So exactly how much open source do you have to have to qualify as open source? As ridiculous as that sounds, thats kind of how you are trying to compare it too. Hint: Its more then zero.
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.
No more nvidia chips in Apple's computer line
- Thread starter netkas
- Start date
- Sort by reaction score
So exactly how much open source do you have to have to qualify as open source? As ridiculous as that sounds, thats kind of how you are trying to compare it too. Hint: Its more then zero.
There are more than a few complete programs/applications.
http://www.opensource.apple.com/
However, if talking about "Mac OS X" with an emphasis on the "Macintosh" aspect that is what normal users directly interact with day-to-day, then no. Both in that it technically not being an Operating System in the strict sense and most normal users don't know the difference ("The Finder isn't just an app, it is the heart of the OS". ). Similar to how X windows isn't Linux. Neither is the Android layer.
The stuff that makes a Mac really a Mac is closed. With a finite amount of work Apple could layer all of the Mac stuff on top of FreeBSD or Linux if really wanted to. At one point Nextstep was being layered on top of Solaris. It would be painful (to driver and low level apps), but possible. It wouldn't buy much for the pain though.
I didn't know that Apple is in the business of doing PR for Nvidia.
And, I don't see that Metal is doing more than supplement OpenCL.
Last edited:
That sounds like an beast on paper. I am thinking of an 6700K, 32gb ram and an GTX980ti. I think that will deliver a lot of muscle to Photoshop
Actually, it's pretty much 100%. If somew of your code isn't open source then you can't call it open source.
Some of Mac OSX is open sourced. No, its not a complete operating system. But there are several operating systems based on Darwin.
Live, moderately active ones? Stuff like PureDarwin and SEDarwin seem largely comatose.
There is open source as a technical category and then there is open source religion. For example, the Free Software Foundation folks rail against Linux because Linus and others allow closed source kernel drivers. Kind of like Christian versus Catholic , Orthodox Catholic , Calvinist , Lutherans , Episcopalians , etc .... Not worth the drama of trying to get everyone under one single definition. There isn't one.
You guys can continue cuda open source discussion in separate thread or in PM
I stopped but the poster that said that is tend to be nonsensical and not exactly popular so one thing lead to another. I think we should let it drift so folks can get stuff off their chests.
While Darwin being open source maybe off topic, not so with the talk of open sourcing CUDA. We are also theorizing on the reason why Apple has been using more AMD and Intel rather then Nvidia. Since Nvidia uses more closed CUDA rather then a more openCL might be one of the reasons why. Since it is in the realm of possibility on a Apple rumor site makes it a valid topic of discussion on this thread about Nvidia.
If I see something that I think is wrong or inaccurate I will say so. If it does not fit with popular opinion, so be it. Yes, lets move on instead of attacking posters.
I believe that poster is both wrong and inaccurate and that doesn't fit with popular or at least semi-popular opinion, if you view that as an attack so be it.
So I can take that tack to.
what.. you don't think cuda being proprietary has anything to do with:
No more nvidia chips in Apple's computer line
?
If people here really objected to proprietary lock-in - they wouldn't be here. They'd have moved to Windows (open hardware, open apps, proprietary OS) or Linux (free-for-all).
ps: Note that "free-for-all" doesn't mean "no charge for anyone" - it means "no payment, no support".
my point was more @OP saying the conversation is off topic when it's actually very on topic.
Here's an idea: Instead of trying to port CUDA to AMD gear, how about people port their apps to OpenCL?
Seriously, waiting for better CUDA support is a giant waste of time. OpenCL, Metal, and Vulkan are all viable ways forward. Why bother trying to save CUDA? It's not even current compared to Vulkan. Time to move on.
I don't know why everyone is going back and forth on trying to save some aging technology instead of just getting apps that are actually written on top of real, current standards.
Seriously, waiting for better CUDA support is a giant waste of time. OpenCL, Metal, and Vulkan are all viable ways forward. Why bother trying to save CUDA? It's not even current compared to Vulkan. Time to move on.
I don't know why everyone is going back and forth on trying to save some aging technology instead of just getting apps that are actually written on top of real, current standards.
fwiw.. from what i can gather via a handful of (applicable) applications in my circle.. that is what's happening.
developers who were using cuda before are moving on to other standards that will moreOrLess work on any gpu/cpu.
Right, that's why I don't get this discussion. Everyone I know of is dropping CUDA support. Why even go through the effort of open sourcing it or porting it to AMD at this point? Even Nvidia is adopting Vulkan and Metal. You'd have to be crazy to do an app in CUDA at this point, both in theory and in practice.
So why bother with worrying about CUDA's future?
Next people are going to start complaining that the Mac Pro doesn't support Glide because "I don't have any PCI-E slots to stick my 3DFX card in!"
100%
Calling something open-source implies that it's completely open source, just as calling something organic implies that all of the ingredients are organic, not just ⅓ of them.
Fortunately, the issue of the PCIe slot can be solved by Thunderbolt PCIe enclosures, at least for low- and mid-range cards.