Daily reminder

No one cares fuck off

>>55811545
the board is literally full of "muh vulkan" and "muh dx12" every single day. would not hurt if people read an article about them.

Fuck off no one cares

If you are going to post documents on what it is don't post a link from that shit.

Also fuck off with these shit threads
https://en.m.wikipedia.org/wiki/Vulkan_(API)

>>55811604
did daddy touch you back then?
please, show us in paint where did he touch you.

File: 1465686927441.gif (2MB, 400x206px) Image search: [Google]

2MB, 400x206px

>>55811615
>I got btfo'dbso I have to resort to memery

Every time

>>55811537
>believing nvidia's lies
daily reminder they have no ability to do hardware async, yet claim full dx12 compatibility
daily reminder they are paying 30$ per card for lying about the 970's capabilities
daily reminder they have lied about power draw and faked benches in the past

>>55811636
> btfo
> throwing words around you don't even know

10/10

File: g_in_charge_of_anything.jpg (252KB, 1049x569px) Image search: [Google]

252KB, 1049x569px

>>55811661
>yet claim full dx12 compatibility

Daily reminder #2:
No vendor besides Intel have "Full dx12 compatibility". Only Skylake does.

https://en.wikipedia.org/wiki/Feature_levels_in_Direct3D#Direct3D_12

File: 1469617315581.jpg (51KB, 514x536px) Image search: [Google]

51KB, 514x536px

>>55811667
Ironic desu

>>55811661
>believing nvidia's lies
> how to spot the person without any 3D or coding experience

>>55811682
He said nothing about others. He said nvidia claims full dx12 compatibility when it does not.

File: lolpng.png (149KB, 1120x977px) Image search: [Google]

149KB, 1120x977px

>>55811682
>skylake gpu is more compatible than jewvidia and amdpoors

OpenGL/Vulkan(I can only imagine) is the most annoying and tedious kind of programming. You spent like hundreds of lines of code setting up contexts and buffers through an obscure C interface with hundreds of GL_ENUM_CONTANTs. And apparently it's even worse in Vulkan? Fuck that, fuck making vidya games in C++, I'll be using Unity or something that isn't designed to be a pain in the ass.

>>55811843
A lot of what you mention is boilerplate code. You write that once, then abstract it away into higher level functions.

Enjoy your Unity though, should be right up your alley.

>>55811843
Well Vulkan is like C in my opinion. Room for optimization, but with freedom comes human error.

>>55811537
>QQ software development is hard QQ

>>55812128
> computers evolve
> programming languages evolve
> we can now code programs easily and have less bugs

guys, guys, what if... we go back to the stone age?!

10/10

>>55812150
Next thing you're going to bitch about how complex integrated circuit design is.

Unity and other babby tier abstractions are widely available for Pajeets like you.

>>55811537
Who the fuck made those retarded charts? Jesus Christ.

>>55812573
>>55812578
what is this "tl;dr pulled out of my ass"?

>>55812585

Read it and educate yourself - it cites sources for all the claims made so no rectal spelunking is required.

>>55812661
Read the first, it's an "amd_paid_posting.jpg", basically what you can read in every AMD or anti-Nvidia thread.

>>55811537
>actually believing there is a meme miracle API that will boost performance by 15-20% without being as stable as a schizophrenic rabid chimp
>actually thinking any developer is going to sink time and money when most are still using 2 gen old engines to make games.

File: 2828467.jpg (55KB, 637x375px) Image search: [Google]

55KB, 637x375px

>>55812723

Just because you may dislike the information presented does not change the veracity of it. People have been talking about this sorto f thing for the best part of a year and /g/ STILL continues to deny some of the shortcomings Nvidia has for DX12 despite quite a mountain of evidence.

File: xBgSksM.jpg (18KB, 340x260px) Image search: [Google]

18KB, 340x260px

>>55812748

Oh boy do I have the picture for you.

>>55812661
>read it and educate yourself
>literally AMD propaganda materials

>>55812785
>Cheat
Yes it makes AMD cards run faster under certain conditions,
No, nobody cares because the adoption rate is laughably small and people actually still own and play current gen games.
So yes please purchase an AMD card if you enjoy waiting two years for your fans/mobo to die before actually getting a library of games that utilize the meme APIs.
Or be a smart consumer and buy what you need when you need it.
>mfw retards buy 480s day one when I'll pick one up on sale for $100 in a years time for meme async compute.

