I understand all that, but there still has to be something for the consumer to go, WOW!
Well then we need to find something other than clock rate. Even at sub-2 GHz these things will perform really well due to the on-chip memory controller. If they go dual as well they should be outstanding. Long battery life or lower weight (due to less battery) are also big selling points.
Well then we need to find something other than clock rate. Even at sub-2 GHz these things will perform really well due to the on-chip memory controller. If they go dual as well they should be outstanding. Long battery life or lower weight (due to less battery) are also big selling points.
I have mentioned it before, Apple Marketing on it's computers leaves something to be desired. Those nuances only go so far.
FYI: I just received an invite to an IBM iSeries presentation on May 4th. They are going to announce POWER5 based iSeries on that date. No details yet on when they will ship. May 2 - 6 is the COMMON conference which is a big twice a year trade show for iSeries folks (their MacWorld).
From the email:
Next week, May 4, the iSERIES will have a MAJOR ANNOUNCEMENT regarding POWER5 and the iSeries supporting every major operating environment (except mainframe)-in a single footprint: LINUX, AIX, OS/400 and WINDOWS.
Next week, May 4, the iSERIES will have a MAJOR ANNOUNCEMENT regarding POWER5 and the iSeries supporting every major operating environment (except mainframe)-in a single footprint: LINUX, AIX, OS/400 and WINDOWS.
Woah? Are they going to support Windows on iSeries? Support as in, "We'll have Windows running on iSeries"?
That means either that there is some emulation layer that Windows will run in or that Windows is ported to PPC. Both scenarios would be quite astonishing.
And.. I do consider Mac OS X a major operating system and it would be fairly easy implementing that on iSeries, but heck.. who cares?
It is interesting your focus on the bus efficiency, it reminds me of American management of production facilities.
What on Earth are you talking about?
Someone made an incorrect statement about the inefficiency of the G4 architecture and I corrected it. That's all.
That doesn't mean that I wouldn't prefer to see it replaced with a bus that has lower efficiency and higher real throughput. In fact, if you've read any of my other posts on the subject, you'd be aware that I'm all for moving the memory controller on-die and going to a packet-based bus like HyperTransport or RapidIO, which would be less efficient than MaxBus, but much higher throughput. So, in context, your criticism is absurd.
All I did was clarify that the problem with the G4 is bandwidth (throughput) rather than inefficiency. I also alluded to the paucity of FP execution units, which is the other stumbling block in the design.
More on topic, it's nice to see that IBM's business model still hasn't changed since the 1960s. $10,000 to activate a processor, indeed.
One of those documents mentions CMP (Chip MultiProcessing). Is this Motorolas answer to to the vearious multi threaded processors coming onto the market or is it just their reference to dual core chips?
Quote:
Originally posted by PB
OK, it seems that Motorola actually reads these boards and delivers information on command :
One of those documents mentions CMP (Chip MultiProcessing). Is this Motorolas answer to to the vearious multi threaded processors coming onto the market or is it just their reference to dual core chips?
I have only done a little googling on CMP, and haven't as of yet come up with anything that defines it.
That sounds like a name for dual-core, which is in itself not a bad answer to SMT.
One pleasant consequence of the Reign of MaxBus is that there's a huge legacy of existing code that's designed to squeeze every last drop out of a G4, far more pervasive threading of applications than on other platforms, and a large pool of talent skilled at the necessary arts. A dual-core G4 with an on-die memory controller and a fat pipe to RAM would quickly find itself being used at close to full potential, as the Evil Bandwidth Cap is removed. Meanwhile, the point of SMT is to use a given architecture at closer to full potential.
I think I've seen Motorola discuss CMP before and it meant multiple cores. The G4's limited OoOP capabilities mean it wouldn't be particuarly effective at SMT, so I doubt that would be a worthwhile approach.
i forgot the version (3.5?) of NT supported PPC way back when...
wouldnt be surprised if they could get NT/XP running on latest PPC again.
Also, with all the work MS is doing with IBM and the XBOX2 processor... It seems kinda "easy" for Windoze to support G5/Power5
What do you want the bloatwear for? I personally don't want it running nativly on my macs. That is why I own a dust collecting door stop. So when I do have to use Windows or RedHat, it is not messing with the good stuff.
i forgot the version (3.5?) of NT supported PPC way back when...
wouldnt be surprised if they could get NT/XP running on latest PPC again.
