Why is nVidia being a dick and not dropping CUDA and supporting Vulkan?

Why is nVidia being a dick and not dropping CUDA and supporting Vulkan?

It seems like it would benefit the whole industry to move to one common graphics API.

Other urls found in this thread:

pcper.com/reviews/General-Tech/Breaking-OpenCL-Merging-Roadmap-Vulkan
home.otoy.com/render/octane-render/
hexus.net/tech/news/software/105895-vulkan-opencl-will-merge-single-api/
twitter.com/SFWRedditVideos

Nvidia has excellent support for Vulkan. It would not make sense to drop CUDA because Vulkan is not a replacement for CUDA. CUDA is not a graphics API.

>benefit the whole industry to move to one common graphics API


How does that help nvidia?

CUDA is not a graphics api
What's up with a those illiterates on Sup Forums

gaymers

>Vulkan is not a replacement for CUDA.
pcper.com/reviews/General-Tech/Breaking-OpenCL-Merging-Roadmap-Vulkan
because muh proprietary apis

No I mean drop CUDA in favor of Vulkan, which will also include OpenCL.

It is a known fact that certain applications that support CUDA are CUDA only and don't support OpenCL (see OctaneRenderer)

home.otoy.com/render/octane-render/

Now until nVidia drops CUDA supports and joins Vulkan, this will not change.

By having a standard?

No shit, it's for compute just like OpenCL (which is being merged into Vulkan).

>By having a standard?

How does having a standard help Nvidia?

AMD passed off Mantle to help the industry. nVidia is a kike chink company.

If Vulkan is superior, why would Nvidia need to drop support for it? People will start using it of their own accord, and CUDA will become a legacy API for backwards compatibility with old software.

because nvidia loves closed ecosystem

just look how they are going with g sync they started with 60% of the market now its on 20% and they still dont wanna give up

Because CUDA and Vulkan aren't exactly the same thing?

Then why don't you say "drop CUDA in favor of OpenCL"?
Are you retarded?

Same reason why nvidia doesn't stop supporting directx. They're fucking garbage company.

Because CUDA is the industry standard with huge amounts of professional software built on top of it?

do you even follow news?
hexus.net/tech/news/software/105895-vulkan-opencl-will-merge-single-api/

Are you mad that amd will never surpass Nvidia in the gpu market again?

But did they merged it already in single API?
Or are you talking from the future?

Nope, nvidia is just cancerous company.

Not my fault your reading comprehensions are subhuman level.

Why the fuck would I mention CUDA if I wasn't talking about it's equivalent OpenCL?

Are you this much of an autist to not realize that?

Open CL and Vulkan are completely different things, go back to

OpenCL is fully supported by Nvidia, along with Cuda and Vulkan. The question is why haven't AMD paid the licence for Cuda support? Because they don't care about none gaming is why.

Hey idiot, OpenCL is being merged into Vulkan.

OpenCL & CUDA are compute APIs.

->

If they use Vulcan, it would run on the competitor's GPU and help them and reduce their market share. No company is trying to help the "industry", they're trying to help themselves and their shareholders.

Open CL and CUDA are completely different things, go back to

It's a known fact AMD has better OpenCL support than nVidia.

nVidia has subpar OpenCL support on purpose to force applications to use CUDA (ie see Davinci Resolve). It's literally being an asshole for this reason.

I don't think nVidia will ever license out CUDA.

The industry is moving to open standards, and nVidia is being a dick just like Microshit with DX12 (although DX12 is for a different purpose). But Vulkan supports both DX12 low level Graphics API and Compute (OpenCL replacement).

They are in the process.

They are literally similar you fucking sperg. Kys

Because CUDA is better.
Just like DX12 is better than Vulkan

AMD gives away their stuff all the time, ie Mantle and even uses open source stuff like FreeSync.

nVidia is being a dick with CUDA, Hairworks, etc.etc.

Also, directx is alive because of nvidia supporting it (by developing with M$ together).