File: dx12-support-645x366.jpg (58KB, 645x366px) Image search: [Google]

58KB, 645x366px

>>55812812

I can't wait for my gtx 580 to get its DX12 support.

>>55811537
There's this thing called "abstraction". Premature optimization is the root of all evil, after all. Vulkan is important because some things are simply impossible with OpenGL.

>>55811537
Its not verbose..
It is harder but for good reason. I'm not sure what sort of replies OP wants.

>>55813015

The kind that downplays the advantages these new APIs can bring.

>>55812573
>>55812578
interesting read.
guess that answers my question about my 1080 running dx12.
just fuck my shit

>>55812748
> how to spot someone who didn't even open the link

The point Nvidia is trying to make that Vulkan enables developers to do more, but also requires more effort.

>>55812770
There are still no proper DX12 titles around. Most of them runs the same as DX11 (Memes of Singularity) did, or even worse (Tomb Raider).

And, if you checked the OP link, DX12 is 100% supported only on Skylake. Good luck gaming with that.

>>55812829
Again, bullshit.
Even 9 series GTX cards only have the basic feature set. The additional features? Barely supported. Same on AMD.

>>55813015
I tried to raise awareness?
People have no idea what they shitpost about (what a surprise).

>>55813147

>Again, bullshit.

Don't look at me, that picture was straight from Nvidia themselves.

>>55813196
I know, and it's stupid how they shill this.

>>55811682
All hail Intel!!
Opengl 4.5 support on linux when?

>>55813196
>>55813368
Fuck hotkeys.
And it's stupid how they can claim all the 6xx (or whatever) series will have DX12 support.
It's false advertising, if anything.

>>55812150
>computers evolve
>programming languages evolve
>we waste resources on bad optimization
this is actually stopping progress tbhwy.

File: 1080 vs 980ti identical clocks.jpg (154KB, 1920x869px) Image search: [Google]

154KB, 1920x869px

>>55813381

I do believe they genuinely were going to support DX12 all the way back to fermi until it became clear internally that doing so would show up the change in direction their architecture has had. Why do you think /g/ shilled so much for maxwell having "full" DX12 support? It is an easy thing to say but few people realise how much of it is bullshit. I just suspect Nvidia thought they had it all sewn up but MS (and now khronos) have moved faster than they anticipated given (relatively) software has a shorter development cycle than hardware.

>>55813147
>The additional features? Barely supported. Same on AMD.
Maxwell 2nd Gen and up supports 12.1 feature set, where has AMD supports 12.0

Nvidia supports 3d sparse textures and Rasterizer-ordered views on the 9 series, where AMD does not, and that's a pretty big deal.

>>55814376

Correction, 12_1 and 12_0 respectively This is very different from what 12.1 may end up being (see also: 11.1, 11.2 and 11.3).

I have been in graphics programming for years. Haven't gotten in dx12 yet (which is assumed to be similar to vulkan) but as far as I saw, Vulkan is much more primitive which makes it harder to learn and harder to build an architecture on it. You have many ways to do something but just deciding for the correct way is another effort. Vulkan is definitely not for indies.

No one gives a shit, it has huge performance benefits and it's still probably easier than DX12.

Go watch the Vulkan n64 emulator preview. Accurate LLE at playable speed? Yes please.

>>55814505

I wonder if that dude who made a DX12 backend for dolphin will bother porting it to vulkan.

>>55814523
Wouldn't surprise me. Team Retroarch has a raging hardon for Vulkan, so probably most of the cores will get re-worked to use it over time, and those guys work closely with other emulator teams. I wouldn't be surprised at all if someone stepped in to make plugins for PS1 and PS2 also.

>>55811537
>AMD shows that Time Spy is shit
>LELELEL AMD is salty

>Nvidia shows that Vulkan is shit
>Oh my god, can't you see it's a conspiracy

>>55814486
DX12 and Vulkan are the same, low level.
DX11 and OpenGL are more alike.

