InDesign CS2 has improved 16 bpc support as well; though it's not full featured as of yet (layer transparency modes if I recall correctly behave differently at the least on screen with 16 bpc images. As far as photography with DSLRs goes, 16 bpc is highly desirable, and iPhoto would certainly benefit from some level of support at the OS level.
The major issue I find with consumer grade CM however, is that consumer grade products rarely have good profiles associated with them. User generated profiles are nearly always better, oddly enough this is more important the further down a product line you go due to manufacturing tolerances.
OS level color management certainly makes life easier from a professional standpoint, though, I would agree. Ironically I remember the first mention of software able to handle 48bit RGB images that I saw was somewhere around the 1994 timeframe. A computer teacher I then talked to tried to argue why no one needed more than 256 or 32,000 colors (can't remember which).
InDesign CS2 has improved 16 bpc support as well; though it's not full featured as of yet (layer transparency modes if I recall correctly behave differently at the least on screen with 16 bpc images. As far as photography with DSLRs goes, 16 bpc is highly desirable, and iPhoto would certainly benefit from some level of support at the OS level.
The major issue I find with consumer grade CM however, is that consumer grade products rarely have good profiles associated with them. User generated profiles are nearly always better, oddly enough this is more important the further down a product line you go due to manufacturing tolerances.
OS level color management certainly makes life easier from a professional standpoint, though, I would agree. Ironically I remember the first mention of software able to handle 48bit RGB images that I saw was somewhere around the 1994 timeframe. A computer teacher I then talked to tried to argue why no one needed more than 256 or 32,000 colors (can't remember which).
Other than for marketing reasons, most low end consumers don't need 16 bit channels. But advanced amateurs and the commercial markets do.
To properly scare off some, there is even a 64 bit per channel HDR standard. It's used with custom made equipment by museums and for scientific work.
I'm surprised nobody talked about this yet: Woodcrest is the server version of the portable chip Merom. Moreover, Woodcrest is the successor of Sossaman, the server version of Yonah (all are Pentium-M derivatives) and NOT Dempsey (a Pentium-4D chip, design Apple will never use).
Apple said they will switch the laptop and consumer Mac line (think iBook, PowerBook, Mac mini) first, in June 2006. And they will finish the transition with high-end (think Power Mac & Xserve) by the end of 2007. So there is NO WAY that Apple would make an Intel-based PM or Xserve in 2006. For the Pro lines, they are waiting for the quad-core Whitefield.
Quote:
Originally posted by aegisdesign
On the other hand, I'm not buying a crappy Yonah based laptop so launching with Merom and Woodcrest appeals much more to me than the stopgap Yonah/Conroe chips.
You don't get it. Conroe (a desktop chip) does not come from this family. It will go for sure inside iMac, because it comes from the same newer design as Merom/Woodcrest (and not Yonah/Sossaman).
to Melgross: no offense but? Aren't Apple's <Power>Macs for the "pro" ?
I don't understand how in this boards people say "we need more power" and other say "most don't need that power"? This is Future Hardware isn't it ?
I wouldn't like Apple to focus anymore on ipods and imac, I need power and I had to switch from 3/4 macs to 3/4 PCs because of that I don't care a PowerMac is 4000 $, a highend PC is too?
Now, all my computers have hit the 32 bit wall for a while : the G5 because there is no real OS support (so no apps) and the PC because their proc doesn't support XP64. I need to buy new ones, and though I'd love to try those future intel PowerMacs, the first thing I need to now is : will mac OS X have a 64bit compliant GUI or not?
This is not a marketing issue for DCC people. (I like to think that Apple sees DCC as a market)?
I'm talking more of 50% to 100% speed advantage like 3D rendering speed? for the same price ! (this is for instance what I see when I compare my dual 2,5Ghz G5 with my dual 3GHz xeon though the mac has twice the ram and have cost roughly the same)?
