner.
"Take a 7455, add an FP unit, a memory controller and RIO support, keep it MP friendly, and die-shrink it to .09. There's your screaming 32-bit chip. The biggest advances would be the additional FP and (especially) the RapidIO, since the current 7455 is hobbled by low bandwidth."
A part of me says that is where we should be now...if Apple wants to 'innovate its way out of this recession'. Innvotion should be more than skin deep.
Still, it that's what they put under the G5 nom da plum in 6 months time then I'll take it on the chin. I've waited years for a decent 'power'Mac. I won't leave it longet than this next year.
I appreciate Gamblor's 'soothing'. When I first heard the 'Nasa' rumour, I almost blew a gasket. I thought, Motorola won't deliver, Apple are now turning down a great Chip from IBM...what now?
In calmer morning light I guess Apple will probably create a healthy competition from both IBM/Moto' and leave their options open ala Nvidia and Ati. Having only one supplier for graphics cards for years was very damaging...and there's no doubt being roped into Moto' only for cpus has done Apple immense damage. The sales of 'power'Macs speak for themselves. The comment on a 'core' will modular elements and Rio adoption being virtually mandatory between IBM/Moto then I guess there's cause to be optimistic. Apple can 'choose' what's right for them in consumer and workstations.
Choice of supplier is the watchword. We can see what competition has done for graphics cards and cpus on the x86 side. There's alot to be said for competition.
Part of me, inside feels like JD's post. It scary...despite all the great things Apple are doing...all the publicity they're getting for their machines, stores and os, they're still barely able to stand still! Dell are giving them the beating of a lifetime in education. If you extrapolate that to publishing then Apple are in serious trouble.
They MUST address the performance issue in workstations.
They MUST address the price issue in education.
The only way you can make things cheaper is to sell more of them. 10% is the holy grail for Apple.
They've got great OS, great software, beyond the box thinking, adopted open standards, have more games and developers than they've ever had...yet...why do I get the feeling that they're still sinking?
Lemon Bon Bon
Your rabid pessimism is unwarranted. Apple's global market share has grown and their software has done a 180. Although the hardware is lagging and their educational sales have dropped, they are doing well and are poised to succeed.
Apple is a thriving company with a niche market, they will not likely overturn MS, and so they will never overtake Dell. They could succeed in making MS operating systems non-ubiquitous, however. And for every Powermac sale they are losing due to hardware, they are probably gaining another plus some due to the OS and iApps. I'm sure the Powermac drought is temporary, albeit lengthy, but you will survive
Dell support is well above Apple for people like me:
Me: "My hard drive is busted, and I did this, this and this just to make sure."
Dell: " OK, where do you want the replacement sent?"
Next morning there is a replacement hard drive at my office and a prepaid return sticker ready to send the broken drive back. All free of additional charges. That is why we buy Dell over any PC vendor or Apple if Apple doesn't provide a reason to use them over a PC.
(Yes, the display was still covered by its one-year--BTW, isn't 3-year standard industry-wide?--warranty!)
I can understand very well now why Apple's .edu share is less than half than Dell's... <img src="graemlins/oyvey.gif" border="0" alt="[No]" />
ZoSo
But I can only forgive lack in raw performance up to a certain point.
Apple better have caught up with Intel/AMD in a year or two, or my next computer ain't gonna be a Mac.
If you search my post history (which you probably won't bother to do), you will see that I have belittled JYD, Lemon Bon Bon etc for being so anti-Apple due to the lack in raw performance in the Power Mac range.
Mac OS X makes up for some of that loss, but it's getting to the point where it doesn't make up for the G4's inability to compete with Intel/AMD. The G4 is great for consumer/portable products (cheap, low temp), but not for the Power Mac.
The Power Mac demo I saw at the Apple Road Show a few weeks ago didn't floor me like the latest PCs I've seen.
Apple, I'm a customer. Treat me better. Find a way to give me more performance, or I'm leaving.
