Ah! But if it were that simple, everyone would have already done it, including Apple. This is really quite complex.
Well, everyone has kinda already done it... the idea of using queues to manage parallel processing is not new (e.g. bank queues). Apple's innovation is moving it from programmer control to OS control - grand centralizing it. With knowledge of overall system load the OS is where this job should have been done all along.
But moving it from here to there shouldn't require a new API to be designed...
I hope the word "Grand" isn't going to anyone's head in Cupertino. Keep it simple and it will be a major success, but the path is littered with APIs that developers did not adopt due to complexity.
Well, everyone has kinda already done it... the idea of using queues to manage parallel processing is not new (e.g. bank queues). Apple's innovation is moving it from programmer control to OS control - grand centralizing it. With knowledge of overall system load the OS is where this job should have been done all along.
But moving it from here to there shouldn't require a new API to be designed...
If Apple does go for a 64 bit kernel, then we will NEED all new 64 bit drivers. Remember how that screwed Win 64. I've got some expensive equipment here, I want to know they will work properly, or I won't upgrade. If people won't upgrade because of that, what advantage will they have gotten by coming out with the OS early?
That's easy. None!!!
In fact, it will be a marketing fiasco.
For people who are complaining on the other thread about lack of FW on the new MB, and the use of a mini DisplayPort connector on another, this will be a hurricane compared to the breeze of those complaints.
And the Firewire thread was a 1500 comment thread.
I don't know, but OS development at Apple is getting me depressed. I was so disappointed by Leopard. When it was released it was clear that they rushed it out - and they were late, by their own admission because it OS X 10.5 just was not a priority compared to iPhone development.
To me the entire reason SL exists is that this is what Leopard should have been to begin with.
I have simply skipped Leopard. I'm still on 10.4.11 - and from looking at my friend's adventures with Leopard to this day, Tiger is STILL more stable. And frankly there are no features of Leopard that I'm salivating over.
I wanted to buy some new hardware - iMac - but I wanted to wait for SL before I did that. Now it means I'll have to wait longer. Anyhow, I wasn't going to rush out the moment 10.6 hit the street. I'll wait for 10.6.2 at least.
I truly hope SL will be worth waiting for, since Leopard was a giant disappointment. I've thought 10.2, 10.3, 10.4 were all significant jumps and improvements, but 10.5 was a huge disappointment. I'm keeping my fingers crossed for SL.
Who ever wrote this article did a pathetically terrible job at it.
Nothing of Snow Leopard is 'building out' - learn what building out means, building out would be the source code is being built and features that were promised are nothing being build (aka build out).
Please, don't use terms unless you know what the hell it means, its pathetic to see it in action so many damn times on this site. The correct term is that they're still building *IN* features. *IN* means features going *IN*, *OUT* means that features are being taken *OUT*.
Apple no longer has the burden of devoting considerable resources to carbon developement. I'd expect that that has freed up many an engineer to work on
other tasks.
OpenCL is a group led effort with much input from Nvidia, Intel, AMD/ATI and Khronos Group. Apple's load here should be more minimal than home grown stuff.
I think Snow Leopard should be delivered when it lives up to its promise to polish and optimize OS X to the point where there aren't a bunch of complaints about x feature not working as purported.
I see a spring delivery perhaps in late April.
OpenCL was written by Apple and introduced to the Group at Khronos to adopt to, not the other way around which you imply as, ``I have this idea and can you do the heavy lifting for us to then write our APIs and put into the OS, thanks so much.''
Apple wrote OpenCL. They did the heavy lifting as a multi-year project. They had Nvidia, ATi and Intel in scope when they wrote it. They then brought in those bodies when it made sense to then work out the rest to ratify it as quickly as possible.
If it weren't for Apple doing the heavy lifting it would have stuck in limbo for a few years before finalization. That is what made it amazing to Nvidia, ATi and Intel. They didn't have to do much to have their stuff work with it.
