Silk.NET•3y ago

dxvk exists but idk if the developer experience on linux would be great

FFurball d3d11 is saner than opengl imo but it has the issue of kinda being windows speci...

Machinarius•4/8/23, 2:26 PM

I wonder if D3D11 + DXVK would be easier then?

Aqua•4/8/23, 2:26 PM

there is also my pie abstraction layer that has a similar API to dx11 but runs on both opengl + dx11 so its fully cross platform

Machinarius•4/8/23, 2:26 PM

I imagine I'd have to run my dotnet code through Wine/mono to get DXVK tho?

FurballOP•4/8/23, 2:27 PM

dxvk alone would be enough iirc

FurballOP•4/8/23, 2:27 PM

im pretty sure you lose the ability to debug with renderdoc though

FurballOP•4/8/23, 2:27 PM

or you have to debug but still partially have to know what vulkan does

Perksey•4/8/23, 2:27 PM

yeah d3d11 is best developed on windows, as at least you know your code works against the real thing before brining dxvk in the loop

Perksey•4/8/23, 2:28 PM

but dxvk pretty much works ootb with silk.net 2.17, no mono/wine needed

Machinarius•4/8/23, 2:28 PM

Gotcha. For now I'll stick to opengl and take it easy, baby-steps

Perksey•4/8/23, 2:28 PM

yeah fair

Perksey•4/8/23, 2:29 PM

if you do want to give yourself a little treat you could use OpenGL 4.5, the glNamed* functions are very nice to use

Howie•4/8/23, 2:29 PM

I definitely agree with that, they are super nice

FurballOP•4/8/23, 2:29 PM

but also come with the issue that renderdoc absolutely hates DSA

Perksey•4/8/23, 2:29 PM

did not know that

Perksey•4/8/23, 2:29 PM

that sucks

Howie•4/8/23, 2:29 PM

oof

FurballOP•4/8/23, 2:29 PM

beyley had issues with that

FurballOP•4/8/23, 2:29 PM

people in the GP server had issues

FurballOP•4/8/23, 2:29 PM

and apperantly its known and baldurk doesnt wanna bother or smth

Perksey•4/8/23, 2:30 PM

Howie•4/8/23, 2:30 PM

the DSA stuff does get you a little better prepped for Vulkan imo

Perksey•4/8/23, 2:30 PM

it also just makes OpenGL better to use as well

Perksey•4/8/23, 2:31 PM

i die everytime i need to do a glBind*

Howie•4/8/23, 2:31 PM

hehe

Howie•4/8/23, 2:31 PM

back in MY day...

Perksey•4/8/23, 2:32 PM

oh god here comes the glVertex3

Howie•4/8/23, 2:32 PM

how did you know?

Perksey•4/8/23, 2:32 PM

you awoken the era of graphics i've been trying to pretend doesn't exist

Perksey•4/8/23, 2:33 PM

speaking of we really need to delete Silk.NET.OpenGL.Legacy when

FurballOP•4/8/23, 2:33 PM

grDrawTriangle

Machinarius•4/8/23, 2:34 PM

Ok I finally understand! You can pack multiple bits of data in the vertex data like coordinates and base colors, and then tell OGL how to map that to the different positionposition values in the shader!

Aqua•4/8/23, 2:35 PM

yep

Machinarius•4/8/23, 2:36 PM

I was struggling with the example code having stride 77 until I saw it packing XYZ + RGBA on a single row

Machinarius•4/8/23, 2:46 PM

Does this abstraction make sense?

أأمير الدخميسي ahhhhhhhhhhh i very bad in engligh

innoxious•4/8/23, 2:52 PM

I can really see

Iinnoxious I can really see

أمير الدخميسي•4/8/23, 2:53 PM

everyone here able to that

MasterMan•4/8/23, 3:37 PM

How do I check the ComPtr<> object for validity? With regular pointers it's simple, but with a wrapper I've no idea
Google gives this (screenshot), but it's not valid in Silk.net