Vulcan

what good will ever come of this?
name one thing vulcan can do that openGL can't

Attached: khronosgay.png (1165x652, 392K)

Other urls found in this thread:

graphicsfuzz.com/benchmark/android-v1.html
en.wikipedia.org/wiki/Vulkan_(API)#Video_games
blogs.nvidia.com/blog/2016/02/16/vulkan-graphics-api/
twitter.com/SFWRedditGifs

>opengayl
back to the nursing home gramps

>name one thing vulcan can do that openGL can't
tp be fast and easy

This, and async

Receive precompiled shaders
Work on more than a single thread
Not require bloated vendor drivers with individual tweaks for every new game

Require 1k loc for the simplest rendering task

>have to ship shader source
w-why is this bad?

OpenGL is fine. If it's not broken don't fix it.

Attached: 1542891266129.jpg (771x723, 60K)

OpenGL is fine in most cases. There's a reason why they aren't abandoning OpenGL with Vulkan out there now.

Not an argument

that's just technical inertia, it has nothing to do with how good or bad a technology is
t. works at a company still using VSS for some projects

Actually it's severely broken on all operating systems. DirectX as well. GPU drivers are simply garbage.
graphicsfuzz.com/benchmark/android-v1.html
Vulkan is a solution to this problem.

I mean it does though. Unless you're getting into something that really needs raw power OpenGL works really well. Especially for prototyping. Even Khronos says that Vulkan isn't for everyone and they are right which is why they will continue to support and develop both.

that has nothing to do with rate of adoption

What I want to know is why it's been so slow to develop for, it's been almost 3 years since Doom came out and developers are still pushing this direct X11 60fps targeted bullshit

Attached: 1380479625116.gif (499x235, 766K)

because winbabbies and microshaft. also most games devs are either pajeets or code like pajeets

there are plenty of games using it
en.wikipedia.org/wiki/Vulkan_(API)#Video_games
big companies can't be assed to implement vulkan support for their game engines, though. maybe with next gen consoles. who knows

>there are plenty of games using it
>a list of incomplete games that claim support is "planned"

nice

>he needs to settle on a singular universal choice because his mind can't comprehend the pros and cons of different solutions
different tools for different purposes nibba
also what everyone in this thread just spoon-fed you

OpenGL act as a rendering server and only has one thread to accept commands, while Vulkan doesn't suffer from such a bottleneck.

>want to make a triangle with OpenGl
>30 lines of code

>want to make a triangle with Vulkan
>1000 lines of code

Attached: 1508343767495.jpg (800x807, 340K)

>not implementing everything in software

Hardware companies shouldn't be allowed to write software.
See the amd drivers that were opensourced

>name one thing vulcan can do that openGL can't
Multithread drawcalls
Trace rays
parallel computing
GPGPU compute capabilities
Have real time ray tracing API

But that's a faster triangle

It gives me terrible performance with mpv on linux, it runs good on windows.

Imagine judging a language solely on how long the hello world in it is.

command lists are controlled by the application (user-land) instead of the driver.

opengl supports spirv shaders nowadays but needs an extension.

how do I know I can use Vulkan? will my 10yo toaster use it?

okay 1 fcking thing
RPCS3

When are we gassing all /v/ niggers on this board?

Kek

AZDO is pretty much the same as vk
> throwing around buzzwords without understanding what they mean
ogl can do async
>Receive precompiled shaders
wrong
>Work on more than a single thread
also wrong, depending on what you mean
>Not require bloated vendor drivers with individual tweaks for every new game
vk drivers are still bloated, but less than ogl
vk gives you effectively nothing over ogl

Make running games through wine faster than native OpenGL ports.

>>vk gives you effectively nothing over ogl
>be persona 5
>on rpcs3
>core i9-9900K
>gtx 1080ti
>opengl shits itself
>vulkan solid 30fps
>dx12 crush
>>(((vk gives you effectively nothing over ogl)))
fucking Kek

>implying this is entirely the fault of the framework and not just the way they’ve implemented it

opengl is slow to handle shit
even on mobile systems
ppsspp doesnt lag at all on my S7 While opengl shits itself

Try writing that 1000 line triangle. You'll realise how much control you actually have over everything and the potential to optimise it significantly more.

how do I know if my ancient laptop supports vulcan?

look up the chipset

amd radeon r5

well I guess you can play freecell on ultra

k so google that
jesus christ this board is dead

hello gramps

It's weird seeing anons seriously shill for shit on Jow Forums

the actual code for a triangle is basically the same.
its the up front initialization that you have to do just to get the gfx card to do anything is where all those extra lines comes from.
You're telling the gfx card what to do instead of drivers dictating 60% of everything.

It's more power.
reduces cpu overhead by offloading shit onto the gpu which is where it should be.

most of it is memory management done by the programmer instead of the driver.

