Apple to release first ARM Mac without Intel processor in next 18 months

[UnknownSouljer]

Ahh, because creating and then abandoning OpenCL is just too much work for Apple, RIGHT?

Which is why they did abandon OpenCL in-favor of Metal? A to-the-metal proprietary API that they created as a competitor to NVIDIA's CUDA.. eight year after Cuda launched.

Native Metal uptake hasn't been that huge so if Apple hadn't gotten a bug-up-their-ass (as is their tendency,) they could have supported Vulkan instead.

A number of games / applications are just running a Vulkan ->Metal wrapper through, so why not make it official?

Fuck Apple, and their shitty choices for APIs; they have a long history of choosing the wrong ones, and then taking forever to undo their mistakes. But the "Apple does no wrong" True Believers™ think everything they do is Gospel.

Like you Metal Cheerleaders, who suddenly think that %100 of their apps all run it natively, and Apple hasn't totally abandoned a proprietary 3D API suddenly before

And the worst part of the Metal transition is, Apple is abandoning BOTH OpenGL and OpenCL in favor of their One True API (this time for sure!)

They also don't support Vulkan on Apple, so the world is working around them.

Whenever Apple has had more money than brains, they have always gone back to "Not Invented Here." And their users just pretend it's not happening again. Sometimes it's for the greater good (Apple custom ARM cores), but most of the time it's not.

Cool. Just don’t buy a Mac. Problem solved.

 

[Deleted member 243478]

Ahh, because creating and then abandoning OpenCL is just too much work for Apple, RIGHT?

Which is why they did abandon OpenCL in-favor of Metal? A to-the-metal proprietary API that they created as a competitor to NVIDIA's CUDA.. eight year after Cuda launched.

Native Metal uptake hasn't been that huge so if Apple hadn't gotten a bug-up-their-ass (as is their tendency,) they could have supported Vulkan instead.

A number of games / applications are just running a Vulkan ->Metal wrapper through, so why not make it official?

Fuck Apple, and their shitty choices for APIs; they have a long history of choosing the wrong ones, and then taking forever to undo their mistakes. But the "Apple does no wrong" True Believers™ think everything they do is Gospel.

Like you Metal Cheerleaders, who suddenly think that %100 of their apps all run it natively, and Apple hasn't totally abandoned a proprietary 3D API suddenly before

And the worst part of the Metal transition is, Apple is abandoning BOTH OpenGL and OpenCL in favor of their One True API (this time for sure!)

They also don't support Vulkan on Apple, so the world is working around them.

Whenever Apple has had more money than brains, they have always gone back to "Not Invented Here." And their users just pretend it's not happening again. Sometimes it's for the greater good (Apple custom ARM cores), but most of the time it's not.

That is because their mobile platform was running on Metal for ages so they aligned their desktop to use it also so developers can write apps for both platforms easily with no learning curve.

When will people learn that multiple standards is bad, developers will all do their own thing leading to wildly inefficient software and hardware fragmentation like Windows is plagued with across rendering platforms.

At least with Apple, all software can leverage their hardware equally with no need to use shitty inefficient processing on a CPU or CUDA which will only run on 1 vendor and 1 or 2 platforms only.

The battle between CUDA, OpenCL has been utter BS with no one budging so Apple stepped in and said you want to use our platform, use our API so users get a good experience regardless of application or hardware.

 

[IdiotInCharge]

By 2 years. But more to the point, they chose to develop a closed standard rather than an open one. AMD creates Vulkan and gives it away. nVidia creates Cuda and tries to force not everyone just to use the codebase, but also to force everyone to use their hardware.
Apple developed OpenCL (I guess one of those "walled garden things full of other people's inventions" as you would say) and gives it away. You can think whatever you want. But Cuda hasn't offered an advantage over alternatives since near after its inception, except that it plays nice with the current leader in graphics cards.

Being an optimized API for the fastest hardware, with the full support of Nvidia's development resources, is one hell of an advantage. And with respect to creating an 'open' standard, you'll need to present more than a utopian ideology to show that Nvidia was being unreasonable by creating an API for their hardware, which was necessary to use their hardware, without other hardware manufacturers in mind.

Further, with respect to AMD, their only chance at any traction for what they create with their market position is to give it away and hope for the best. Apple the same, if they want it used outside of their ecosystem.