Who ever wrote this article did a pathetically terrible job at it.
Nothing of Snow Leopard is 'building out' - learn what building out means, building out would be the source code is being built and features that were promised are nothing being build (aka build out).
Please, don't use terms unless you know what the hell it means, its pathetic to see it in action so many damn times on this site. The correct term is that they're still building *IN* features. *IN* means features going *IN*, *OUT* means that features are being taken *OUT*.
Are you familiar with this term in a technical way? Because it's also used here, in the USA, to mean making larger, as in "building out to there."
Removing, is the term you mean. At least that would be the way we would put it.
We were also promised QuickTime X on the same time line yet developers have not seen hide nor hair of that.
System-level stuff will come first. Quicktime is mainly user-level.
Quote:
Originally Posted by melgross
Really? Who that you know is working on saying that? The possible red circle around the date the we read of? That's certainly nothing. It's already been asserted that it could have been a mistake.
It was the head 'BSD-side of OS X' developer who wrote it. I thought it could have been a mistake but someone pointed out he wrote it twice. It could have been another error rather than a typo though.
Quote:
Originally Posted by melgross
If Apple does go for a 64 bit kernel, then we will NEED all new 64 bit drivers. Remember how that screwed Win 64. I've got some expensive equipment here, I want to know they will work properly, or I won't upgrade. If people won't upgrade because of that, what advantage will they have gotten by coming out with the OS early?
I wonder if it will be all that bad. Reports have said that SL will run on the original Core Duo 32-bit Intel Macs so SL should have both 32-bit and 64-bit support. Maybe they have a way to ensure 32-bit drivers can still run.
I don't have 3rd party drivers for any of my devices though so if they do the built-in ones, everything should work fine and they'll want top get the all 64-bit even if they can support 32-bit ones. Vista's problem is that it supports much more drivers than the Mac and a lot of 3rd party drivers broke.
Quote:
Originally Posted by melgross
Well, FCS comes to mind. I think Shake is still in Carbon, though I'm not 100% on that.
Yeah Shake is still Carbon. I don't think they are updating it to Cocoa. It looks like they are dismantling it to use with a dumbed down Cocoa app like Motion much to my annoyance. No word on any major improvements in well over 3 years now I think. One of the stupidest moves they've ever made IMO. Buying one of the most important pieces of software the film industry has seen and then killing it.
It would be like buying Renderman and using the code to do render the reflections on the Dock.
They could of course be rewriting these apps from the ground up but I'll reserve the right to hate them every day until they say it's the case.
I actually find it a bit hypocritical how they tell developers like Adobe to transition to Cocoa when they haven't done it themselves.
Quote:
Originally Posted by ascii
Apple's innovation is moving it from programmer control to OS control - grand centralizing it. With knowledge of overall system load the OS is where this job should have been done all along.
But moving it from here to there shouldn't require a new API to be designed...
I think it will need a change to programming conventions. If standard code could be parallelized easily, people would have done it ages ago. I'm not sure an API will do it but there may be guidelines as to how to program for parallel processing. It won't work for all code but hopefully there will be some automated compiling that does a better job than auto-vectorization on top that deals with threading.
Quote:
Originally Posted by TBaggins
And the Firewire thread was a 1500 comment thread.
Otwayross made half of those though. I think he lives there now.
Quote:
Originally Posted by FineWine
I have simply skipped Leopard. I'm still on 10.4.11 - and from looking at my friend's adventures with Leopard to this day, Tiger is STILL more stable. And frankly there are no features of Leopard that I'm salivating over.
Same here. I also find the graphics to be a bit to in my face. Drop shadows, dock etc. Same the iphone SDK is Leopard-only. Yet another example of Apple forcing upgrades on people.
I wonder if it will be all that bad. Reports have said that SL will run on the original Core Duo 32-bit Intel Macs so SL should have both 32-bit and 64-bit support. Maybe they have a way to ensure 32-bit drivers can still run.
