587
questions
0
votes
1
answer
27
views
Compute shader occlusion test only works with scene camera
Using Unity 2022.3.40f1 and URP.
I'm trying to implement occlusion culling with the depth buffer in a compute shader.
I don't know what's wrong here but my code only works fine with the scene camera ...
0
votes
1
answer
34
views
Raymarching compute shader error in Unity VR
I am trying to apply a raymarching shader to my Unity VR project.
It is supposed to apply a shader to particles that represent a fluid using Smooth Particle Hydrodynamics.
However, I am running into ...
0
votes
0
answers
52
views
Smoothing normal on GPU (using geometry shader) without knowledge to adjacent faces
Smoothing normal on GPU might be a common question because you really can generate flat normal by triangle primitives , but you don't have adjacent faces so it is not possible to smooth normal . I ...
0
votes
0
answers
19
views
Perform multiple lines in atomic way in compute shader
I am implementing a marching cubes algorithm in Unity using a compute shader. In my current implementation all the necessary data is stored in RWStructuredBuffer's with a theoretical maximum size. The ...
1
vote
0
answers
28
views
Metal Compute Shader Incorrect Output
I am writing an FFT in GLSL and compiling to MSL using glslang and SpirvCross to support macOS. The code works fine on Windows with GLSL, but on Metal I see output that makes no sense. However, I have ...
0
votes
0
answers
21
views
Unity: Set a mesh directly on the GPU
In unity I'm constructing a mesh on the GPU. Currently, I have to write the vertices / indices back to the CPU before setting them again in the mesh. This is very inefficient and takes about 90% of ...
1
vote
0
answers
20
views
Unity compute shader thread syncing issue
So I am writing a marching cubes compute shader. It’s going reasonably well and does work. I am in the processing of modifying it to remove all duplicate vertices. I want to do this in the shader. I ...
1
vote
0
answers
46
views
Why did my optimization of compute shader cache hit ratio have the opposite effect?
I wrote a compute shader in ue4 that each thread needs to get the value of a structured buffer corresponding to its own id.As shown in the figure below, for example, my texture is 8x8, the thread ...
0
votes
0
answers
37
views
WGPU compute shader inexplicably returns no data
I have a wgsl compute shader, that essentially checks pairs of datapoints and counts the number pairs that meet a certain requirement (passes(i,j) is true) for each point. (ie. if there are 5 passing ...
3
votes
0
answers
88
views
Can I use a GPUBuffer which is the output of a compute shader as a Three.js or Babylon.js vertex position buffer?
I am experimenting whith webgpu. I found out that transferring data between cpu and gpu can be quite slow.
My goal is to simulate particle physics on the gpu using webgpu's compute shaders, then use ...
0
votes
0
answers
34
views
Is it better to have multiple small textures or one large texture as input texture for compute shader?
My goal is to use 5 1080x1080 textures and use the algorithm in the compute shader to generate a new 1080x1080 texture. Each pixel will most take the colors of 16 adjacent pixels from the same input ...
0
votes
1
answer
74
views
How do I achieve the ping-ponging technique in HLSL compute shaders?
I am doing a fluid simulation, and implementing such simulation with the CPU is very slow, especially when the simulation grid size gets bigger or is a 3d grid. So I used shaders to get the ...
0
votes
0
answers
75
views
Unexpected results from Metal computing
EDIT:
In the comment below, pmdj pointed out 2 bugs in my code.
For clarity, I will not copy all the code and output buffer contents. Instead I modified both and left the originals as comments.
...
1
vote
1
answer
156
views
Is this WebGPU compute shader badly designed?
I am having unexpected results while comparing a WebGPU compute shader with it's cpu equivalent.
This is a CPU version of a function that does Verlet Integration for simulating particles.
...
0
votes
0
answers
29
views
How to retain bits on lowp float when using `imageLoad` and `imageStore`
I would like to use all 8-bits as flags on a lowp float stored in an image. However, for some reason the data is never retained when I read it back. I use uintBitsToFloat and floatBitsToUint like so.
(...