First off, its Apple who better get used to Intel's business, rather then intel getting used to Apple. Intel doesn't NEED Apple. They like the added sales, but Apple's sales are a drop in the bucket compared to all the other PC makers. Intel's not going to kiss Apple's ass anymore then they have to. IBM and Motorola used to do that, and then they realized it wasn't worth it. You think Intel doesn't see this up-front?
Second, what exactly HAS to be different between Macs and Wintels besides design and OS? Reviewers aren't going to go "Yeah, but Dell sells the same box, so who cares?" or "Wow, Apple included this fancy new port and technology that no one else has. Sure, its of no use until peripherals come about, but why wait. Get your mac now!".
To put it another way, what exactly is different between a Windows/Dell box and an OS X/Apple box right now? Well, design and the OS. That's about it. And people keep saying "I don't care what chip is inside, its the OS that matters", or even "I don't even know what processor is in my computer, its the OS and performance, stupid!". So the fact that its a G5 vs a P4 is only important now to the lame few who get geeked up over hardware specs. So why does the new Mac-Intel boxes all of a sudden need some great 'extra' feature?
Right now, Intel gives its chips out to EVERYONE at the same time. Dell, HP, Gateway, etc, all have products waiting to sell when Intel announces a new chip. Dell doesn't get preferential treatment (because if they did, then the other big players could all go AMD, making things worse for Intel if just Dell went AMD). And Dell would only go Opteron if they could get as good a deal and the volumes that intel offers.
But, conversely, this is also why Apple WON'T get preferential treatment or early chips. They'll get chips when everyone else does, and release them when Intel allows it (hell, if Apple can tell Motorola when and how to show off a cell phone, Intel can tell Apple when to sell a new computer).
Ok. I really don't have time at the moment to argue through this with you but every single paragraph you have written is fundementally incorrect.
I don't understand how in this boards people say "we need more power" and other say "most don't need that power"? This is Future Hardware isn't it ?
I think it was a result of somebody saying apple was "doomed" if they didn't get such and such technology in very next powermac. This statement polarized responses.
I would be willing to bet that every single person here thinks that we will eventually have 64bit (at least) everything, and that there will be common uses for 64bit computing.
The dispute is about when it makes sense to adopt and push certain technologies.
I'm surprised nobody talked about this yet: Woodcrest is the server version of the portable chip Merom.
Actually, the Intel slides also said it was for 'Workstations' too. ie. PowerMacs.
Quote:
Originally posted by Cosmos 1999
Moreover, Woodcrest is the successor of Sossaman, the server version of Yonah (all are Pentium-M derivatives) and NOT Dempsey (a Pentium-4D chip, design Apple will never use).
Yes, but Sossaman won't be quicker than today's Xeons. Which incidentally get beaten to a pulp by today's Opterons. It's advantage is in low power/heat systems like blade servers not in workstations or full servers.
Quote:
Originally posted by Cosmos 1999
Apple said they will switch the laptop and consumer Mac line (think iBook, PowerBook, Mac mini) first, in June 2006.
They've never said the above. It's repeated over and over again but it's not attributable to anyone at Apple. Jobs didn't say it. Schiller neither. It's a fairly good bet since the Intel low powered Yonah based chips fit the most immediate product need - laptops. And that PowerMac users will still demand speed and the G5 will still beat Intel's offerings for some time to come, especially if they release a 970MP based PowerMac.
Quote:
Originally posted by Cosmos 1999
And they will finish the transition with high-end (think Power Mac & Xserve) by the end of 2007. So there is NO WAY that Apple would make an Intel-based PM or Xserve in 2006. For the Pro lines, they are waiting for the quad-core Whitefield.
The only difference between the Woodcrest and Whitefield is the size of the cache. They're the same chip.
Quote:
Originally posted by Cosmos 1999
You don't get it. Conroe (a desktop chip) does not come from this family. It will go for sure inside iMac, because it comes from the same newer design as Merom/Woodcrest (and not Yonah/Sossaman).
Sorry, yes, I was just thinking laptops and PowerMacs. IMHO we might still see Merom in the iMac just because of Apple's love of cool and quiet computers. G5 iMacs aren't exactly either of those.