A 64 bit kernel will require 64 bit drivers. There's no way around that at all. If it will somehow run on a 32 bit machine, then that machine, and version of the OS can use a 32 bit kernel. If they can somehow do that, then for those machines only, 32 bit drivers will suffice.
Quote:
I don't have 3rd party drivers for any of my devices though so if they do the built-in ones, everything should work fine and they'll want top get the all 64-bit even if they can support 32-bit ones. Vista's problem is that it supports much more drivers than the Mac and a lot of 3rd party drivers broke.
Some of the more sophisticated devices need their own drivers and software for full feature support, even if Apple has their own.
My Canon iPF5100 printer requires Canon's software and Photoshop driver. My old, but still pretty good Hp CLS8500N printer can use Apple's drivers, but Hp's are much better.
Same with my scanners, cameras etc.
Some of these drivers are very complex, and Apple won't be able to duplicate drivers for all the machines out there. That's the same problem MS had with Vista. Ad MS has far more people to do this than Apple does. Apple will need these companies to do the drivers for their own machines. What happens when a machine is discontinued? One can't expect companies to work on drivers for machines they haven't sold for some time, even if they are still in top notch condition operationally, and quality wise.
Quote:
Yeah Shake is still Carbon. I don't think they are updating it to Cocoa. It looks like they are dismantling it to use with a dumbed down Cocoa app like Motion much to my annoyance. No word on any major improvements in well over 3 years now I think. One of the stupidest moves they've ever made IMO. Buying one of the most important pieces of software the film industry has seen and then killing it.
It would be like buying Renderman and using the code to do render the reflections on the Dock.
They could of course be rewriting these apps from the ground up but I'll reserve the right to hate them every day until they say it's the case.
I actually find it a bit hypocritical how they tell developers like Adobe to transition to Cocoa when they haven't done it themselves.
Well, I keep telling people who dump on MS and Adobe about this situation that Apple is in the same boat, but it doesn't wake them up.
I hope they aren't killing Shake. Hopefully they are re-writing it. They certainly must be working hard on FCS though. At least they should be, considering how it has the video industry in such a stranglehold these days.
Quote:
I think it will need a change to programming conventions. If standard code could be parallelized easily, people would have done it ages ago. I'm not sure an API will do it but there may be guidelines as to how to program for parallel processing. It won't work for all code but hopefully there will be some automated compiling that does a better job than auto-vectorization on top that deals with threading.
That's right. What's even worse is programming for hyperthreading. If that isn't done right, the program actually slows down. If Apple wants to include that in their specs, they've got a lot of work cut out for them.
A 64 bit kernel will require 64 bit drivers. There's no way around that at all. If it will somehow run on a 32 bit machine, then that machine, and version of the OS can use a 32 bit kernel. If they can somehow do that, then for those machines only, 32 bit drivers will suffice.
Of course there are ways around that. Just because nobody does it (maybe it's even already done) doesn't mean it's impossible.
It seems fairly certain that it will likely be in March at the earliest though and this is stretching Q1 to the limit so we'll just have to see how developments go over the next couple of months and maybe get some info at MW.
The urgency is for the apps and drivers. I've only recently considered moving to Leopard because I don't want my apps to start breaking and introducing bugs that weren't there before.
My view is if it ain't broke don't upgrade unless it offers something major. Snow Leopard should offer something major but I want it to be in the hands of all developers soon so that they can make sure the apps work properly.
Otherwise, I might not upgrade for another year from now. If it's out in March then I might be able to upgrade earlier.
Are you saying that you haven't made up your mind and are still using Tiger or something earlier?
Since there is no new features and it's pretty much a leopard bug fix, why rush it? The whole snow leopard thing is making leopard faster and smaller in file size. There should be nothing wrong with this release. So take your time Apple and make sure it's all 100%.
Since there is no new features and it's pretty much a leopard bug fix, why rush it? The whole snow leopard thing is making leopard faster and smaller in file size. There should be nothing wrong with this release. So take your time Apple and make sure it's all 100%.
