Yes, this is what I've been hoping for. I hope it will be redesigned as well as written in cocoa.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
OpenCL 1.0 Specification Completed and Released
- Thread starter MacRumors
- Start date
- Sort by reaction score
Yes, this is what I've been hoping for. I hope it will be redesigned as well as written in cocoa.
This thing is definitively coming to the iPhone. Can't tell whether the SDK will allow openCL for us mere mortal iPhone devs though.
Let's have a look at the acknowledgements section from http://www.khronos.org/registry/cl/specs/opencl-1.0.29.pdf
Tally:
Apple 21
Intel 14
ARM 3
Broadcom 1
Freescale 1
Qualcomm 1
Texas Instruments 2
-- leftovers from the PC world --
AMD 12
Blizzard 1
Codeplay 2
NVIDIA 7
Electronic Arts 2
Ericsson 3
IBM 5
Let's have a look at the acknowledgements section from http://www.khronos.org/registry/cl/specs/opencl-1.0.29.pdf
Tally:
Apple 21
Intel 14
ARM 3
Broadcom 1
Freescale 1
Qualcomm 1
Texas Instruments 2
-- leftovers from the PC world --
AMD 12
Blizzard 1
Codeplay 2
NVIDIA 7
Electronic Arts 2
Ericsson 3
IBM 5
Gooooooooo wrapper for primitive GPU operations^W^W^W^WCUDA^Wrevolutionary Apple-divined OpenCL! Now the limited subset of embarrassingly GPU-parallel algorithms can be expressed in quasi-C!
Software engineering: because tools are better if they're all made to look like a familiar hammer. Meanwhile, real data-parallel languages are relegated to those whose control freakery doesn't marry them to imperative programming.
Software engineering: because tools are better if they're all made to look like a familiar hammer. Meanwhile, real data-parallel languages are relegated to those whose control freakery doesn't marry them to imperative programming.
We didn't say realtime operating system!
I wasn't making a reference to realtime operating systems which at this point I don't believe have a place on the desktop. What I'm talking about is apps that can respond instantly to user input while managing large data sets. If not instantly at least allow for interactive processing that isn't hobbled with extreme waits for data processing.
I suspect that you know the difference here but needed to highlight the vastly different uses of realtime. For many of the novice here the idea that a computer action may have a deadline is a bit strange.
Dave
I wasn't making a reference to realtime operating systems which at this point I don't believe have a place on the desktop. What I'm talking about is apps that can respond instantly to user input while managing large data sets. If not instantly at least allow for interactive processing that isn't hobbled with extreme waits for data processing.
I suspect that you know the difference here but needed to highlight the vastly different uses of realtime. For many of the novice here the idea that a computer action may have a deadline is a bit strange.
Dave
This are historic news, we should all be very proud of the company we love and support that they are once again in the forefront of innovation.
OpenCL and Grand Central are single handedly going make Snow Leopard the best OS ever. With M$ scratching their heads, still trying to figure out how to patch the Vista and sell it as Windows-7 they are going to loose the race. Linux and OSX will emerge as two heavyweight OSs in next few years. M$ is history.
I know there will be many of the opinion that most commercial / development houses still use Windows. But from the present trend and companies trying not to rely on flawed windows have already started to migrate to Linux and OSX. I work for a leading edge technology company and half of my companies computers are already using linux and osx for core development stuff.
For Apple to win, M$ does not have to loose..... Steve Jobs ====> coz M$ will eventually fall out anyways
Maybe in that imaginary world of yours that's what is happening. In the real world, there's absolutely no chance that a company is gonna spend $1500 in a computer to do something that can be done with a $800 one from Dell. Imagine if they have to buy 100 computers... Only very specialized companies (Video or sound editing, graphic design, etc) will go for Mac. And even that is being jeopardized by Apple's poor hardware decisions lately.
People could not car less about OpenCL and Grand Central. They want check their email and browse the internet. That's 85% of the people buying computers. Some of them also want to play videogames, but this guys usually customize their computers a lot, something you can't do with Macs.