Oh, so you're not aware. Hairworks was intentionally designed to run poorly on AMD hardware.

This means that AMD GPUs were designed to run DX poorly, given that Hairworks (and all of Nvidia's Gameworks middleware) is written for DX.

It was then leveraged by nVidia with their status to get as many devs to use it in order to create more titles that would give nVidia a speed advantage. Hairworks by definition is the epitomy of anticompetitive. Actually all of nVidia's "works" products are.

Again, they're all written in DX, which their competitors have full access to. You're complaining about a company profiting from their work.

Apple has invented quite a few things and still does. But people's perception of what they do (and whether they like them or not) directly changes the perception of them. And because they are as big as they are, they're almost immediately polarizing. Your statements to me are a direct reflection of that.
It's okay if the company you like does it, but not okay if the company you don't also does. And that can be explained by whatever you deem to be 'relevant'.

That may be what you read from what I said, but it's not reflective of what I said. I have quite a bit of respect for what Apple produces; not enough to jump into their ecosystem as it stands, but that's more a function of personal preference.

On the discussion point at hand, Apple and Nvidia aren't really comparable except that they're both doing what leading technology companies do. Nvidia's gardens are so minimally 'walled' in comparison that they're simply a product of... actually releasing their products. GPU hardware isn't an exception in needing an ecosystem.

Also, don't get me wrong here -- the main point is that technology leaders generally don't plan a way ahead for their competition when they're trying to get their innovations, their hard work, to market.

 

[Mazzspeed]

That is because their mobile platform was running on Metal for ages so they aligned their desktop to use it also so developers can write apps for both platforms easily with no learning curve.

When will people learn that multiple standards is bad, developers will all do their own thing leading to wildly inefficient software and hardware fragmentation like Windows is plagued with across rendering platforms.

At least with Apple, all software can leverage their hardware equally with no need to use shitty inefficient processing on a CPU or CUDA which will only run on 1 vendor and 1 or 2 platforms only.

The battle between CUDA, OpenCL has been utter BS with no one budging so Apple stepped in and said you want to use our platform, use our API so users get a good experience regardless of application or hardware.

Standards are bad...

...So lets make another standard and make it proprietary as possible.

C'mon, you're going to have to do better than that. Fact is: We don't need another proprietary standard.

 

[Deleted member 243478]

Standards are bad...

...So lets make another standard and make it proprietary as possible.

C'mon, you're going to have to do better than that. Fact is: We don't need another proprietary standard.

Yes a standard with a huge market share across mobile and desktop platforms. I love the double standards with how AMD and NVIDIA are entitled to make their own APIs just for their GPUs but Apple aren’t allowed to make one to support their entire ecosystem with open support for any hardware, from any manufacturer.

OpenCL is old and CUDA is just plain shit. Neither can do the job of Metal. You also seem to forget it is one standard across all Apple products, which significantly reduces complexity and improves performance.

There is a reason why PCs are split between CUDA, OpenCL and CPU encoding, because no one has stepped in and put a stop to everyone doing their own thing and lazy developers not wanting to rewrite their apps. No one should be relying on a bloody CPU to do encoding in 2020.

 

[Mazzspeed]

Yes a standard with a huge market share across mobile and desktop platforms. I love the double standards with how AMD and NVIDIA are entitled to make their own APIs just for their GPUs but Apple aren’t allowed to make one to support their entire ecosystem with open support for any hardware, from any manufacturer.

OpenCL is old and CUDA is just plain shit. Neither can do the job of Metal. You also seem to forget it is one standard across all Apple products, which significantly reduces complexity and improves performance.

There is a reason why PCs are split between CUDA, OpenCL and CPU encoding, because no one has stepped in and put a stop to everyone doing their own thing and lazy developers not wanting to rewrite their apps. No one should be relying on a bloody CPU to do encoding in 2020.

A standard with a comparatively minuscule share across Apple desktop and mobile platforms. I applaud Nvidia for telling them to stick it and there are still scenarios where CPU's are more suited to the task at hand, where they are not we already have API's in place and they work just as well as Metal on more platforms.

This is an attempt at a vendor lock in, nothing more. Developers aren't going to suddenly show a spark of enthusiasm over the addition of yet another proprietary API.

 

[UnknownSouljer]

This is an attempt at a vendor lock in, nothing more. Developers aren't going to suddenly show a spark of enthusiasm over the addition of yet another proprietary API.