They've never said the above. It's repeated over and over again but it's not attributable to anyone at Apple. Jobs didn't say it. Schiller neither. It's a fairly good bet since the Intel low powered Yonah based chips fit the most immediate product need - laptops. And that PowerMac users will still demand speed and the G5 will still beat Intel's offerings for some time to come, especially if they release a 970MP based PowerMac.
Quote from 09/20/2005 at Apple Expo : "Apple is on track to ship Intel-based computers as targeted by June 2006", Chief Executive Steve Jobs said on Tuesday. "We are on track to do that," Jobs told a news conference in Paris".
I personally think that this was the press who asked about the targeted timeframe of "June 2006", and that Jobs only answered the words "We are on track to do that". I concede this is very elusive and subject to many interpretations.
Oh, and Yonah will be available in January 2006.
Quote:
The only difference between the Woodcrest and Whitefield is the size of the cache. They're the same chip.
I thought that:
- Woodcrest is a dualcore chip
- Whitefield is a quadcore chip, first of the Nehalem family, with CSI (Common System Interface), competitor of AMD HyperTransport, and possibly with integrated memory controller.
I think that Apple has some good reasons to wait for the new processors.
- Speed : the 64 bits processor are faster than the 32 bits processors. Not because they deal with 64 bits values, but because they have more registers, and compiler will generate faster code. So when compared with regular windows software, the OSX software will be faster.
- Development cost : No need to support the 32 bit processors. It will save on testing and development for both Apple and software developers
I think that Apple has some good reasons to wait for the new processors.
- Speed : the 64 bits processor are faster than the 32 bits processors. Not because they deal with 64 bits values, but because they have more registers, and compiler will generate faster code. So when compared with regular windows software, the OSX software will be faster.
- Development cost : No need to support the 32 bit processors. It will save on testing and development for both Apple and software developers
So far, Intel 64bit offers no speed advantage over 32bit I gather. AMD on the other hand does. Intel had their hands forced to support AMD's 64bit ISA by Microsoft deciding to support that instead of Intel.
Development cost - totally agree but if that was the case developers would have seen 64bit compiler tools and the transition kit wouldn't be a Pentium4 32bit.
We'll have the Intel 32bit to 64bit transition to do some point in the future again. Roll on Universal Binaries with 3 ISAs in them - IA32, X64 and PPC.
to Melgross: no offense but? Aren't Apple's <Power>Macs for the "pro" ?
I don't understand how in this boards people say "we need more power" and other say "most don't need that power"? This is Future Hardware isn't it ?
I wouldn't like Apple to focus anymore on ipods and imac, I need power and I had to switch from 3/4 macs to 3/4 PCs because of that I don't care a PowerMac is 4000 $, a highend PC is too?
Now, all my computers have hit the 32 bit wall for a while : the G5 because there is no real OS support (so no apps) and the PC because their proc doesn't support XP64. I need to buy new ones, and though I'd love to try those future intel PowerMacs, the first thing I need to now is : will mac OS X have a 64bit compliant GUI or not?
This is not a marketing issue for DCC people. (I like to think that Apple sees DCC as a market)?
toma
I'm sorry if I don't have that "stars in my eyes" quality to my musings. I try to be realistic in my writing. Perhaps I've been in the pro area for too many decades, but there it is!
My experience is that a Powerbook *as it is now* is perfectly capable of doing pro level work .Most people on these boards are not pro's, but would like to think so. We've been happily doing feature level editing on Apple's laptops for years now. NO ONE editing on a PB thinks that it should be as fast as a hi-end desktop. It never will be. We don't use our Macs for this purpose because they are the fastest machines around. We know that they are not. There are other more intangible reasons why we use them. And those reasons will continue to be the reasons why we will use them.
Would it be better if they were twice as fast? Three times? Ten times? Yes of course. And you know what? They will be!
But for now, if someone needs a laptop to do work, the present ones work well enough so that it's foolish to pospone getting one if it is needed. Waiting for something new never allows you to get your work done NOW. And that's what I'm talking about. Doing your work now.