Fucking and Sperg are completely different things, go back to

>DX12 is better than Vulkan

Doom 2016 says otherwise.

And Nvidia has better CUDA support.

CUDA is far superior to Vulkan or OpenCL garbage

In the process doesn't mean it's already here you fucking retard.

No because Microshit has a monopoly on gaymes.

Vulkan is a fine API.

Take a shower, neckbeard.

That's because they lost control over the market, they can't make it proprietary even if they want to.
By giving away free stuff they can destabilize Nvidia's grip on the market. Mantle has lead to DX12 and the shift away from DX11, which shift the market away from AMD's shitty DX11 drivers and benefited itself.

>Game with shitty graphics and closed environment says otherwise
kek

You're dumb.

You probably never used a CUDA enabled application in your life, faggot.

A few months away and OpenCL will dissolve into Vulkan. Faggot.

If Linux is superior, why would Intel need to drop support for it? People will start using it of their own accord, and Windows will become a legacy OS for backwards compatibility with old software.

OpenCL is still garbage

CUDA will forever be superior in ease of use and getting results faster

You don't even know what CUDA is and/or never used an application capable of using CUDA.

Now kindly fuck off. Let the grown men talk.

But is it already merged?
Where do you live fucking morron? In timezone where it's already autumn?
Fucking imbeciles can't even communicate.
Go to fucking you morron.

because cuda is a paraller api and nvidia cards compared to amd when it comes to multithreading is like comparing a boar to a bear nvidia will lose insane amount of money if they actually do this

Isn't CUDA just a whole lot easier to work with compared to OpenCL, hence greater adoption? Or something along those lines.

So what? Did they forced nvidia to help developing directx?

Because close standards and software are always better.
Who the fuck would want to use something develop by some weoboo furry neckbeard

AMD confirmed for Forerunners.

Yes and it's faster and easier to develop in.

You're the moron that has never used CUDA or programmed in CUDA and it shows

It's obvious your cunt whore mother never taught you to not butt in when the adults are talking

>its faster
WTF DOES THAT EVEN MEAN did you ever coded a program that run fast on cuda and you made a opencl program that is slow on opencl? LOL
>its easy
no its not east cuda is what steam is it was the first op lang to exist that make use of gpus that DOESNT make it easier the HIP tools amd released can literally translate all the code code into opencl and requires minimal human interraction for it this is what amd will start pushing when the FE cards are out and given how amd cards naturally are workhorse monsters this is only going to get better with theit also new HSA platform..

And yet look at each company's market share and profits. Nvidia aren't a charity or a humanitarian organisation. They're a business that exists to make as much money as possible, and they do it very well. Pushing their proprietary ecosystem and technologies is one of the primary ways they accomplish this.

I just mentioned Davinci Resolve and OctaneRenderer with Cinema 4D which both use CUDA you fucking idiot.

I'm not a pajeet programmer, but I've been using CUDA enabled applications for years.

Now go kys.

Not the same things

>They are literally similar
They are not. OpenCL is too generic, while CUDA is specialized and reflect the hardware.

Microsoft and gaymers deserve each other. Apple and gays deserve each other. It keeps subhumans outside Linux and BSD.

I use Windows on my desktop at home (because gaymes), macOS on my laptop (because battery and comfiness), Linux at work (because embedded development) and FreeBSD on my file server at home (because ZFS).

Come at me bro.

two different things

Currently owning a nvidia GPU. Here's to hoping that amd delivers and releases a framework which allows neural network training of current frameworks with their GPUs.

>nVidia being a dick
What's new?

Vulkan and OpenCL begin backend in SPIR-V looks like asm for GPU, news you get about Vulkan merge OpenCL, is about let Vulkan make a lot computer and inter operability with OpenCL things but OpenCL still begin computer language.

About CUDA, CUDA is used in scientific and machine learning camps, a lot of software begin make in it, and Nvidia could add new feature for each new GPU gen.