I think we all agree that an x86 port ain't happening nor should it. But while we're thinking processor realignment... how much work would it take to adapt OSX to Cell?
I have had the same feelings when the G4s stagnated at 500 MHz and Intel and AMD where scaling into the GHz as if there was no tomorrow. Speedwise, the stagnation threw Apple's desktop line back almost 2 or 3 processor iterations.
And the knowledge made by developing the M-series processors will finally make it into Intel's desktop line.
I still wonder why Apple is unable to put enough pressure onto their suppliers to improve certain aspects of their CPU designs (faster FSB in Freescales case, dynamic power and speed adjustments Ã* la Pentium M in both cases).
I'm not going to question the length of time it has taken for Apple to get decent processor speeds out of their suppliers...I wonder if this will feature in a chapter of Steve Jobs autobiography
You'd have to say that this MHz gap has improved over the past 2 years...both IBM and Freescale finally seem to have realised that there is a wider market for RISC chips...and quicker ones at that. Whilst the 7447B is ok, I have a sneaking suspicion the next PowerBook will see a greater performance jump.
The 7448 (see http://www.freescale.com/files/32bit.../MPC7450UM.pdf) is a step along the way to Dothan-like performance...increased Dynamic Frequency Switching (DFS), increased cache, improvements to Altivec, and perhaps a 20% improvement in clockspeed in the move to a 90nm process.
When we see the chip, hopefully at WWDC or just after, it will face a Pentium-M/Sonoma combo that is good but only slightly better than the Pentium-M/Centrino ( with a marginal clockspeed improvement and a quad pumped bus). The gap will have closed considerably in that time between the supposed inadequacy of the PowerBook and the pretender to the notebook throne. When the 8641D arrives at Macworld 2006, everyone will be happy. Optimistic? People, things are on the up
No more limited FSB and the G5 fanboys can go back to their mothers for coke and hugs
agreed. on this note, what's the group consensus on how likely 7448 is for WWDC?
yesterday, i was pretty resolved to get one of the new 15" models. only downside is the lingering 167mhz bus.
but now i'm thinking i might do better with a stopgap (cheap) ibook til WWDC, then get a 7448 PB, and resell the ibook.
thoughts?
hey, go for it, you might end up holding on to your iBook a little longer than you expect (maybe enough to upgrade when the 7448/G5 powerbook Rev B comes out)
agreed. on this note, what's the group consensus on how likely 7448 is for WWDC?
I would not bet on that. And by the way, let me remind the official Freescale roadmap from the other thread that died in Temporary Insanity:
(1) 7448, single e600 core, >1.5 GHz, 1 MB L2 cache with ECC, 200 MHz system bus, less than 10 W power consumption at 1.4 GHz.
(2) 8641D, dual e600 core, >1.5 GHz, 1 MB L2 cache with ECC (per processor core), integrated MPX bus up to 667 MHz (MUCH MUCH faster than the one we have for years in the G4 as a major bottleneck) with dual 64-bit DDR and DDR-II support, PCIe interface.
Availability is scheduled for 2H-05 and 1H-06 respectively, see the SNDF presentation (page 41).
What does this mean. With some good luck, we will have a 7448-based Powerbook by this year's summer-autumn and a 8641D-based (dual-core) Powerbook by the beginning of next year, say in one year from now. Now, the 7448 and 8641D are based on the e600 core which is 32-bit. But as Freescale says, there will be a 64-bit successor to the e600, with full 32-bit support, the e700. If Apple takes the Freescale road (and with all current evidence they have no other choice), I guess this e700 would perhaps be the G5-mobile for the Powerbook. Unfortunately, there are no available details on sampling and production for the e700 (personally, I would not expect it before 2007).
For the time being though, it seems that the Powerbook will follow a smooth evolution to a very capable 32-bit dual-core portable machine, without the system bottlenecks that plagued the platform for years. It remains to see if IBM comes with a response to Freescale. It would be very interesting indeed if they did. But there is zero evidence at this moment that they would do that before 2006.
originally was going to wait till wwdc...now decided i'll be happy with the last g4 update - it won't have any rev a bugs and is more than fast enough for what i need to do.
will probably order in april/may, bit of a cash flow problem at the moment.