I think that Apple has some good reasons to wait for the new processors.
- Speed : the 64 bits processor are faster than the 32 bits processors. Not because they deal with 64 bits values, but because they have more registers, and compiler will generate faster code. So when compared with regular windows software, the OSX software will be faster.
- Development cost : No need to support the 32 bit processors. It will save on testing and development for both Apple and software developers
Let's separate out the differences between 64 bits on the Mac with the G5 and 64 bits on the PC with the 64 bit extensions.
The Mac sees no advantage in speed going to 64 bits because the PPC family has no legacy coding to overcome in going to 64 bits. The PPC family was designed as a 32 bit *clean* design from the beginning with the intended goal of moving easily and logically to 64 bits. It has plenty of registers etc. Therefore going to 64 bits doesn't result in a massive change to the cpu. It does result in much more memory needed in RAM.
The PC is different. The 32 bit x86 designs are filled with the detritus of leftover micro code 8 bit, 16 bit, and 32 bit code and other flotsam floating around on the substrate. This is a result of its 8 bit origins. It also has fewer registers that are smaller.
For this and other reasons when they came out with the 64 bit extensions, they had to fix those problems so that 64 bit code could work. 64 bit code doesn't use a fair amount of the old chips legacy wonkiness. It also has many new large registers etc. This means that 64 bit code works faster on a 64 bit x86 chip compared to the older 32 bit code. You do, therefore, get a speed advantage from going 64 bit. More RAM is required as well.
This is also good for Apple. Since x86 chips are now faster than the equivalent PPC chips, when Apple takes advantage of the newer 64 bit ones, things will only get faster still.
I'm certainly not putting the PPC's down here. As I said they have a much cleaner design which has been an advantage, but we do have to be realistic here. Development has not proceeded apace with the x86 side, and will not in the future. That is the business side of the picture.
So far, Intel 64bit offers no speed advantage over 32bit I gather. AMD on the other hand does. Intel had their hands forced to support AMD's 64bit ISA by Microsoft deciding to support that instead of Intel.
You shoul try to compile and run the same code compiled in 32 and 64 bit mode on an Intel processor ...
Except for memory moves, the use of the extra registers in the 64 bits mode improves speed.
Quote:
Development cost - totally agree but if that was the case developers would have seen 64bit compiler tools and the transition kit wouldn't be a Pentium4 32bit.
The processor used in the transition kit is a 64 bits processor.
..concludes that if applications are compiled with Intel's compiler then AMD32 apps run quicker than AMD64.
There are other reports.
I'm running CentOS 4.1 x64 on an Opteron 244 here and it's faster than x32 but not greatly so. And it has a number of application problems where some code just doesn't compile and run x64.
Quote:
Originally posted by MacAddict84
The processor used in the transition kit is a 64 bits processor.
But the OS, tools, compiler and all the documentation say it's 32bit. The transition kit isn't just the hardware. Apple aren't aiming at x64, which I find somewhat sad as we'll have all that legacy crap hanging about.
Its also only on these internet list do I see people value the productivity of a computer based on its processor speed.
I suppose this is because a 3Ghz computer is more productive than a 1.8Ghz computer is an easy concept to digest.
In a real work environment I?ve never seen anyone solely place value of a productive machine on its processor speed. Generally many factors are taken into account when deciding which platform to use, what OS, and which machine. All of this combined is taken into account and judged as productivity.
But the OS, tools, compiler and all the documentation say it's 32bit. The transition kit isn't just the hardware. Apple aren't aiming at x64, which I find somewhat sad as we'll have all that legacy crap hanging about.
That's one of the big mysteries for me.
Tiger already supports some 64 bit function in PPC.
But the x86 Tiger does not address AMDx64.
Because Intel will release x64 chips next year, most assume it eventually it will.
Looking at how Intel really feels about x64 its possible Apple will not go in that direction.
Comments
The major issue I find with consumer grade CM however, is that consumer grade products rarely have good profiles associated with them. User generated profiles are nearly always better, oddly enough this is more important the further down a product line you go due to manufacturing tolerances.