Also, with all the work MS is doing with IBM and the XBOX2 processor... It seems kinda "easy" for Windoze to support G5/Power5
Of course they could, but I doubt they would. The work they are doing for the XBox2 processor would help such a tiny amount compared to the total effort to bring the entire OS to PowerPC that it is basically irrelevent. Wrong development group within MS anyhow.
Of course they could, but I doubt they would. The work they are doing for the XBox2 processor would help such a tiny amount compared to the total effort to bring the entire OS to PowerPC that it is basically irrelevent. Wrong development group within MS anyhow.
Yeah, and this time they'll let NT and its applications use more than 16 general purpose registers at a time.
Has anyone considered the possibility that the 970FX is destined for Apples low end machines and that the 97X, Power 5 dirivative is destied for the PowerMacs?
I mean lets be honest the 970 is not that magnificent of a processor. Sure it does well in some situations, but it really is not the bleeding edge that Steve would want out of a PowerMac. Futher the 970FX was not much of an improvement over the original which is probally the strongest argument for another processor being readied for the PowerMac.
Just a thought, that comes from an over active imagination. The reality is that the 970FX would allow Apple to produce a nice machine to compete against the i86 consumer world. Having a significantly faster chip ready for the PowerMac would allow Apple to move decisively to the front of the performance race.
I know that some have disagreed with me, but the day that the 970FX came out I was rather shocked that Apple / IBM did not do more to take advantage of the space available to enhance performance. The conclusion to be reached is that this is not a chip earmarked for high performance computing rather one targetted at the consumer.
Has anyone considered the possibility that the 970FX is destined for Apples low end machines and that the 97X, Power 5 dirivative is destied for the PowerMacs? . . .
It sounds reasonable to me. It's a small chip so the price should be lower than the Power5 derivative, once yields are up. Fewer transistors, so the power will be lower at a given clock rate.
Wizard, I think there's something in what you're saying.
A few things.
The 970 is a good enough rescue act for the PowerMac range. No doubts there. But they can hardly moan about disappointing PowerMac sales when they don't build upon their early success. No updates and it's been nearly 11 months since they were announced. Stagnation eroded g4 power Mac sales. If they're having problems at least drop the 1.6 gig model, have the dual 1.8 as the low end model, stick a dual 2 gig model in the middle and a loaded dual 2 gigger as the top model. But no...11 months and Apple doesn't do anything like that...bar the update of the 1.8 to dual.
But, to be honest, the 970 doesn't really come across to me as a from the ground up Wintel Intel/AMD killer.
How's the 970 stand up to the latest AMD64s in benchmarks?
If it wasn't for the compelling bandwidth, I don't think the 970 would do as well as it does. It's competitive. Or was at June 03.
But...as always...the PC World has moved on while Apple have been static. Don't they ever learn?
And...considering that the AMD64s can be had for half the price of the dual 2 gig G5 PowerMac then surely it would be prudent to address the consumer range with the G5 as soon as is possible.
Personally, considering the G4 debacle(insert word of choice...) I would have thought, hoped that Apple would have co-developed a decent PPC roadmap with the new A.I Alliance. As a catch-up strategy, I would have hoped Apple would have the '975' out for the PowerMacs this Summer and quickly bung the 970fx in everything from eMacs to iMacs to Powerbooks. Do that? And Apple would have a decent chance of breaking the 1 million units per quarter barrier.
The fact that it's May. No iMac update. No G5s in Powerbooks (I was amazed at how the recent G4 Powerbooks were lauded and raved over on these boards...it may be stunningly slim but it's getting whomped in 3d benchmarks and photoshop by latest Wintel laptops and their screens are brighter... But I'm not sure a 1.6 G5 fx would offer compelling performance edge over the 1.5 gig G4. Worth waiting the extra time then and get a 2 gig G5fx or G5 II in there?) eMac updated but no G5.
It's not a disaster. But it's disappointing in my view. I thought Apple would have aggressively followed on from the G5 launch.
Hopefully, 2nd half 2004 will have the entire line G5 fx with Professional line G5 II.