It remains to see if IBM comes with a response to Freescale. It would be very interesting indeed if they did. But there is zero evidence at this moment that they would do that before 2006.
you have not seen anything regarding a Power5-based mobile PPC that could make it into a PowerBookG5 by summer/fall 2005? just curious, not even a hint anywhere?
you have not seen anything regarding a Power5-based mobile PPC that could make it into a PowerBookG5 by summer/fall 2005? just curious, not even a hint anywhere?
Sorry, no. If you or anyone else found/find anything (I mean some IBM document), please feel free to let us know .
I think we all agree that an x86 port ain't happening nor should it. But while we're thinking processor realignment... how much work would it take to adapt OSX to Cell?
The answer is, that it depends on the Cell.
The first iteration of the Cell line is, according to Hannibal@Ars, not suitable for inclusion in a Mac. At any rate, as he points out, all (or at least, much) will be revealed at the ISSCC presentation by IBM.
My unfounded guess is that the PowerPC at the heart of the Cell is a PPC400 class, or equivalent.
Generally, I'd imagine that it would not be hard to get OS X to limp along on Cell, given that every Cell has a PowerPC core as far as I've read. Getting OS X to exploit Cell's capabilities... well, that's a different matter entirely, and since Cell's capabilities are not publicly known in any detail, it's hard even to speculate.
I think Power5 derivatives are more interesting possibilities for Apple in the near term. If IBM's engineers have been able to cash the rather large checks written by their PR flacks, those will be coming sooner rather than later. The claimed performance delta of a Power5 core over a Power4 core is staggering.
you have not seen anything regarding a Power5-based mobile PPC that could make it into a PowerBookG5 by summer/fall 2005? just curious, not even a hint anywhere?
The only news I know of came from Mr. Macphisto, who said that something was in the pipeline @ IBM a couple o'months ago. But details...nope.
If you feel the need to speculate, have a look at Power4 and the revisions that were made to it before it became the 970. Then repeat the same process for the Power5.
A lot of people have posted a lot of positive and negative comments on the powerbook updates. Does anyone have comments/considerations on the updates in relation to film editing? I?m editing in final cut pro?
If you feel the need to speculate, have a look at Power4 and the revisions that were made to it before it became the 970. Then repeat the same process for the Power5.
Actually, reading from the checks written by IBM's PR flacks, they engineered the Power5 so that a 900-class PowerPC could be derived from it immediately. The Power5 is a fundamentally different creature than the Power4, not in terms of its architecture, but in terms of its design goals: Relatively low cost, low power, etc. The Power4 was an ogre of a CPU in terms of size, cost, power consumption and (low) yields. The Power5 is not.
Again, if IBM's boffins (I love that word) have kept up with their marketing people—and I've been in the business way to long to take that for granted—then look for a PowerPC spinoff of the Power5 sooner rather than later.
Actually, reading from the checks written by IBM's PR flacks, they engineered the Power5 so that a 900-class PowerPC could be derived from it immediately. The Power5 is a fundamentally different creature than the Power4, not in terms of its architecture, but in terms of its design goals: Relatively low cost, low power, etc. The Power4 was an ogre of a CPU in terms of size, cost, power consumption and (low) yields. The Power5 is not.
Again, if IBM's boffins (I love that word) have kept up with their marketing people—and I've been in the business way to long to take that for granted—then look for a PowerPC spinoff of the Power5 sooner rather than later.
I agree. I would imagine that we will never see a 970 derived PPC in an Apple portable. The 970 has been around for quite a while. I would hope to see products based on a new processor reasonably soon (summer for PMs, maybe next Macworld for PBs).
And wasn't there always some murmurings that the 970 was a temporary measure to bring Apple up to parity (and even slightly exceed on launch) offerings from both Intel and AMD?
I agree. I would imagine that we will never see a 970 derived PPC in an Apple portable. The 970 has been around for quite a while. I would hope to see products based on a new processor reasonably soon (summer for PMs, maybe next Macworld for PBs).
And wasn't there always some murmurings that the 970 was a temporary measure to bring Apple up to parity (and even slightly exceed on launch) offerings from both Intel and AMD?