People don't like change. If you are used to Vista, you don't wanna use OSX (let alone Linux) because it feels weird to adapt. And again, most people don't care if it takes advantage of the GPU or manages the RAM better, because most people don't even know what RAM is.
I've been to many graphic design companies as part of my course. I've seen a lot of Macs, more then windows system, which were mainly used by the secretary and managing part of the companies. I know these companies could care less about the new Apple led display, since I've seen almost none.
iPhone? Maybe not the current one!
See the thing is OpenCL can be used to leverage other hardware like DSP's and such.
One thing OpenCL demonstrates is that Apple is looking well out into the future here. They are showing sound leadership. Now my greatest wish is that they don't rush baking this cake, we have had to many burnt offerings from Apple lately. So lets hope that SL stays in the oven until done just right.
Dave
I'm not sure there are useful resources on the current iPhone to make use of OpenCL. Long term though it does look like they have intentions to exploit embedded devices. The question in my mind is this will OpenCL be using GPU's on future iPhones or more specialized hardware.
See the thing is OpenCL can be used to leverage other hardware like DSP's and such.
Nice fine!
You shouldn't be so hard on AMD, I could see Apple using their hardware in future devices. AMD would be an especially good fit in a Mini replacement.
One thing OpenCL demonstrates is that Apple is looking well out into the future here. They are showing sound leadership. Now my greatest wish is that they don't rush baking this cake, we have had to many burnt offerings from Apple lately. So lets hope that SL stays in the oven until done just right.
Dave
Ahh ye of little experience. I'll get you started here on your journey to understanding how a corporate works. Or any business for that matter.
1. Applying the way you as a consumer purchases a computer system or service or product should never ever be applied to a business entity. 2 very different ROI's. The initial cost is a very small part of the solution as a whole.
2. Development programs care very much about the frameworks and environments that are presented.
3. People liking or disliking change is irrelevant. Without change and a dynamic poise, a company is doomed to failure or a monotonous decline into obscurity.
OpenCL and Grand Central are single handedly going make Snow Leopard the best OS ever. With M$ scratching their heads, still trying to figure out how to patch the Vista and sell it as Windows-7 they are going to loose the race. Linux and OSX will emerge as two heavyweight OSs in next few years. M$ is history.
I know there will be many of the opinion that most commercial / development houses still use Windows. But from the present trend and companies trying not to rely on flawed windows have already started to migrate to Linux and OSX. I work for a leading edge technology company and half of my companies computers are already using linux and osx for core development stuff.
For Apple to win, M$ does not have to loose..... Steve Jobs ====> coz M$ will eventually fall out anyways
Agreed. Even in the CAD Industry the growing consensus is for Autodesk to move their applications to Linux and OS X, while EOL on XP and wait until Microsoft proves something with Microsoft 7.
Leftovers? Besides APPLE, the main designers are AMD, Intel and Nvidia. Obviously EA wants their gaming systems to leverage it, along side other game players. IBM will use it for advanced research and application services. Broadcom, TI, Freescale and Qualcomm will use it for all sorts of market spaces.
Intel, of all the people are behind the curve when it comes to GPGPU work and aren't targeting the heavy end work that AMD and Nvidia will continue to dominate.
AMD has turned the corner and their cards and more are going to give Nvidia a serious run for their money.
Copy-and-paste is, of course, a glaring omission on the iPhone. However, cut-and-paste in the Finder sounds like a potential nightmare. What if someone "cuts" an important application or file and is interrupted by a phone call or steps away from the computer, forgetting to "paste" it somewhere? Where does it go? It makes more sense to drag and drop. For the PC types, Apple has already implemented copy-and-paste for document and file icons (OS X 10.4).
On the matter of OpenCL, this sounds very promising! Cross-platform products would not have to write custom code to take advantage of whatever processing power is available (CPUs+GPUs).
That is a very old variant of Cut and paste. It no longer works like this in windows (and hasn't done for a long time). When you cut a file it becomes greyed out then if you decide not to paste it anywhere the file then ungreys and is perfectly normal.