Hey, at least they got the eMac closer to the magic 495 mark here in the UK. An eMac for £545 inc VAT. No bad. So much for Apple being 'unable' to enter the low end or be unable to make money off the low end. It's getting some nice reviews as a machine that dispels the 'expensive'/'premium' Apple tag.
I think Apple have really got to sort out the desktop line above the eMac. (Which, itself must be going LCD soon? It's still the ugly duckling of the Apple line.)
iMac. G5. Alu' design. Mini-tower format. LCD angle lamp design to go with entire desktop line. Maybe a dockable LCD. And fer gadsakes. Get a hair cut. Sorry, a price cut. Yeesh. It's laughable how laptop sales are accused of taking away iMac sales. Nothing to do with being massively overpriced and obselete? And Apple's Powerbook sales weren't that good either. If laptops are taking off, why aren't Apple getting 250K worth of Powerbook sales and 350K of iBook sales? Because you can get a cheap crappy Wintel brick for £350? Twice less the entry iBook?
I just hope Steve doesn't have to climb down on the 3 gig promise like he did the Pepsi 100 million promise. Ouch.
Still, I'm still shouting for Cupertino boys..., 'Go Apple..!'
(Is it me or has 2004 been disappointing so far? Not in software. Great. But Mac hardware? Lame. G4 Powerbooks still. Consumer desktops languishing on 1 gig ish G4 processors...STILL. F&&**(*(inBleep. It's nearly 2005!!!!) (Ey, I do like a good moan...)
Well the eMac was not going to get a G5, yet. The iMac/or what ever replaces it, is getting one, but Apple has to wait for IBM. It should have happened in March, but the IBM issues put a stop to that. The PowerMac has to have an upgrade at the same time the iMac gets the G5.
September was Steve's goal for releasing the 3.0Ghz PM. If IBM has the Power5 dirivative ready togo then we look to June. IBM is going to start announcing their Power5 servers today I believe. If I remember correctly they will start shipping in early June.
I believe that Apple will skip a generation of PowerMacs and go with the WOW factor. If one thing is on Steve's mind, it is that Apple will release a 3.0Ghz PM before October 1st.
Comments
Originally posted by oldmacfan
I understand all that, but there still has to be something for the consumer to go, WOW!
Well then we need to find something other than clock rate. Even at sub-2 GHz these things will perform really well due to the on-chip memory controller. If they go dual as well they should be outstanding. Long battery life or lower weight (due to less battery) are also big selling points.
Originally posted by Programmer
Well then we need to find something other than clock rate. Even at sub-2 GHz these things will perform really well due to the on-chip memory controller. If they go dual as well they should be outstanding. Long battery life or lower weight (due to less battery) are also big selling points.
I have mentioned it before, Apple Marketing on it's computers leaves something to be desired. Those nuances only go so far.
From the email:
Next week, May 4, the iSERIES will have a MAJOR ANNOUNCEMENT regarding POWER5 and the iSeries supporting every major operating environment (except mainframe)-in a single footprint: LINUX, AIX, OS/400 and WINDOWS.
Originally posted by CodeWarrior
From the email:
Next week, May 4, the iSERIES will have a MAJOR ANNOUNCEMENT regarding POWER5 and the iSeries supporting every major operating environment (except mainframe)-in a single footprint: LINUX, AIX, OS/400 and WINDOWS.
Woah? Are they going to support Windows on iSeries? Support as in, "We'll have Windows running on iSeries"?
That means either that there is some emulation layer that Windows will run in or that Windows is ported to PPC. Both scenarios would be quite astonishing.
And.. I do consider Mac OS X a major operating system and it would be fairly easy implementing that on iSeries, but heck.. who cares?
Originally posted by wizard69
It is interesting your focus on the bus efficiency, it reminds me of American management of production facilities.
What on Earth are you talking about?
Someone made an incorrect statement about the inefficiency of the G4 architecture and I corrected it. That's all.
That doesn't mean that I wouldn't prefer to see it replaced with a bus that has lower efficiency and higher real throughput. In fact, if you've read any of my other posts on the subject, you'd be aware that I'm all for moving the memory controller on-die and going to a packet-based bus like HyperTransport or RapidIO, which would be less efficient than MaxBus, but much higher throughput. So, in context, your criticism is absurd.