Possibly. I think the original intent was to have the 970 trickle down into portables, but it seems to be putting up a fight. We'll see. In the worst case, it's Freescale to the rescue with the 7448, 8641D, and then 87xx "G5" (because they're 64bit capable) CPUs. PB's post above covers them well.
A lot of people have posted a lot of positive and negative comments on the powerbook updates. Does anyone have comments/considerations on the updates in relation to film editing? I?m editing in final cut pro?
Tom
Tom, i think it is best to go to 'current hardware' part of the forums and ask people there, there's definitely a lot of FCP people around, they'll tell you what they've been doing on their 1ghz to 1.5ghz powerbooks ...
I think you'll need some real-world answers rather than getting bogged down in 'my ghz is bigger than your ghz' or 'my fsb kicks your fsb's ass' type discussions
1.67ghz 15" powerbook with superdrive, 128mb vram is now Current Hardware
What is important with PCIe, and I think it will find its way into Core Image/Video, Quartz 2D Extreme or whatever, is the capability to talk back to the CPU very fast.
I have convinced myself that with Core Video, Apple will need the downstream capability of the graphics bus. Now the question is if PCIe provides that big of a difference compared to AGP.
If Apple needs PCIe x16 for its pro-level machines post-Tiger, then that would present a problem for G4 systems. 74xx systems are limited by the FSB. 8461 (e600) systems only have PCIe x8. The G5 systems shouldn't have a problem.
Tom, i think it is best to go to 'current hardware' part of the forums and ask people there, there's definitely a lot of FCP people around, they'll tell you what they've been doing on their 1ghz to 1.5ghz powerbooks ...
I think you'll need some real-world answers rather than getting bogged down in 'my ghz is bigger than your ghz' or 'my fsb kicks your fsb's ass' type discussions
1.67ghz 15" powerbook with superdrive, 128mb vram is now Current Hardware
Comments
To me that is a far more relevant question.
Originally posted by RolandG
I have had the same feelings when the G4s stagnated at 500 MHz and Intel and AMD where scaling into the GHz as if there was no tomorrow. Speedwise, the stagnation threw Apple's desktop line back almost 2 or 3 processor iterations.
And the knowledge made by developing the M-series processors will finally make it into Intel's desktop line.
I still wonder why Apple is unable to put enough pressure onto their suppliers to improve certain aspects of their CPU designs (faster FSB in Freescales case, dynamic power and speed adjustments Ã* la Pentium M in both cases).
I'm not going to question the length of time it has taken for Apple to get decent processor speeds out of their suppliers...I wonder if this will feature in a chapter of Steve Jobs autobiography
You'd have to say that this MHz gap has improved over the past 2 years...both IBM and Freescale finally seem to have realised that there is a wider market for RISC chips...and quicker ones at that. Whilst the 7447B is ok, I have a sneaking suspicion the next PowerBook will see a greater performance jump.
The 7448 (see http://www.freescale.com/files/32bit.../MPC7450UM.pdf) is a step along the way to Dothan-like performance...increased Dynamic Frequency Switching (DFS), increased cache, improvements to Altivec, and perhaps a 20% improvement in clockspeed in the move to a 90nm process.
When we see the chip, hopefully at WWDC or just after, it will face a Pentium-M/Sonoma combo that is good but only slightly better than the Pentium-M/Centrino ( with a marginal clockspeed improvement and a quad pumped bus). The gap will have closed considerably in that time between the supposed inadequacy of the PowerBook and the pretender to the notebook throne. When the 8641D arrives at Macworld 2006, everyone will be happy. Optimistic? People, things are on the up
No more limited FSB and the G5 fanboys can go back to their mothers for coke and hugs
Originally posted by a j stev
Whilst the 7447B is ok, I have a sneaking suspicion the next PowerBook will see a greater performance jump.
agreed. on this note, what's the group consensus on how likely 7448 is for WWDC?
yesterday, i was pretty resolved to get one of the new 15" models. only downside is the lingering 167mhz bus.
but now i'm thinking i might do better with a stopgap (cheap) ibook til WWDC, then get a 7448 PB, and resell the ibook.
thoughts?