I'm glad to know you speak for all devs. In either case, this affects you not at all. Thank your for voicing your opinion about a platform you aren't on and will never be on.

 

[Deleted member 243478]

A standard with a comparatively minuscule share across Apple desktop and mobile platforms. I applaud Nvidia for telling them to stick it and there are still scenarios where CPU's are more suited to the task at hand, where they are not we already have API's in place and they work just as well as Metal on more platforms.

This is an attempt at a vendor lock in, nothing more. Developers aren't going to suddenly show a spark of enthusiasm over the addition of yet another proprietary API.

It doesn’t lock anyone in, any manufacturer can write for the Metal API, Apple even provide tools for easy transition and doesn’t impact other APIs. NVIDIA are the ones trying to lock people out.

and yes it is a massive marketshare. iphone sales are a huge market alone.

Apple are doing fine without NVIDIA anyway and don’t need to put up with their shit.

 

[Mazzspeed]

It doesn’t lock anyone in, any manufacturer can write for the Metal API, Apple even provide tools for easy transition and doesn’t impact other APIs. NVIDIA are the ones trying to lock people out.

and yes it is a massive marketshare. iphone sales are a huge market alone.

Apple are doing fine without NVIDIA anyway and don’t need to put up with their shit.

Apple are the only ones supporting their API. By definition, that's a vendor lock in.

Furthermore, the loss of Nvidia is a big kick in the guts for a great many Nvidia based Apple users and they're not one bit happy about it. You are drowning in the Apple Kool Aid.

 

[Mazzspeed]

I'm glad to know you speak for all devs. In either case, this affects you not at all. Thank your for voicing your opinion about a platform you aren't on and will never be on.

You're 100% right. I will never support a vendor lock in. Apple as a company can kiss my ass.

 

[Deleted member 243478]

Apple are the only ones supporting their API. By definition, that's a vendor lock in.

Furthermore, the loss of Nvidia is a big kick in the guts for a great many Nvidia based Apple users and they're not one bit happy about it. You are drowning in the Apple Kool Aid.

You need to read the definition of what a “lock in” is. Any vendor is welcome to support Metal with no affect to their hardware or other APIs, Apple doesn’t have a gun to their heads and NVIDIA chose not to adopt it so they are not a supplier anymore. Everyone else was more than happy to meet Apples needs so they get the business instead.

To be honest you are the typical PC user who has NFI what you are talking about and think everything is easily interchangeable. The Metal API was developed to make it easy for developers to code for iOS, OSX and tvOS. Vulcan, OpenCL and CUDA are either too generic or too proprietary for them to be even remotely suitable for such a task.

PC is the perfect example of what happens if you let everyone do their own thing without unifying the platform. You end up in the wild west with morons doing video rendering with a generic CPU instruction set, OpenCL or CUDA. Which is why Adobe Premiere is such a POS.

Rather than using one standard and optimising it to within an inch of its life.

 

[pendragon1]

so lock into apple's metal and youre gtg.

 

[Mazzspeed]

You need to read the definition of what a “lock in” is. Any vendor is welcome to support Metal with no affect to their hardware or other APIs, Apple doesn’t have a gun to their heads and NVIDIA chose not to adopt it so they are not a supplier anymore. Everyone else was more than happy to meet Apples needs so they get the business instead.

To be honest you are the typical PC user who has NFI what you are talking about and think everything is easily interchangeable. The Metal API was developed to make it easy for developers to code for iOS, OSX and tvOS. Vulcan, OpenCL and CUDA are either too generic or too proprietary for them to be even remotely suitable for such a task.

PC is the perfect example of what happens if you let everyone do their own thing without unifying the platform. You end up in the wild west with morons doing video rendering with a generic CPU instruction set, OpenCL or CUDA. Which is why Adobe Premiere is such a POS.

Rather than using one standard and optimising it to within an inch of its life.

No other vendor needs Metal as we have API's currently that work just fine. TBH, you're the typical Apple user that's sucked too much Apple teet. Metal is just as proprietary as CUDA, what we need is more open solutions and the last thing we need is a vendor stalling OGL support and blatantly refusing to support Vulkan natively just so they can force their own proprietary API that's just fragmenting the situation even further.

Adobe Premier is a POS because as a company Adobe suck balls about as much as Apple suck balls.