Correct in part. But it is not a "…a leopard bug fix,…"
Since 2001, Mac OS X has delivered more than a thousand innovative new features. With Snow Leopard, the next major version of the world’s most advanced operating system, Mac OS X changes more than its spots, it changes focus. Taking a break from adding new features, Snow Leopard — scheduled to ship in about a year — builds on Leopard’s enormous innovations by delivering a new generation of core software technologies that will streamline Mac OS X, enhance its performance, and set new standards for quality. Snow Leopard dramatically reduces the footprint of Mac OS X, making it even more efficient for users, and giving them back valuable hard drive space for their music and photoshttp://www.apple.com/macosx/snowleopard/
And the only official press release from Apple says in part:
SAN FRANCISCO—June 9, 2008—Apple® today previewed Mac OS® X Snow Leopard, which builds on the incredible success of OS X Leopard and is the next major version of the world’s most advanced operating system. Rather than focusing primarily on new features, Snow Leopard will enhance the performance of OS X, set a new standard for quality and lay the foundation for future OS X innovation. Snow Leopard is optimized for multi-core processors, taps into the vast computing power of graphic processing units (GPUs), enables breakthrough amounts of RAM and features a new, modern media platform with QuickTime® X. Snow Leopard includes out-of-the-box support for Microsoft Exchange 2007 and is scheduled to ship in about a year. http://www.apple.com/pr/library/2008...owleopard.html
I would suggest that any report indicating a specific introductory time is conjecture. As Apple has published on two sites, it "is scheduled to ship in about a year" (dated June 9, 2008). The question now would be, "How long is about a year?" Some would say that it could lie anywhere between 6 months plus a day to 18 months less a day. In any event, it is somewhere within that period, it depends to what you round time off to, and it is relative.
I just hope they don't release it until its ready. Too many times Apple has rushed out an OS. Who cares if its released this time next year. Its not like Tiger or Leopard are terrible OS's or something. Now is the time to take their time and get it 99% right the first time and not have to rush around and release 10.6.1 as quickly as possible to fix bugs that should have been caught before it was released.
I just hope they don't release it until its ready. Too many times Apple has rushed out an OS. Who cares if its released this time next year. Its not like Tiger or Leopard are terrible OS's or something. Now is the time to take their time and get it 99% right the first time and not have to rush around and release 10.6.1 as quickly as possible to fix bugs that should have been caught before it was released.
1) You always have the option to not buy it until it has matured a bit and reports are showing that it's quite stable.
2) Hopefully Apple learned from the iPhone 3G/iPhone v2.0/App Store/MobileMe fiasco.
OpenCL was written by Apple and introduced to the Group at Khronos to adopt to, not the other way around which you imply as, ``I have this idea and can you do the heavy lifting for us to then write our APIs and put into the OS, thanks so much.''
Apple wrote OpenCL. They did the heavy lifting as a multi-year project. They had Nvidia, ATi and Intel in scope when they wrote it. They then brought in those bodies when it made sense to then work out the rest to ratify it as quickly as possible.
If it weren't for Apple doing the heavy lifting it would have stuck in limbo for a few years before finalization. That is what made it amazing to Nvidia, ATi and Intel. They didn't have to do much to have their stuff work with it.
actually, OpenCL is mostly based on CUDA with some Apple people doing some nifty dynamic compilation on top.
CUDA has been ported to the CPU as well to NVIDIA GPUS (and ATIs are getting there). Basically they are going to be doing super-threading algorithms that can run on any streaming device, or on a normal device that is not as fast but does not break the programming model so it can scale across different cores in the system (from different vendors, but it is geared towards NVIDIA initially)
Although, I always get a kick about working in a project and getting corrected from an outside source :-) hint.. hint...
I don't know if this guy is blowing smoke or not but it is interesting that OpenCL seems to favor Nvidia GPU. I would have thought that equal support for AMD/ATI would have been developed concomittantly seing as how Apple's lineup always has a model with ATI graphics.