OS level color management certainly makes life easier from a professional standpoint, though, I would agree. Ironically I remember the first mention of software able to handle 48bit RGB images that I saw was somewhere around the 1994 timeframe. A computer teacher I then talked to tried to argue why no one needed more than 256 or 32,000 colors (can't remember which).
Originally posted by ChevalierMalFet
InDesign CS2 has improved 16 bpc support as well; though it's not full featured as of yet (layer transparency modes if I recall correctly behave differently at the least on screen with 16 bpc images. As far as photography with DSLRs goes, 16 bpc is highly desirable, and iPhoto would certainly benefit from some level of support at the OS level.
The major issue I find with consumer grade CM however, is that consumer grade products rarely have good profiles associated with them. User generated profiles are nearly always better, oddly enough this is more important the further down a product line you go due to manufacturing tolerances.
OS level color management certainly makes life easier from a professional standpoint, though, I would agree. Ironically I remember the first mention of software able to handle 48bit RGB images that I saw was somewhere around the 1994 timeframe. A computer teacher I then talked to tried to argue why no one needed more than 256 or 32,000 colors (can't remember which).
Other than for marketing reasons, most low end consumers don't need 16 bit channels. But advanced amateurs and the commercial markets do.
To properly scare off some, there is even a 64 bit per channel HDR standard. It's used with custom made equipment by museums and for scientific work.
Apple said they will switch the laptop and consumer Mac line (think iBook, PowerBook, Mac mini) first, in June 2006. And they will finish the transition with high-end (think Power Mac & Xserve) by the end of 2007. So there is NO WAY that Apple would make an Intel-based PM or Xserve in 2006. For the Pro lines, they are waiting for the quad-core Whitefield.
Originally posted by aegisdesign
On the other hand, I'm not buying a crappy Yonah based laptop so launching with Merom and Woodcrest appeals much more to me than the stopgap Yonah/Conroe chips.
You don't get it. Conroe (a desktop chip) does not come from this family. It will go for sure inside iMac, because it comes from the same newer design as Merom/Woodcrest (and not Yonah/Sossaman).
I don't understand how in this boards people say "we need more power" and other say "most don't need that power"? This is Future Hardware isn't it ?
I wouldn't like Apple to focus anymore on ipods and imac, I need power and I had to switch from 3/4 macs to 3/4 PCs because of that
Now, all my computers have hit the 32 bit wall for a while : the G5 because there is no real OS support (so no apps) and the PC because their proc doesn't support XP64. I need to buy new ones, and though I'd love to try those future intel PowerMacs, the first thing I need to now is : will mac OS X have a 64bit compliant GUI or not?
This is not a marketing issue for DCC people. (I like to think that Apple sees DCC as a market)?
toma
Originally posted by Louzer
First off, its Apple who better get used to Intel's business, rather then intel getting used to Apple. Intel doesn't NEED Apple. They like the added sales, but Apple's sales are a drop in the bucket compared to all the other PC makers. Intel's not going to kiss Apple's ass anymore then they have to. IBM and Motorola used to do that, and then they realized it wasn't worth it. You think Intel doesn't see this up-front?
Second, what exactly HAS to be different between Macs and Wintels besides design and OS? Reviewers aren't going to go "Yeah, but Dell sells the same box, so who cares?" or "Wow, Apple included this fancy new port and technology that no one else has. Sure, its of no use until peripherals come about, but why wait. Get your mac now!".
To put it another way, what exactly is different between a Windows/Dell box and an OS X/Apple box right now? Well, design and the OS. That's about it. And people keep saying "I don't care what chip is inside, its the OS that matters", or even "I don't even know what processor is in my computer, its the OS and performance, stupid!". So the fact that its a G5 vs a P4 is only important now to the lame few who get geeked up over hardware specs. So why does the new Mac-Intel boxes all of a sudden need some great 'extra' feature?