Originally posted by imiloa
agreed. on this note, what's the group consensus on how likely 7448 is for WWDC?
yesterday, i was pretty resolved to get one of the new 15" models. only downside is the lingering 167mhz bus.
but now i'm thinking i might do better with a stopgap (cheap) ibook til WWDC, then get a 7448 PB, and resell the ibook.
thoughts?
hey, go for it, you might end up holding on to your iBook a little longer than you expect
Originally posted by imiloa
agreed. on this note, what's the group consensus on how likely 7448 is for WWDC?
I would not bet on that. And by the way, let me remind the official Freescale roadmap from the other thread that died in Temporary Insanity:
(1) 7448, single e600 core, >1.5 GHz, 1 MB L2 cache with ECC, 200 MHz system bus, less than 10 W power consumption at 1.4 GHz.
(2) 8641D, dual e600 core, >1.5 GHz, 1 MB L2 cache with ECC (per processor core), integrated MPX bus up to 667 MHz (MUCH MUCH faster than the one we have for years in the G4 as a major bottleneck) with dual 64-bit DDR and DDR-II support, PCIe interface.
Availability is scheduled for 2H-05 and 1H-06 respectively, see the SNDF presentation (page 41).
What does this mean. With some good luck, we will have a 7448-based Powerbook by this year's summer-autumn and a 8641D-based (dual-core) Powerbook by the beginning of next year, say in one year from now. Now, the 7448 and 8641D are based on the e600 core which is 32-bit. But as Freescale says, there will be a 64-bit successor to the e600, with full 32-bit support, the e700. If Apple takes the Freescale road (and with all current evidence they have no other choice), I guess this e700 would perhaps be the G5-mobile for the Powerbook. Unfortunately, there are no available details on sampling and production for the e700 (personally, I would not expect it before 2007).
For the time being though, it seems that the Powerbook will follow a smooth evolution to a very capable 32-bit dual-core portable machine, without the system bottlenecks that plagued the platform for years. It remains to see if IBM comes with a response to Freescale. It would be very interesting indeed if they did. But there is zero evidence at this moment that they would do that before 2006.
will probably order in april/may, bit of a cash flow problem at the moment.
Originally posted by PB
It remains to see if IBM comes with a response to Freescale. It would be very interesting indeed if they did. But there is zero evidence at this moment that they would do that before 2006.
you have not seen anything regarding a Power5-based mobile PPC that could make it into a PowerBookG5 by summer/fall 2005? just curious, not even a hint anywhere?
Originally posted by sunilraman
you have not seen anything regarding a Power5-based mobile PPC that could make it into a PowerBookG5 by summer/fall 2005? just curious, not even a hint anywhere?
Sorry, no. If you or anyone else found/find anything (I mean some IBM document), please feel free to let us know
Originally posted by Dave J
I think we all agree that an x86 port ain't happening nor should it. But while we're thinking processor realignment... how much work would it take to adapt OSX to Cell?
The answer is, that it depends on the Cell.
The first iteration of the Cell line is, according to Hannibal@Ars, not suitable for inclusion in a Mac. At any rate, as he points out, all (or at least, much) will be revealed at the ISSCC presentation by IBM.
My unfounded guess is that the PowerPC at the heart of the Cell is a PPC400 class, or equivalent.
Generally, I'd imagine that it would not be hard to get OS X to limp along on Cell, given that every Cell has a PowerPC core as far as I've read. Getting OS X to exploit Cell's capabilities... well, that's a different matter entirely, and since Cell's capabilities are not publicly known in any detail, it's hard even to speculate.
I think Power5 derivatives are more interesting possibilities for Apple in the near term. If IBM's engineers have been able to cash the rather large checks written by their PR flacks, those will be coming sooner rather than later. The claimed performance delta of a Power5 core over a Power4 core is staggering.
Originally posted by sunilraman
you have not seen anything regarding a Power5-based mobile PPC that could make it into a PowerBookG5 by summer/fall 2005? just curious, not even a hint anywhere?
The only news I know of came from Mr. Macphisto, who said that something was in the pipeline @ IBM a couple o'months ago. But details...nope.