All I did was clarify that the problem with the G4 is bandwidth (throughput) rather than inefficiency. I also alluded to the paucity of FP execution units, which is the other stumbling block in the design.
More on topic, it's nice to see that IBM's business model still hasn't changed since the 1960s. $10,000 to activate a processor, indeed.
Originally posted by PB
OK, it seems that Motorola actually reads these boards and delivers information on command
Freescale Semiconductor reveals PowerPC® core roadmap and scalable system-on-chip platforms
High-Performance PowerPC Processors
Roadmap
Originally posted by wizard69
One of those documents mentions CMP (Chip MultiProcessing). Is this Motorolas answer to to the vearious multi threaded processors coming onto the market or is it just their reference to dual core chips?
I have only done a little googling on CMP, and haven't as of yet come up with anything that defines it.
One pleasant consequence of the Reign of MaxBus is that there's a huge legacy of existing code that's designed to squeeze every last drop out of a G4, far more pervasive threading of applications than on other platforms, and a large pool of talent skilled at the necessary arts. A dual-core G4 with an on-die memory controller and a fat pipe to RAM would quickly find itself being used at close to full potential, as the Evil Bandwidth Cap is removed. Meanwhile, the point of SMT is to use a given architecture at closer to full potential.
It's that efficiency thing again.
Originally posted by Amorph
That sounds like a name for dual-core, which is in itself not a bad answer to SMT.
Might also be their name for SMT.
Originally posted by Tomb of the Unknown
Might also be their name for SMT.
I think I've seen Motorola discuss CMP before and it meant multiple cores. The G4's limited OoOP capabilities mean it wouldn't be particuarly effective at SMT, so I doubt that would be a worthwhile approach.
wouldnt be surprised if they could get NT/XP running on latest PPC again.
Also, with all the work MS is doing with IBM and the XBOX2 processor... It seems kinda "easy" for Windoze to support G5/Power5
Originally posted by ZO
i forgot the version (3.5?) of NT supported PPC way back when...
wouldnt be surprised if they could get NT/XP running on latest PPC again.
Also, with all the work MS is doing with IBM and the XBOX2 processor... It seems kinda "easy" for Windoze to support G5/Power5
What do you want the bloatwear for? I personally don't want it running nativly on my macs. That is why I own a dust collecting door stop. So when I do have to use Windows or RedHat, it is not messing with the good stuff.
Originally posted by ZO
i forgot the version (3.5?) of NT supported PPC way back when...
wouldnt be surprised if they could get NT/XP running on latest PPC again.
Also, with all the work MS is doing with IBM and the XBOX2 processor... It seems kinda "easy" for Windoze to support G5/Power5
Of course they could, but I doubt they would. The work they are doing for the XBox2 processor would help such a tiny amount compared to the total effort to bring the entire OS to PowerPC that it is basically irrelevent. Wrong development group within MS anyhow.
Originally posted by Programmer
Of course they could, but I doubt they would. The work they are doing for the XBox2 processor would help such a tiny amount compared to the total effort to bring the entire OS to PowerPC that it is basically irrelevent. Wrong development group within MS anyhow.
Yeah, and this time they'll let NT and its applications use more than 16 general purpose registers at a time.
Originally posted by Amorph
Yeah, and this time they'll let NT and its applications use more than 16 general purpose registers at a time.
I mean lets be honest the 970 is not that magnificent of a processor. Sure it does well in some situations, but it really is not the bleeding edge that Steve would want out of a PowerMac. Futher the 970FX was not much of an improvement over the original which is probally the strongest argument for another processor being readied for the PowerMac.
Just a thought, that comes from an over active imagination. The reality is that the 970FX would allow Apple to produce a nice machine to compete against the i86 consumer world. Having a significantly faster chip ready for the PowerMac would allow Apple to move decisively to the front of the performance race.
I know that some have disagreed with me, but the day that the 970FX came out I was rather shocked that Apple / IBM did not do more to take advantage of the space available to enhance performance. The conclusion to be reached is that this is not a chip earmarked for high performance computing rather one targetted at the consumer.
Dave
Originally posted by wizard69
Has anyone considered the possibility that the 970FX is destined for Apples low end machines and that the 97X, Power 5 dirivative is destied for the PowerMacs? . . .