Right now, Intel gives its chips out to EVERYONE at the same time. Dell, HP, Gateway, etc, all have products waiting to sell when Intel announces a new chip. Dell doesn't get preferential treatment (because if they did, then the other big players could all go AMD, making things worse for Intel if just Dell went AMD). And Dell would only go Opteron if they could get as good a deal and the volumes that intel offers.
But, conversely, this is also why Apple WON'T get preferential treatment or early chips. They'll get chips when everyone else does, and release them when Intel allows it (hell, if Apple can tell Motorola when and how to show off a cell phone, Intel can tell Apple when to sell a new computer).
Ok. I really don't have time at the moment to argue through this with you but every single paragraph you have written is fundementally incorrect.
Originally posted by toma
I don't understand how in this boards people say "we need more power" and other say "most don't need that power"? This is Future Hardware isn't it ?
I think it was a result of somebody saying apple was "doomed" if they didn't get such and such technology in very next powermac. This statement polarized responses.
I would be willing to bet that every single person here thinks that we will eventually have 64bit (at least) everything, and that there will be common uses for 64bit computing.
The dispute is about when it makes sense to adopt and push certain technologies.
Originally posted by Cosmos 1999
I'm surprised nobody talked about this yet: Woodcrest is the server version of the portable chip Merom.
Actually, the Intel slides also said it was for 'Workstations' too. ie. PowerMacs.
Quote:
Originally posted by Cosmos 1999
Moreover, Woodcrest is the successor of Sossaman, the server version of Yonah (all are Pentium-M derivatives) and NOT Dempsey (a Pentium-4D chip, design Apple will never use).
Yes, but Sossaman won't be quicker than today's Xeons. Which incidentally get beaten to a pulp by today's Opterons. It's advantage is in low power/heat systems like blade servers not in workstations or full servers.
Quote:
Originally posted by Cosmos 1999
Apple said they will switch the laptop and consumer Mac line (think iBook, PowerBook, Mac mini) first, in June 2006.
They've never said the above. It's repeated over and over again but it's not attributable to anyone at Apple. Jobs didn't say it. Schiller neither. It's a fairly good bet since the Intel low powered Yonah based chips fit the most immediate product need - laptops. And that PowerMac users will still demand speed and the G5 will still beat Intel's offerings for some time to come, especially if they release a 970MP based PowerMac.
Quote:
Originally posted by Cosmos 1999
And they will finish the transition with high-end (think Power Mac & Xserve) by the end of 2007. So there is NO WAY that Apple would make an Intel-based PM or Xserve in 2006. For the Pro lines, they are waiting for the quad-core Whitefield.
The only difference between the Woodcrest and Whitefield is the size of the cache. They're the same chip.
Quote:
Originally posted by Cosmos 1999
You don't get it. Conroe (a desktop chip) does not come from this family. It will go for sure inside iMac, because it comes from the same newer design as Merom/Woodcrest (and not Yonah/Sossaman).
Sorry, yes, I was just thinking laptops and PowerMacs. IMHO we might still see Merom in the iMac just because of Apple's love of cool and quiet computers. G5 iMacs aren't exactly either of those.
Originally posted by aegisdesign
They've never said the above. It's repeated over and over again but it's not attributable to anyone at Apple. Jobs didn't say it. Schiller neither. It's a fairly good bet since the Intel low powered Yonah based chips fit the most immediate product need - laptops. And that PowerMac users will still demand speed and the G5 will still beat Intel's offerings for some time to come, especially if they release a 970MP based PowerMac.
Quote from 09/20/2005 at Apple Expo : "Apple is on track to ship Intel-based computers as targeted by June 2006", Chief Executive Steve Jobs said on Tuesday. "We are on track to do that," Jobs told a news conference in Paris".
I personally think that this was the press who asked about the targeted timeframe of "June 2006", and that Jobs only answered the words "We are on track to do that". I concede this is very elusive and subject to many interpretations.
Oh, and Yonah will be available in January 2006.
The only difference between the Woodcrest and Whitefield is the size of the cache. They're the same chip.