I am interested in seeing how this directly relates to overall computer use. It is an interesting technology and no doubt will make advances in software quicker.
Yeah why don't you ask the guys over the microsoft's multibillion ad campains where they design their pictures for the ads, because all of them where found in the metadata to be from macs!!!
Then you will love 10.7 when they finally work the kinks out of the rewritten Finder from 10.6.No, a new and better Finder will make SL the best OS ever...
Its a task that needs to be done and moving to Cocoa will provide a foundation to address a number of issues but expecting the version 1.0 of any rewrite to be as smooth, stable, and polished as the mature version it replaces is asking for disappointment.
If the Finder is rewritten in SL then it will take advantage of all the 'comes for free' benefits of the true Cocoa frameworks but will ship with annoying bugs, features regressions, and UI inconstancies compared to the current version. While the user base screams bloody murder Apple will keep patching away at it and by the time SL starts to reach the end of its update cycle the new Finder will be close to as stable, functional, and UI complete as the one it replaced.
Much like OpenCL, it will take until the NEXT full OS revision before the new and more powerful foundation truly begins to be taken advantage of.
While I agree with most of what you say I look back in history and always see people wanting/needing more speed. Where the browser is today isn't where it will be tomorrow. Web 2.0+ is gonna require beefier hardware and hopefully Apple will be there waiting.
That said, has anybody seen any remarks about how Grand central and now OpenCL is affecting the speed in the latest Snow Leopard builds? I would think we'd start seeing something soon..
The only reason cut and paste is useful in Windows is because the Windows style of displaying files and the entire method of window management sucks. That's not true on Mac OS.
Love the hype machine....
Once again, the forum is full of over-hyped posts about how the "next thing" is going to be revolutionary...
...and when 10.6 ships the forum will be filled with the wailings of those disappointed by the reality.
___
Face it, most people won't get any significant benefit from OpenCL - because most systems don't have enough cores and GPU pipelines to exploit it, and because most applications won't be able to use it. (And on launch day, even those apps which could exploit it won't have been re-written yet....)
In the mid term, people using the maxi-tower with optional graphics cards for rendering will be quite happy. Eventually, when quad/octo core CPUs are common, and Larrabee embedded graphics is here, even the people with Apple consumer machines will see a boost for some apps.
So many people here are setting themselves up to be let down by 10.6.
And, by the way, if Apple is all about "innovation", why has it taken them so long to simply use the MPEG/H.264 hardware decoders that have been in the graphics cards all along?
Once again, the forum is full of over-hyped posts about how the "next thing" is going to be revolutionary...
...and when 10.6 ships the forum will be filled with the wailings of those disappointed by the reality.
___
Face it, most people won't get any significant benefit from OpenCL - because most systems don't have enough cores and GPU pipelines to exploit it, and because most applications won't be able to use it. (And on launch day, even those apps which could exploit it won't have been re-written yet....)
In the mid term, people using the maxi-tower with optional graphics cards for rendering will be quite happy. Eventually, when quad/octo core CPUs are common, and Larrabee embedded graphics is here, even the people with Apple consumer machines will see a boost for some apps.
So many people here are setting themselves up to be let down by 10.6.
And, by the way, if Apple is all about "innovation", why has it taken them so long to simply use the MPEG/H.264 hardware decoders that have been in the graphics cards all along?
These kinds of articles always make me wonder how efficient operating systems/software/even programming languages are these days. When many of these languages & basic operating systems (C, C++, Unix) were originally developed before multi-core, gpus, etc. were created, much less available in personal computers.
Well they were by far the first to use h.264 endoding a standard in both mp3 players and computers, and h.264 decoding in cards is fairly recent.
So, Apple using AAC in an Igadget defends their lack of innovation in OSX?
Meanwhile, all your MPEG-2 DVDs that you have now have been stuttering and skipping....
And if people think Redmond is asleep on the parallel computing front, you should check outApple's main advantage in parallel computing may simply be their "catchy naming department", as Microsoft embeds parallel computing in .NET itself....