Comments
Isn't iTunes still written in Carbon? When will that ever be moved to Cocoa? That has to be the largest app right now that still uses Carbon.
Well, FCS comes to mind. I think Shake is still in Carbon, though I'm not 100% on that.
Ah! But if it were that simple, everyone would have already done it, including Apple. This is really quite complex.
Well, everyone has kinda already done it... the idea of using queues to manage parallel processing is not new (e.g. bank queues). Apple's innovation is moving it from programmer control to OS control - grand centralizing it. With knowledge of overall system load the OS is where this job should have been done all along.
But moving it from here to there shouldn't require a new API to be designed...
I hope the word "Grand" isn't going to anyone's head in Cupertino. Keep it simple and it will be a major success, but the path is littered with APIs that developers did not adopt due to complexity.
Well, everyone has kinda already done it... the idea of using queues to manage parallel processing is not new (e.g. bank queues). Apple's innovation is moving it from programmer control to OS control - grand centralizing it. With knowledge of overall system load the OS is where this job should have been done all along.
But moving it from here to there shouldn't require a new API to be designed...
Apparently, it does.
If Apple does go for a 64 bit kernel, then we will NEED all new 64 bit drivers. Remember how that screwed Win 64. I've got some expensive equipment here, I want to know they will work properly, or I won't upgrade. If people won't upgrade because of that, what advantage will they have gotten by coming out with the OS early?
That's easy. None!!!
In fact, it will be a marketing fiasco.
For people who are complaining on the other thread about lack of FW on the new MB, and the use of a mini DisplayPort connector on another, this will be a hurricane compared to the breeze of those complaints.
And the Firewire thread was a 1500 comment thread.
...
And the Firewire thread was a 1500 comment thread.
...
Exactly!
To me the entire reason SL exists is that this is what Leopard should have been to begin with.
I have simply skipped Leopard. I'm still on 10.4.11 - and from looking at my friend's adventures with Leopard to this day, Tiger is STILL more stable. And frankly there are no features of Leopard that I'm salivating over.
I wanted to buy some new hardware - iMac - but I wanted to wait for SL before I did that. Now it means I'll have to wait longer. Anyhow, I wasn't going to rush out the moment 10.6 hit the street. I'll wait for 10.6.2 at least.
I truly hope SL will be worth waiting for, since Leopard was a giant disappointment. I've thought 10.2, 10.3, 10.4 were all significant jumps and improvements, but 10.5 was a huge disappointment. I'm keeping my fingers crossed for SL.
Nothing of Snow Leopard is 'building out' - learn what building out means, building out would be the source code is being built and features that were promised are nothing being build (aka build out).
Please, don't use terms unless you know what the hell it means, its pathetic to see it in action so many damn times on this site. The correct term is that they're still building *IN* features. *IN* means features going *IN*, *OUT* means that features are being taken *OUT*.
Apple no longer has the burden of devoting considerable resources to carbon developement. I'd expect that that has freed up many an engineer to work on
other tasks.
OpenCL is a group led effort with much input from Nvidia, Intel, AMD/ATI and Khronos Group. Apple's load here should be more minimal than home grown stuff.
I think Snow Leopard should be delivered when it lives up to its promise to polish and optimize OS X to the point where there aren't a bunch of complaints about x feature not working as purported.
I see a spring delivery perhaps in late April.
OpenCL was written by Apple and introduced to the Group at Khronos to adopt to, not the other way around which you imply as, ``I have this idea and can you do the heavy lifting for us to then write our APIs and put into the OS, thanks so much.''
Apple wrote OpenCL. They did the heavy lifting as a multi-year project. They had Nvidia, ATi and Intel in scope when they wrote it. They then brought in those bodies when it made sense to then work out the rest to ratify it as quickly as possible.