I thought that:
- Woodcrest is a dualcore chip
- Whitefield is a quadcore chip, first of the Nehalem family, with CSI (Common System Interface), competitor of AMD HyperTransport, and possibly with integrated memory controller.
He's never even said the words Yonah, Merom, or Woodcrest.
Any of it is speculation.
- Speed : the 64 bits processor are faster than the 32 bits processors. Not because they deal with 64 bits values, but because they have more registers, and compiler will generate faster code. So when compared with regular windows software, the OSX software will be faster.
- Development cost : No need to support the 32 bit processors. It will save on testing and development for both Apple and software developers
Originally posted by MacAddict84
I think that Apple has some good reasons to wait for the new processors.
- Speed : the 64 bits processor are faster than the 32 bits processors. Not because they deal with 64 bits values, but because they have more registers, and compiler will generate faster code. So when compared with regular windows software, the OSX software will be faster.
- Development cost : No need to support the 32 bit processors. It will save on testing and development for both Apple and software developers
So far, Intel 64bit offers no speed advantage over 32bit I gather. AMD on the other hand does. Intel had their hands forced to support AMD's 64bit ISA by Microsoft deciding to support that instead of Intel.
Development cost - totally agree but if that was the case developers would have seen 64bit compiler tools and the transition kit wouldn't be a Pentium4 32bit.
We'll have the Intel 32bit to 64bit transition to do some point in the future again. Roll on Universal Binaries with 3 ISAs in them - IA32, X64 and PPC.
Originally posted by toma
to Melgross: no offense but? Aren't Apple's <Power>Macs for the "pro" ?
I don't understand how in this boards people say "we need more power" and other say "most don't need that power"? This is Future Hardware isn't it ?
I wouldn't like Apple to focus anymore on ipods and imac, I need power and I had to switch from 3/4 macs to 3/4 PCs because of that
Now, all my computers have hit the 32 bit wall for a while : the G5 because there is no real OS support (so no apps) and the PC because their proc doesn't support XP64. I need to buy new ones, and though I'd love to try those future intel PowerMacs, the first thing I need to now is : will mac OS X have a 64bit compliant GUI or not?
This is not a marketing issue for DCC people. (I like to think that Apple sees DCC as a market)?
toma
I'm sorry if I don't have that "stars in my eyes" quality to my musings. I try to be realistic in my writing. Perhaps I've been in the pro area for too many decades, but there it is!
My experience is that a Powerbook *as it is now* is perfectly capable of doing pro level work .Most people on these boards are not pro's, but would like to think so. We've been happily doing feature level editing on Apple's laptops for years now. NO ONE editing on a PB thinks that it should be as fast as a hi-end desktop. It never will be. We don't use our Macs for this purpose because they are the fastest machines around. We know that they are not. There are other more intangible reasons why we use them. And those reasons will continue to be the reasons why we will use them.
Would it be better if they were twice as fast? Three times? Ten times? Yes of course. And you know what? They will be!
But for now, if someone needs a laptop to do work, the present ones work well enough so that it's foolish to pospone getting one if it is needed. Waiting for something new never allows you to get your work done NOW. And that's what I'm talking about. Doing your work now.
Originally posted by MacAddict84
I think that Apple has some good reasons to wait for the new processors.
- Speed : the 64 bits processor are faster than the 32 bits processors. Not because they deal with 64 bits values, but because they have more registers, and compiler will generate faster code. So when compared with regular windows software, the OSX software will be faster.
- Development cost : No need to support the 32 bit processors. It will save on testing and development for both Apple and software developers
Let's separate out the differences between 64 bits on the Mac with the G5 and 64 bits on the PC with the 64 bit extensions.
The Mac sees no advantage in speed going to 64 bits because the PPC family has no legacy coding to overcome in going to 64 bits. The PPC family was designed as a 32 bit *clean* design from the beginning with the intended goal of moving easily and logically to 64 bits. It has plenty of registers etc. Therefore going to 64 bits doesn't result in a massive change to the cpu. It does result in much more memory needed in RAM.