Barto
Reading this through shows that Rivina and guTS are 2 different prototypes. Appears that guTS is a precursor to Rivina and Rivina itself is not a finished processor but a precursor to something else. I think the prototype (?) before Rivina is the Pulsar. It has larger L1 caches (128KB vs 64KB) but more L1 cache latency. Rivina was designed to test an experimental CPU floorplan to improve chip frequency beyond the improvements that extending the pipeline can provide. You have to admit, on a .22 micron process (220nm) 1.15GHz is impressive albiet at 120+ watts. With die shrinks to 180nm, 130nm (should yield 1.8-2.0GHz processors), and eventually 90nm, this new family of processors should scale well.
And here is an interesting quote from the article:
If one wants to predict what a next generation of microprocessors might look like, examining the system in which they are used usually gives a good clue. Present-day designs have migrated L2 caches and multiple cores onto the chip. It is thus interesting to examine integrating DRAM on the die with the processor. IBM has demonstrated the capability of combining high-performance DRAM, the next off-chip component, with high-performance logic on the same die [38] (Figure 34). When logic transistor performance no longer suffers, the power and density advantages of DRAM make it a good candidate for an on-chip, higher-level cache. For a large enough cache size, a DRAM can even outperform an SRAM in latency because the delay to cover the distance over the array to the processor dominates the time needed to access the subarray. Promising as this technology may be, it is by and large complementary to the improvements that are expected in the core of the processor.
<strong>Why couldn't IBM have a SIMD unit that used existing registers and still decode Altivec instructions for processing in its own SIMD unit?</strong><hr></blockquote>
Code that uses the AltiVec registers doesn't expect the standard registers to get clobbered when AltiVec instructions are used, yet this is what would happen if the AltiVec instructions were mapped onto the standard ones.
Adding a SIMD instructions that use the standard registers is much simpler than adding a whole external vector unit, and it would generally be transparent to all software that didn't use the SIMD instructions -- including the OS. It doesn't increase the amount of processor state that needs to be preserved across context switches and function calls. It makes it easier to "do a little bit of SIMD" since the data is already in your standard registers. This approach makes more sense in the embedded space (hence Motorola uses it in the 85xx series), whereas a desktop or server processor can afford the full vector unit which is a relatively small amount of transistors compared to the huge caches and other parts of a high performance chip.
IBM goes ahead and manufactures the chip, ostensibly for Linux boxes and everyone runs Mac OS(including X) inside Linux.
<a href="http://maccentral.macworld.com/news/0209/07.mol.php"; target="_blank">Mac-on-Linux supports Mac OS X (including Jaguar)</a>
Your rabid pessimism is unwarranted."
Hmmmm. (Think Marge, Simpsons.) Thanks Bartman, you almost cheered me up there...
"Apple's global market share has grown and their software has done a 180. Although the hardware is lagging and their educational sales have dropped, they are doing well and are poised to succeed."
There are more Macs out there than ever before. Compared to 1996...maybe they are selling more Macs due to the smaller slice of a larger pie sydrome. But Apple with 2.5 World Wide Market share have had a shrinking share of the the growing pie. They're slipping faster than they're climbing relatively. It'll keep them in business for the forseeable future. That and the 4 billion in cash.
"Apple is a thriving company with a niche market, they will not likely overturn MS, and so they will never overtake Dell. They could succeed in making MS operating systems non-ubiquitous, however."
Maybe the 'switch' campaign will help get the 'Mac OS' message across. Apple doen't seem to show their OS, their products off in the Switch ads. Perhaps they should along with the emotive PCs make life hard tac. Getting to that 10% market share looks to be a long haul.
"And for every Powermac sale they are losing due to hardware, they are probably gaining another plus some due to the OS and iApps. I'm sure the Powermac drought is temporary,"
Perhaps. But the 'power'Mac drought has been going on for so long it's not funny. What's worse? Waiting for the G5 or sexual frustration?
"albeit lengthy, but you will survive"
I'm gagging for it. A G5 that is.