If it weren't for Apple doing the heavy lifting it would have stuck in limbo for a few years before finalization. That is what made it amazing to Nvidia, ATi and Intel. They didn't have to do much to have their stuff work with it.
Get it?
Who ever wrote this article did a pathetically terrible job at it.
Nothing of Snow Leopard is 'building out' - learn what building out means, building out would be the source code is being built and features that were promised are nothing being build (aka build out).
Please, don't use terms unless you know what the hell it means, its pathetic to see it in action so many damn times on this site. The correct term is that they're still building *IN* features. *IN* means features going *IN*, *OUT* means that features are being taken *OUT*.
Are you familiar with this term in a technical way? Because it's also used here, in the USA, to mean making larger, as in "building out to there."
Removing, is the term you mean. At least that would be the way we would put it.
We were also promised QuickTime X on the same time line yet developers have not seen hide nor hair of that.
System-level stuff will come first. Quicktime is mainly user-level.
Really? Who that you know is working on saying that? The possible red circle around the date the we read of? That's certainly nothing. It's already been asserted that it could have been a mistake.
It was the head 'BSD-side of OS X' developer who wrote it. I thought it could have been a mistake but someone pointed out he wrote it twice. It could have been another error rather than a typo though.
If Apple does go for a 64 bit kernel, then we will NEED all new 64 bit drivers. Remember how that screwed Win 64. I've got some expensive equipment here, I want to know they will work properly, or I won't upgrade. If people won't upgrade because of that, what advantage will they have gotten by coming out with the OS early?
I wonder if it will be all that bad. Reports have said that SL will run on the original Core Duo 32-bit Intel Macs so SL should have both 32-bit and 64-bit support. Maybe they have a way to ensure 32-bit drivers can still run.
I don't have 3rd party drivers for any of my devices though so if they do the built-in ones, everything should work fine and they'll want top get the all 64-bit even if they can support 32-bit ones. Vista's problem is that it supports much more drivers than the Mac and a lot of 3rd party drivers broke.
Well, FCS comes to mind. I think Shake is still in Carbon, though I'm not 100% on that.
Yeah Shake is still Carbon. I don't think they are updating it to Cocoa. It looks like they are dismantling it to use with a dumbed down Cocoa app like Motion much to my annoyance. No word on any major improvements in well over 3 years now I think. One of the stupidest moves they've ever made IMO. Buying one of the most important pieces of software the film industry has seen and then killing it.
It would be like buying Renderman and using the code to do render the reflections on the Dock.
They could of course be rewriting these apps from the ground up but I'll reserve the right to hate them every day until they say it's the case.
I actually find it a bit hypocritical how they tell developers like Adobe to transition to Cocoa when they haven't done it themselves.
Apple's innovation is moving it from programmer control to OS control - grand centralizing it. With knowledge of overall system load the OS is where this job should have been done all along.
But moving it from here to there shouldn't require a new API to be designed...
I think it will need a change to programming conventions. If standard code could be parallelized easily, people would have done it ages ago. I'm not sure an API will do it but there may be guidelines as to how to program for parallel processing. It won't work for all code but hopefully there will be some automated compiling that does a better job than auto-vectorization on top that deals with threading.
And the Firewire thread was a 1500 comment thread.
Otwayross made half of those though. I think he lives there now.
I have simply skipped Leopard. I'm still on 10.4.11 - and from looking at my friend's adventures with Leopard to this day, Tiger is STILL more stable. And frankly there are no features of Leopard that I'm salivating over.
Same here. I also find the graphics to be a bit to in my face. Drop shadows, dock etc. Same the iphone SDK is Leopard-only. Yet another example of Apple forcing upgrades on people.
I wonder if it will be all that bad. Reports have said that SL will run on the original Core Duo 32-bit Intel Macs so SL should have both 32-bit and 64-bit support. Maybe they have a way to ensure 32-bit drivers can still run.