The PC is different. The 32 bit x86 designs are filled with the detritus of leftover micro code 8 bit, 16 bit, and 32 bit code and other flotsam floating around on the substrate. This is a result of its 8 bit origins. It also has fewer registers that are smaller.
For this and other reasons when they came out with the 64 bit extensions, they had to fix those problems so that 64 bit code could work. 64 bit code doesn't use a fair amount of the old chips legacy wonkiness. It also has many new large registers etc. This means that 64 bit code works faster on a 64 bit x86 chip compared to the older 32 bit code. You do, therefore, get a speed advantage from going 64 bit. More RAM is required as well.
This is also good for Apple. Since x86 chips are now faster than the equivalent PPC chips, when Apple takes advantage of the newer 64 bit ones, things will only get faster still.
I'm certainly not putting the PPC's down here. As I said they have a much cleaner design which has been an advantage, but we do have to be realistic here. Development has not proceeded apace with the x86 side, and will not in the future. That is the business side of the picture.
Originally posted by aegisdesign
So far, Intel 64bit offers no speed advantage over 32bit I gather. AMD on the other hand does. Intel had their hands forced to support AMD's 64bit ISA by Microsoft deciding to support that instead of Intel.
You shoul try to compile and run the same code compiled in 32 and 64 bit mode on an Intel processor ...
Except for memory moves, the use of the extra registers in the 64 bits mode improves speed.
Development cost - totally agree but if that was the case developers would have seen 64bit compiler tools and the transition kit wouldn't be a Pentium4 32bit.
The processor used in the transition kit is a 64 bits processor.
Originally posted by MacAddict84
You shoul try to compile and run the same code compiled in 32 and 64 bit mode on an Intel processor ...
Except for memory moves, the use of the extra registers in the 64 bits mode improves speed.
I've heard differing reports. Sometimes 32bit is actually quicker and it depends on the OS and compiler.
For instance...
http://www.xbitlabs.com/articles/cpu...-64-rc1_9.html
..concludes that if applications are compiled with Intel's compiler then AMD32 apps run quicker than AMD64.
There are other reports.
I'm running CentOS 4.1 x64 on an Opteron 244 here and it's faster than x32 but not greatly so. And it has a number of application problems where some code just doesn't compile and run x64.
Originally posted by MacAddict84
The processor used in the transition kit is a 64 bits processor.
But the OS, tools, compiler and all the documentation say it's 32bit. The transition kit isn't just the hardware. Apple aren't aiming at x64, which I find somewhat sad as we'll have all that legacy crap hanging about.
I suppose this is because a 3Ghz computer is more productive than a 1.8Ghz computer is an easy concept to digest.
In a real work environment I?ve never seen anyone solely place value of a productive machine on its processor speed. Generally many factors are taken into account when deciding which platform to use, what OS, and which machine. All of this combined is taken into account and judged as productivity.
But the OS, tools, compiler and all the documentation say it's 32bit. The transition kit isn't just the hardware. Apple aren't aiming at x64, which I find somewhat sad as we'll have all that legacy crap hanging about.
That's one of the big mysteries for me.
Tiger already supports some 64 bit function in PPC.
But the x86 Tiger does not address AMDx64.
Because Intel will release x64 chips next year, most assume it eventually it will.
Looking at how Intel really feels about x64 its possible Apple will not go in that direction.
Originally posted by TenoBell
That's the whole question for me.
Tiger already supports some 64 bit function in PPC.
But the x86 Tiger does not address AMDx64.
Because Intel will release x64 chips next year, most assume it eventually it will.
Looking at how Intel really feels about x64 its possible Apple will not go in that direction.
Look guys, If Apple comes out with the first machine(s) by JUNE, there won't be a 64 bit cpu in them. That means that a 64 bit OS serves no purpose.
As we won't be seeing an iMac until at least the 2nd half of 2006 when the 64 bit chips Apple will be using will arrive, Apple is on firm ground here.
I'm willing to bet that Leopard, at the end of 2006 will be 64 bit - just in time for Apple's 64 bit machines.
Stop worrying. Be happy.