If you feel the need to speculate, have a look at Power4 and the revisions that were made to it before it became the 970. Then repeat the same process for the Power5.
Tom
Originally posted by a j stev
If you feel the need to speculate, have a look at Power4 and the revisions that were made to it before it became the 970. Then repeat the same process for the Power5.
Actually, reading from the checks written by IBM's PR flacks, they engineered the Power5 so that a 900-class PowerPC could be derived from it immediately. The Power5 is a fundamentally different creature than the Power4, not in terms of its architecture, but in terms of its design goals: Relatively low cost, low power, etc. The Power4 was an ogre of a CPU in terms of size, cost, power consumption and (low) yields. The Power5 is not.
Again, if IBM's boffins (I love that word) have kept up with their marketing people—and I've been in the business way to long to take that for granted—then look for a PowerPC spinoff of the Power5 sooner rather than later.
Originally posted by Amorph
Actually, reading from the checks written by IBM's PR flacks, they engineered the Power5 so that a 900-class PowerPC could be derived from it immediately. The Power5 is a fundamentally different creature than the Power4, not in terms of its architecture, but in terms of its design goals: Relatively low cost, low power, etc. The Power4 was an ogre of a CPU in terms of size, cost, power consumption and (low) yields. The Power5 is not.
Again, if IBM's boffins (I love that word) have kept up with their marketing people—and I've been in the business way to long to take that for granted—then look for a PowerPC spinoff of the Power5 sooner rather than later.
I agree. I would imagine that we will never see a 970 derived PPC in an Apple portable. The 970 has been around for quite a while. I would hope to see products based on a new processor reasonably soon (summer for PMs, maybe next Macworld for PBs).
And wasn't there always some murmurings that the 970 was a temporary measure to bring Apple up to parity (and even slightly exceed on launch) offerings from both Intel and AMD?
Originally posted by DaveLee
I agree. I would imagine that we will never see a 970 derived PPC in an Apple portable. The 970 has been around for quite a while. I would hope to see products based on a new processor reasonably soon (summer for PMs, maybe next Macworld for PBs).
And wasn't there always some murmurings that the 970 was a temporary measure to bring Apple up to parity (and even slightly exceed on launch) offerings from both Intel and AMD?
Possibly. I think the original intent was to have the 970 trickle down into portables, but it seems to be putting up a fight. We'll see. In the worst case, it's Freescale to the rescue with the 7448, 8641D, and then 87xx "G5" (because they're 64bit capable) CPUs. PB's post above covers them well.
Originally posted by maclayman
A lot of people have posted a lot of positive and negative comments on the powerbook updates. Does anyone have comments/considerations on the updates in relation to film editing? I?m editing in final cut pro?
Tom
Tom, i think it is best to go to 'current hardware' part of the forums and ask people there, there's definitely a lot of FCP people around, they'll tell you what they've been doing on their 1ghz to 1.5ghz powerbooks ...
I think you'll need some real-world answers rather than getting bogged down in 'my ghz is bigger than your ghz' or 'my fsb kicks your fsb's ass' type discussions
1.67ghz 15" powerbook with superdrive, 128mb vram is now Current Hardware
good luck
Originally posted by PB
What is important with PCIe, and I think it will find its way into Core Image/Video, Quartz 2D Extreme or whatever, is the capability to talk back to the CPU very fast.
I have convinced myself that with Core Video, Apple will need the downstream capability of the graphics bus. Now the question is if PCIe provides that big of a difference compared to AGP.
If Apple needs PCIe x16 for its pro-level machines post-Tiger, then that would present a problem for G4 systems. 74xx systems are limited by the FSB. 8461 (e600) systems only have PCIe x8. The G5 systems shouldn't have a problem.
Originally posted by sunilraman
Tom, i think it is best to go to 'current hardware' part of the forums and ask people there, there's definitely a lot of FCP people around, they'll tell you what they've been doing on their 1ghz to 1.5ghz powerbooks ...
I think you'll need some real-world answers rather than getting bogged down in 'my ghz is bigger than your ghz' or 'my fsb kicks your fsb's ass' type discussions
1.67ghz 15" powerbook with superdrive, 128mb vram is now Current Hardware
good luck
thanks for your advise..