A 64 bit kernel will require 64 bit drivers. There's no way around that at all. If it will somehow run on a 32 bit machine, then that machine, and version of the OS can use a 32 bit kernel. If they can somehow do that, then for those machines only, 32 bit drivers will suffice.
I don't have 3rd party drivers for any of my devices though so if they do the built-in ones, everything should work fine and they'll want top get the all 64-bit even if they can support 32-bit ones. Vista's problem is that it supports much more drivers than the Mac and a lot of 3rd party drivers broke.
Some of the more sophisticated devices need their own drivers and software for full feature support, even if Apple has their own.
My Canon iPF5100 printer requires Canon's software and Photoshop driver. My old, but still pretty good Hp CLS8500N printer can use Apple's drivers, but Hp's are much better.
Same with my scanners, cameras etc.
Some of these drivers are very complex, and Apple won't be able to duplicate drivers for all the machines out there. That's the same problem MS had with Vista. Ad MS has far more people to do this than Apple does. Apple will need these companies to do the drivers for their own machines. What happens when a machine is discontinued? One can't expect companies to work on drivers for machines they haven't sold for some time, even if they are still in top notch condition operationally, and quality wise.
Yeah Shake is still Carbon. I don't think they are updating it to Cocoa. It looks like they are dismantling it to use with a dumbed down Cocoa app like Motion much to my annoyance. No word on any major improvements in well over 3 years now I think. One of the stupidest moves they've ever made IMO. Buying one of the most important pieces of software the film industry has seen and then killing it.
It would be like buying Renderman and using the code to do render the reflections on the Dock.
They could of course be rewriting these apps from the ground up but I'll reserve the right to hate them every day until they say it's the case.
I actually find it a bit hypocritical how they tell developers like Adobe to transition to Cocoa when they haven't done it themselves.
Well, I keep telling people who dump on MS and Adobe about this situation that Apple is in the same boat, but it doesn't wake them up.
I hope they aren't killing Shake. Hopefully they are re-writing it. They certainly must be working hard on FCS though. At least they should be, considering how it has the video industry in such a stranglehold these days.
I think it will need a change to programming conventions. If standard code could be parallelized easily, people would have done it ages ago. I'm not sure an API will do it but there may be guidelines as to how to program for parallel processing. It won't work for all code but hopefully there will be some automated compiling that does a better job than auto-vectorization on top that deals with threading.
That's right. What's even worse is programming for hyperthreading. If that isn't done right, the program actually slows down. If Apple wants to include that in their specs, they've got a lot of work cut out for them.
A 64 bit kernel will require 64 bit drivers. There's no way around that at all. If it will somehow run on a 32 bit machine, then that machine, and version of the OS can use a 32 bit kernel. If they can somehow do that, then for those machines only, 32 bit drivers will suffice.
Of course there are ways around that. Just because nobody does it (maybe it's even already done) doesn't mean it's impossible.
Regards
apeiros
It seems fairly certain that it will likely be in March at the earliest though and this is stretching Q1 to the limit so we'll just have to see how developments go over the next couple of months and maybe get some info at MW.
The urgency is for the apps and drivers. I've only recently considered moving to Leopard because I don't want my apps to start breaking and introducing bugs that weren't there before.
My view is if it ain't broke don't upgrade unless it offers something major. Snow Leopard should offer something major but I want it to be in the hands of all developers soon so that they can make sure the apps work properly.
Otherwise, I might not upgrade for another year from now. If it's out in March then I might be able to upgrade earlier.
Are you saying that you haven't made up your mind and are still using Tiger or something earlier?
I keep saying that I don't see this before the ADC. This is much more work than some are assuming.
It's almost a total re-write of much of the OS, as well as the added functionality at the lower levels.
This is a shrewd point.
Lemon Bon Bon.
Since there is no new features and it's pretty much a leopard bug fix, why rush it? The whole snow leopard thing is making leopard faster and smaller in file size. There should be nothing wrong with this release. So take your time Apple and make sure it's all 100%.