It sounds reasonable to me. It's a small chip so the price should be lower than the Power5 derivative, once yields are up. Fewer transistors, so the power will be lower at a given clock rate.
A few things.
The 970 is a good enough rescue act for the PowerMac range. No doubts there. But they can hardly moan about disappointing PowerMac sales when they don't build upon their early success. No updates and it's been nearly 11 months since they were announced. Stagnation eroded g4 power Mac sales. If they're having problems at least drop the 1.6 gig model, have the dual 1.8 as the low end model, stick a dual 2 gig model in the middle and a loaded dual 2 gigger as the top model. But no...11 months and Apple doesn't do anything like that...bar the update of the 1.8 to dual.
But, to be honest, the 970 doesn't really come across to me as a from the ground up Wintel Intel/AMD killer.
How's the 970 stand up to the latest AMD64s in benchmarks?
If it wasn't for the compelling bandwidth, I don't think the 970 would do as well as it does. It's competitive. Or was at June 03.
But...as always...the PC World has moved on while Apple have been static. Don't they ever learn?
And...considering that the AMD64s can be had for half the price of the dual 2 gig G5 PowerMac then surely it would be prudent to address the consumer range with the G5 as soon as is possible.
Personally, considering the G4 debacle(insert word of choice...) I would have thought, hoped that Apple would have co-developed a decent PPC roadmap with the new A.I Alliance. As a catch-up strategy, I would have hoped Apple would have the '975' out for the PowerMacs this Summer and quickly bung the 970fx in everything from eMacs to iMacs to Powerbooks. Do that? And Apple would have a decent chance of breaking the 1 million units per quarter barrier.
The fact that it's May. No iMac update. No G5s in Powerbooks (I was amazed at how the recent G4 Powerbooks were lauded and raved over on these boards...it may be stunningly slim but it's getting whomped in 3d benchmarks and photoshop by latest Wintel laptops and their screens are brighter... But I'm not sure a 1.6 G5 fx would offer compelling performance edge over the 1.5 gig G4. Worth waiting the extra time then and get a 2 gig G5fx or G5 II in there?) eMac updated but no G5.
It's not a disaster. But it's disappointing in my view. I thought Apple would have aggressively followed on from the G5 launch.
Hopefully, 2nd half 2004 will have the entire line G5 fx with Professional line G5 II.
Hey, at least they got the eMac closer to the magic 495 mark here in the UK. An eMac for £545 inc VAT. No bad. So much for Apple being 'unable' to enter the low end or be unable to make money off the low end. It's getting some nice reviews as a machine that dispels the 'expensive'/'premium' Apple tag.
I think Apple have really got to sort out the desktop line above the eMac. (Which, itself must be going LCD soon? It's still the ugly duckling of the Apple line.)
iMac. G5. Alu' design. Mini-tower format. LCD angle lamp design to go with entire desktop line. Maybe a dockable LCD. And fer gadsakes. Get a hair cut. Sorry, a price cut. Yeesh. It's laughable how laptop sales are accused of taking away iMac sales. Nothing to do with being massively overpriced and obselete? And Apple's Powerbook sales weren't that good either. If laptops are taking off, why aren't Apple getting 250K worth of Powerbook sales and 350K of iBook sales? Because you can get a cheap crappy Wintel brick for £350? Twice less the entry iBook?
I just hope Steve doesn't have to climb down on the 3 gig promise like he did the Pepsi 100 million promise. Ouch.
Still, I'm still shouting for Cupertino boys..., 'Go Apple..!'
(Is it me or has 2004 been disappointing so far? Not in software. Great. But Mac hardware? Lame. G4 Powerbooks still. Consumer desktops languishing on 1 gig ish G4 processors...STILL. F&&**(*(inBleep. It's nearly 2005!!!!)
Lemon Bon Bon
September was Steve's goal for releasing the 3.0Ghz PM. If IBM has the Power5 dirivative ready togo then we look to June. IBM is going to start announcing their Power5 servers today I believe. If I remember correctly they will start shipping in early June.
I believe that Apple will skip a generation of PowerMacs and go with the WOW factor. If one thing is on Steve's mind, it is that Apple will release a 3.0Ghz PM before October 1st.