"PS. I don't think it's a price issue in education"
Nope. I simpley don't agree with this. Price for people in education (and I work in a school myself where they're upgrade to Pentium 4s...) is a significant issue. It's the different between an extra few PCs...people go PCs. The mhz issue has hurt them in my school. Price also. I've tried 'bleating' the overall cost of ownership mantra. People just see the 'mainstream' view. Everybody has Windows. Mhz matters and so does the bottom line.
The 'useage' semantics of 'ease of use', iphoto automatically taking photos from yer camera, the iapps, appleworks etc. It's like you're talking a different language to x86 folks. I really think Apple have got to come out with a thin client, something below an eMac (which is simply not cheap enough in the UK I'm afraid...) Headless with onboard graphics etc. Barebones. A cheap plastic cube box. Sub: £399. I see gig x86 clients and in a space restrictive environment, 30 kids needing a computer each... Apple have got to do better with price in Education. Perception counts and mhz and price... Apple better hope they do something or maybe the move to 64 bit will save them as Intel bleats about the mhz myth.
"Me: "My hard drive is busted, and I did this, this and this just to make sure."
Dell: " OK, where do you want the replacement sent?"
The support issue is noted. Dell are very responsive to the needs of their customers. They must be doing something right to sell millions of computers a month in a saturated market. Perhaps Apple should take note. Price premium...just curious...where does the premium become unacceptable for people on these boards? For me, it happened a while back. To get those switchers...in the numbers Apple wants, I think they're going learn a couple of tricks from Dell.
"b8rtm8nn, I think pessimism is well warranted,"
Apple, to me, are 'poised' or 'hanging' on. I just don't see any forward momentum yet of any significance. Software and design, they've got the market cornered...yet... Maybe it's going to take a 'cool' olive branch to the x86 community iPod style gadget to convince x86 heads that Apple really is cool and x86..., 'Jeeze...why have I been putting up with this all my life...'
Yeesh, they need some device...SOMETHING to act as a tactical nuke to shake off that Redmond grip on the market. They're so small that they've got to keep hitting home runs! Maybe a few more cool devices like iPod may be the answer. Advertising blitz would help? They don't advertise nearly enough. I can't remember the last time I saw an iPod ad' on UK TV!
Anyhow. Feel better now I've got that off my chest...carry on...
Lemon Bon Bon
OK, Lemon BB and Bartman... could you guys take this outside? To the "whine and moan" thread perhaps? This thread is supposed to address the issue of whether Apple's next powermac will contain an IBM-designed CPU. And actually, you may note, that apart from the distraction that you have managed to provide, after many pages the thread is still fresh and receiving notable contributions. I'm actually intrigued by what "I Have Questions" and the others have to say on the issue.
</signal-to-noise-ratio-police>
Ford could have sold a lot of Edsel's in the 1962 but no one wanted a medium priced car in the 1958 recession. Timing is everything.
The new G5 will beat Intel to the 64 bit market. x86 is 1980?s technology and in the high end 3D Cad industry 1 ghz v.s. a 2 ghz Pentium is no great speed jump, I know this first hand. Intel keeps getting more RPM but the horsepower gains are disappearing. I suppose Intel is working on its ?ALL NEW P5? with pipelines so deep it will be able to run at 5 to 10 ghz. But 100 horsepower is 100 horsepower whether it take 6000 rpms or 15000 rpms to get the power it is still ONLY 100 horsepower.
Steve has learned his lesson over the last 2 years with the CUBE fiasco and the G4 Moto thing. Look at the new iMac flexible design, bigger display needed, no problem. With IBM and Moto both working for Apple business whom ever has the best will be the winner, but both will get business just the best will go in the Pro machines and runner up in consumer. A G4 with pumped up bus and other tweaks could really gain it real performance and would be great in consumer models. Great things come to those with patience.
It looks like it'll be a wait for the IBM chip, but it'll be worth it when it arrives.
One thing that might account for the confused rumors: As Programmer intimates, IBM could be taking their embrace of onboard SIMD in two directions at once: As a full-on separate unit in the big, POWER4-derived unit, and as a Book E APU for a next-gen Sahara G3. IBM has said that the Sahara will be sprouting a SIMD unit.