That is why OpenGL will kept updated and improved - said Khronos - because most projects don't need the trouble that comes with Vulkan.

>>55814505
>it has huge performance benefits
Depends.
See Tomb Raider, Ashes of Singularity, and so on.
You really never know, unless someone writes a 100% optimized DX11, OpenGL, Vulkan, DX12 test.

>>55814588
> my little shitposter cannot be this retarded

File: randomId.png (16KB, 200x200px) Image search: [Google]

16KB, 200x200px

Vulkan (glsllang) does not support subroutines (function pointer) at the moment. This is an essential part of my real-time raytracer and a show-stopper.

I'm also missing SPIR-V reflections.I have a an auto-bind mechanism that cannot be applied to Vulkan at the moment.

Otherwise the validation layers simplify the debugging.

Another nice feature: With the support of SPIR-V, all graphics cards can now use the same computation shaders (theoretically). For example physic engines can be easier used.

>>55814592
But that's the whole point anon, how hard do you think it is to write some 100% optimized OpenGL or DX11 shit? No one does it, no one even knows HOW to do it, that's why Vulkan was made. Yeah, sure, it might be slightly more difficult to use, but in exchange for that, you can access the hardware directly, and you don't NEED to optimize your shit to get the performance gains.

>>55814566
PCSX2 is broken by design
SSE and AVX/FMA give literally no performance benefit despie Emotion Engine making heavy use of SIMD instructions and even integrated FMA unit

I'm wondering, have any of you here, ever coded in opengl or vulkan? Or dx11 or dx12?

>>55814770
That's depressing, since no alternative exists or is even being considered. Well, I use OpenPS2Loader anyway, so I guess it's not my problem. However, from a game archival perspective, it's not a pleasant situation when the only working emulator for a system is a bad one.

AMD just released a driver that gives 10% boost to RotTR on the RX 480. Bang goes your 'wave in your face' title for the GTX 1060 Nvidiots.

>>55814523
https://github.com/dolphin-emu/dolphin/pull/3935

>>55811537
you mean same nvidia that swore freesync would never ever work on anything but laptops and only their superior $200 chip can do that?

File: old man huang.jpg (25KB, 475x324px) Image search: [Google]

25KB, 475x324px

>>55815708

I have no memory of such a claim.

>>55811537
So, Nvidia cards are lagging in modern API's because their programmers are just lazy?

>>55815924
Pretty much. They only know DX11 and are getting on a bit. They are looking forward to their retirement homes and are not au fait with this newfangled technology.

>>55813425
>identical clocks

The whole fucking point of the new architecture is that they can run at higher clocks with lower power consumption. This is literal statistical gerrymandering on the scale of U.S. politics about guns.

>>55811745
Dx12 is tiered, you don't need to support everything to be compatible.

File: 1439920545027.png (65KB, 211x202px) Image search: [Google]

65KB, 211x202px

>>55811537
Once again, Nvidia is shooting itself in the foot just so it can take a bazooka to AMD's foot.

>>55814802
openGL I would expect so. In my university it's something they teach, or is in option. It's used on most scientific projects that require 3D stuff since it's open source and run on linux which is what people use.

You use DX only when you're working for a company or wants to do something very specific.

And about vulkan I guess it will be too complex to replace openGL

>>55811537
That's what i was saying Vulkan is DOA, it takes way too much time for devs to learn new api while they can just use dx12 without a lot of learning since they worked on dx11/dx10/dx9.
Remember how many AAA devs used OpenGL? TWO fucking two.
In year and half nobody will be using vulkan and after 3 years nobody will remember vulkan except AYYYMD Mad boyz.

File: Motion Pictography.png (153KB, 1254x1138px) Image search: [Google]

153KB, 1254x1138px

>>55817656
>AAA
Bet you're the kind of guy who watches flicks like "Avengers" or "Bat Man" unironically.

Meanwhile, Geforce 980/970/960 is already in legacy status:
http://www.nvidia.com/page/legacy.html

>>55816140
Nvidia already got a lawsuit for their Jewish tricks yet you fucks still suck their dick

Its amazing how many people theyve turned into their cockslave.