AMD Radeon (TM) R7 M260
AMD Radeon(TM) R5 Graphics
why is it listing two cards?
t. gramps

Dom is a good example.
my system
fx8350
8gig ddr3 2133mhz
gtx 970

Doom standard renderer on high settings: 60-70 fps
Doom with vulkan canb run ultra setting at 90-100 fps

probably one is a discrete GPU for performance and the other is for light loads

why do they do that shit?
why not just have a single integrated or a dedicatd / discrete gpu

High-performance GPUs are huge and power hungry, pairing them with a smaller GPU (often optimized for lower power) yields more power savings than underclocking/undervolting the larger one

This. Consider as a counterpoint, the amount of bullshit hacky code you would have to write to fight against all of the internal blackbox weirdness that opengl does to you.

>easy
The state of nu-g

r7 260M should have vulkan support

>that's just technical inertia, it has nothing to do with how good or bad a technology is
technology has to be good enough to overcome technical inertia
Vulkan doesn't offer enough over OpenGL to make it worth switching

>>Work on more than a single thread
>also wrong, depending on what you mean

With Vulkan, you can construct multiple command buffers from multiple command pools and fill them with commands asynchronously before submitting them to a graphics (or compute) queue yourself, you retarded faggot.

Here's a reason to use vulkan. AMD's opengl implementation is still fucking broken

t. Someone who has never used OpenGL nor Vulkan,

I work with OpenGL, give me a good reason to convert all my code to Vulkan

Performance and less driver variability. Depending on your requirements, it may or may not be worth it.
But your claim that it's not worth switching certainly doesn't apply to everybody.

OpenGL is going to die a long and slow death, and you can already see it with the vastly reduced number of extensions and increased time between major versions compared to pre- and post-vulkan, The only "significant" OpenGL revision since then has been OpenGL 4.6, which is basically just Vulkan compatibility.

work with steam play
i've been playing skyrim all night on my linux machine.

>tp be fast and easy
It's fast OR easy, you don't get both. It's not an inaccurate way of describing the choice between Vulkan and OpenGL. It's kind of like assembler and Java, your assembler program will take a magnitude longer to write and run many times faster.

Your code will take longer to write and run faster if you switch to Vulkan. You already know this if you're actually writing OpenGL code regularly. Does this mean I can give you a reason to convert existing code to Vulkan? Hell no. You already wrote that code and it probably works. You shouldn't unless someone offers you a large amount of money to do it. And I can't even recommend writing new code in Vulkan if you're very familiar with and used to OpenGL. It depends on you, obviously, but if I client wants that fancy new feature by a deadline and you can do it in OpenGL using half the allotted time or miss the deadline by using Vulkan then the choice should be obvious.

>But your claim that it's not worth switching certainly doesn't apply to everybody.
It applies to most people right now, Vulkan performance gains are pretty questionable, rewriting all your rendering code isn't a trivial task

When it comes to gaymes, most people aren't writing OpenGL either, and most shit is done with an off-the-shelf engine. Most major engines have Vulkan support, because making all of the games using them faster is important.
>rewriting all your rendering code isn't a trivial task
Nobody claimed that it was. Nobody is claiming that you have to switch now either.
From what I've seen, it seems that Vulkan has a lot of momentum, and will probably become the prevailing API in the future. Replacing extremely well-established APIs takes fucking ages to do, though.
I don't like the attitude that nothing is ever allowed to change because new things are scary.

Vulkan doesn't have alot of momentum, I'm not saying it's bad or anything it's just not good enough to motivate people to switch

There is no definitive way to say which one of us it right, so there is no point arguing this.
It's unquestionable that OpenGL is already stagnating and Khronos members have basically lost interest in it.

Vulkan API reduces overhead and exposes the hardware more readily to developers. OGL is ancient and archaic.

Attached: idiot_clap.png (488x463, 28K)

HURR NO MOMENTUM - just a number of AAA titles, industry leaders, and hardware manufacturers are supporting it. LITERALLY NO ONE!

blogs.nvidia.com/blog/2016/02/16/vulkan-graphics-api/

Attached: welp.gif (320x240, 1.58M)

>literally posting a press release
don't you have a graphics card thread to be in

Don't you have a Depends diaper to soil, Luddite?

The fact that you think I'm cheerleading for OpenGL over Vulkan only makes you look like a retard. I only said it wasn't worth switching to because it doesn't offer enough to make it worth the transition. That may just mean slower adoption, assuming Vulkan doesn't have anything terribly wrong with it

feed the graphics card from multiple threads without making you wish you were dead.

opencl is being killed and its functionality is being ported to vulkan. Better start to like it OP because it isn't going anywhere.

You mean you don't know how to use opengl as a surrogate opencl framework?

Then you should be able to outdo them easily.
Oh, hang on...

Attached: 1544753959154.jpg (570x587, 43K)

oh no game devs actually have to ship software. can't wait for another embarrassing hacker talk from an elite coder talking about how "the tech just isn't there yet" as the stumble through an embarrassing slidedeck of the nothing burger they have been writing in their basement the past 10 years.

Vulkan is capable of being far more efficient in its use of resources, however it takes a far superior understanding to achieve this. So for the average graphical programmer you might be better off sticking with OpenGL for the time being.