Correct in part. But it is not a "…a leopard bug fix,…"
Since 2001, Mac OS X has delivered more than a thousand innovative new features. With Snow Leopard, the next major version of the world’s most advanced operating system, Mac OS X changes more than its spots, it changes focus. Taking a break from adding new features, Snow Leopard — scheduled to ship in about a year — builds on Leopard’s enormous innovations by delivering a new generation of core software technologies that will streamline Mac OS X, enhance its performance, and set new standards for quality. Snow Leopard dramatically reduces the footprint of Mac OS X, making it even more efficient for users, and giving them back valuable hard drive space for their music and photos http://www.apple.com/macosx/snowleopard/
And the only official press release from Apple says in part:
SAN FRANCISCO—June 9, 2008—Apple® today previewed Mac OS® X Snow Leopard, which builds on the incredible success of OS X Leopard and is the next major version of the world’s most advanced operating system. Rather than focusing primarily on new features, Snow Leopard will enhance the performance of OS X, set a new standard for quality and lay the foundation for future OS X innovation. Snow Leopard is optimized for multi-core processors, taps into the vast computing power of graphic processing units (GPUs), enables breakthrough amounts of RAM and features a new, modern media platform with QuickTime® X. Snow Leopard includes out-of-the-box support for Microsoft Exchange 2007 and is scheduled to ship in about a year. http://www.apple.com/pr/library/2008...owleopard.html
I would suggest that any report indicating a specific introductory time is conjecture. As Apple has published on two sites, it "is scheduled to ship in about a year" (dated June 9, 2008). The question now would be, "How long is about a year?" Some would say that it could lie anywhere between 6 months plus a day to 18 months less a day. In any event, it is somewhere within that period, it depends to what you round time off to, and it is relative.
I just hope they don't release it until its ready. Too many times Apple has rushed out an OS. Who cares if its released this time next year. Its not like Tiger or Leopard are terrible OS's or something. Now is the time to take their time and get it 99% right the first time and not have to rush around and release 10.6.1 as quickly as possible to fix bugs that should have been caught before it was released.
1) You always have the option to not buy it until it has matured a bit and reports are showing that it's quite stable.
2) Hopefully Apple learned from the iPhone 3G/iPhone v2.0/App Store/MobileMe fiasco.
OpenCL was written by Apple and introduced to the Group at Khronos to adopt to, not the other way around which you imply as, ``I have this idea and can you do the heavy lifting for us to then write our APIs and put into the OS, thanks so much.''
Apple wrote OpenCL. They did the heavy lifting as a multi-year project. They had Nvidia, ATi and Intel in scope when they wrote it. They then brought in those bodies when it made sense to then work out the rest to ratify it as quickly as possible.
If it weren't for Apple doing the heavy lifting it would have stuck in limbo for a few years before finalization. That is what made it amazing to Nvidia, ATi and Intel. They didn't have to do much to have their stuff work with it.
Get it?
http://forums.macrumors.com/showpost...9&postcount=53
actually, OpenCL is mostly based on CUDA with some Apple people doing some nifty dynamic compilation on top.
CUDA has been ported to the CPU as well to NVIDIA GPUS (and ATIs are getting there). Basically they are going to be doing super-threading algorithms that can run on any streaming device, or on a normal device that is not as fast but does not break the programming model so it can scale across different cores in the system (from different vendors, but it is geared towards NVIDIA initially)
Although, I always get a kick about working in a project and getting corrected from an outside source :-) hint.. hint...
I don't know if this guy is blowing smoke or not but it is interesting that OpenCL seems to favor Nvidia GPU. I would have thought that equal support for AMD/ATI would have been developed concomittantly seing as how Apple's lineup always has a model with ATI graphics.
Of course there are ways around that. Just because nobody does it (maybe it's even already done) doesn't mean it's impossible.
Regards
apeiros
It's pretty much impossible to do properly. I suppose a kludge could be done, but that would cause too many problems of it's own, so no one does it.