>>55811537
>Vulkan is a lot more verbose, lot harder.
Well yeah it's closer to the hardware of course it is. Do you not understand the concept of a low level API?

Nobody cares if it's harder, if it allows an easier time porting games(and therefore saving money) then management will force them to use it over OpenGL or DX12

With OpenGL a lot of things were abstracted away to save the programmer from having to write them, this is boilerplate code. With vulkan there's none of those abstractions, you have to make your own boilerplate code. This gives devs more freedom with the API but they also have to do more work. If you wanted to make a successor to OpenGL you could use vulkan as a base for it and provide your own libraries for commonly used functionality but vulkan alone is not directly comparable to OpenGL

File: penn_jillette.jpg (31KB, 300x400px) Image search: [Google]

31KB, 300x400px

>>55811537
>Daily reminder.
>Vulkan is a lot more verbose, lot harder.
>https://developer.nvidia

File: Lords of the Fallen 07.04.2016 - 15.21.48.04.DVR.webm (3MB, 1920x810px) Image search: [Google]

3MB, 1920x810px

>>55811661
Daily reminder you're just another illiterate nigger from AMD stable and async was never part of DX12, it was just a feature enabled by DX12.

Daily reminder that outside of your meme imaginary hug board 970 gtx has no real performance drops when accessing the last 512mbs of vram.

Daily reminder Nvidia isn't the one making GPUs that literally are a potential risk to your MOBO and faked benches is literally shill fallacy aka "the benches that favour my company are real, the ones that don't must be fake".

>>55818681
It did cause severe dips when using over 3.5gb and under 4gb, thats how it was found out in the first place.
GTX480 and it risk your enter house rather than just the mobo
Nvidia have the worst benchmark bs, pic related

It's like you made a post exactly opposite to reality as if you have a fetish for being told your wrong.

>>55818422
games supported by both Android and Windows are so sparse because of totally different controls between both platforms and Vulkan won't change that. OS/Linux support is a meme, those platforms have terrible driver support for a reason - they're literally less than 3% of market share and most of them run on potato/overpriced rigs which aren't suitable for gaming. Once again Vulkan won't change shit cuz nobody will bother supporting dead platforms. Consoles are a completely different thing and MS will push for their DX12 on every Xbone port while Sony does their own thing which according to Brad Wardell has an API working on much lower level than Mantle or Vulkan.

So what you end up with is an API that is literally harder to learn, harder to use and its cross-platform support is absolutely worthless. Thanks /g/.

File: FGM2.jpg (46KB, 300x419px) Image search: [Google]

46KB, 300x419px

>>55818800
>N-No! This can't be happening! I'm in charge here!

The post.

>>55816074
That was what the point he, adoredTV, tried to make with that slide. Pascal isn't really a new architecture but just Maxwell on finfet. So if they were to preform at the same clocks they would preform equal. But if you were to put Kepler vs. Maxwell at the same clock speed Maxwell would be better because the there had been architectural changes to improve performance.
Pascal have other improvements of course like better memory compression to warrant calling it a new architecture.

Here's the video https://www.youtube.com/watch?v=nDaekpMBYUA

>>55818950
>Linking to this AMD shill

He literally made a 12 mins long video just so he can Pascal old architecture.

>>55818984
>He literally made a 12 mins long video just so he can call Pascal old architecture.

Yes. He made the claim earlier sometime around the release of the 1070/80. This was just him backing up his claim form then.

>>55818984
Except he gives high praise to Nvidia for Maxwell's high clock rates and low power usage. He also says AMD's Vega needs to be a lot higher than a crappy 1400Mhz to even bother with. He shills for AMD but he also praises Nvidia for having the better GPU's in terms of performance.

>>55818984
I don't believe he is a shill. In his first videos he talks a lot about the position AMD is in, the bad CPUs, the lack of a top tier GPU, but also see things on the horizon, like the new ZEN and what it could do with all that GPUOpen thing.

>a lower level and closer to the metal API is harder than a higher level one
Who could have guessed? What's going to be your next discovery, OP? That assembly language is harder than Visual Basic?

>>55811537
>low level API
>more verbose
NO SHIT DUDE, HOW DID YOU FIND OUT?

>>55820230
This.

>>55819802

He's pretty uninformed in that case since the lack of a hardware scheduler is why pascal has lower temps and this allowing high clocks.

The hardware scheduler on amds cards is what gives them the async compute edge. If Nvidia adds one in, the temps would shoot up

Nvidia did compute cards anyways with fermi. They're not new to this shit. Too bad we'll never see fermi at full potential because Nvidia pulled support when they realized it would have performance increase and not newer cards like kepler or maxwell

Maybe Volta will have it?

>>55811661
https://www.youtube.com/watch?v=yZ_ZL7SHi1M

>>55821207

>pulled support

FUN FACT: maxwell is now on legacy support.

>>55811661
you are completely technically illiterate and don't even know what those terms you're throwing around mean

>>55821732
Preemption ≠ async compute. They even show it in the beginning of the fucking video, did you even watch it before posting it you fucktard?

>>55824040
but amd is gay and for niggers with no money you're wrong faggot kill yourself

File: nvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvidiots.png (21KB, 967x357px) Image search: [Google]

21KB, 967x357px

Nvidiots in a nutshell. Fucking high school dropout retards.

>>55811537
>Vulkan is a lot more verbose, lot harder.

I hate gamers. They're all so stupid.

>>55824114
Ikr they should all fuck right off back to /v/ where they belong. Shitting up our board, what happened to good ol' /g/ when we actually understood technology?

>>55824040
What the fuck is async compute even? Why would we give a fuck about a sync. If Nvidia doesn't have it, its not worth our time anyways. AMDfag stay pjavascript:;oor.

>>55811537
So what do you suggest? We stay on DX11 for another decade the same way we stayed on DX9 for a decade?

>>55811537
So a low level api is harder to use...
in other news water is wet

>>55814702
You still need to optimize but you can do it further.

>>55824569
>This anon had a stroke while shilling
may he rest in pieces

>>55814691
I think this is the first real reply if you start from the top

>>55814802
Yes. I wrote a Vulkan based game engine with a team for a research project. My partner did the DX12 renderer. By the way you're all fucking retarded and should stop posting about this shit because none of you know anything. Just play your games and have fun. Real people with jobs will work with Vulkan and DX12 and do things that matter.

>>55814691
For SpirV reflection use the SpirVCross tool. ARM donated it to Khronos. Also use OpenCL 2.0 for raytracing or Vulkan compute shaders if you NEED to use Vulkan. It's designed to handle compute applications but I see most professionals opting to just use OpenCL for raytracing.

I bet none of you assholes have ever even designed a basic deferred renderer so you should probably shut the fuck up about graphics technology. It's sickening that you people even pretend to know anything about this shit.

>.nvidia
Wait a second, the graphics card company whose performance compared to similar AMD cards trailed way behind its competitor when using Vulkan?

Off course they're fucking saying this lmao, they can't keep up with the times and are desperate not to lose market share

>>55817656
It's not about learning it, lol.
Developing a render / game for Vulkan takes more effort than OpenGL, no matter how skilled a developer/team is.

Same applies to DX12.

>>55818150
I do.
Lots of people on /g/ don't.
That's why it's a PSA thread so fuckers at least read about it before they start shitposting again about muh +100000 FPS in doom.

>>55818655
Please do not associate based Penn with dumb children level AMD shitposting. Thank you.

>>55820550
> the point
>
>
>
>
> your head

I mean I understand not everyone can read, not everyone have such a high iq and things like that, but damn, it's crazy how many dumbfucks we have here.

>>55825215
No, thread is about trying to show shitposters what is Vulkan and DX12 about. Because they have zero clue.

Other than that, yes, Khronos and developers say to wait for new OpenGL that will allow to use similar features with none of the trouble. I don't know, I developed only with DirectX11 and OpenGL, and I only made simple games/3D applications, like 3D car games and such, nothing fancy.

But for me, I would use OpenGL. It's very easy to learn and with helping C++ libs you can be up and running in no time.

>>55828620
> can't even read an article
> don't know anything about graphics
> his brain works like a simple grep script
..... this is just sad.

I wonder what his parents think. Did they already give up?