Search found 44 matches

by szellmann
Wed Jun 21, 2017 1:48 pm
Forum: Tools, Demos & Sources
Topic: Visionaray Ray Tracing Framework
Replies: 11
Views: 17975

Re: Visionaray Ray Tracing Framework

Here's the (extended) AO results from the SEARIS paper (https://szellmann.github.io), along with a script to run the benchmark. Results for interesting platforms (CPU,GPU) not listed yet are highly welcome!
by szellmann
Thu Mar 16, 2017 10:40 am
Forum: Tools, Demos & Sources
Topic: Visionaray Ray Tracing Framework
Replies: 11
Views: 17975

Re: Visionaray Ray Tracing Framework

Going to talk about Visionaray at SEARIS (http://www.searis.net), which is a workshop colocated with IEEE VR.

Check out this teaser video: https://youtu.be/bhhwEO4KWII
by szellmann
Fri Jan 20, 2017 9:10 pm
Forum: General Development
Topic: Intersection of ray with quadric surface for hobby ray tracer
Replies: 6
Views: 1950

Re: Intersection of ray with quadric surface for hobby ray tracer

Here's an explanation of how to build up the 4x4 matrix for several quadric types: https://www.siggraph.org/education/materials/HyperGraph/raytrace/rtinter4.htm I believe that Glassner's 1980's book "Introduction to Ray Tracing" also has a paragraph on this. There's a ray/quadric implement...
by szellmann
Fri Dec 30, 2016 1:05 am
Forum: Tools, Demos & Sources
Topic: Visionaray Ray Tracing Framework
Replies: 11
Views: 17975

Re: Visionaray Ray Tracing Framework

It's christmas holidays, and so I had a little fun compiling Visionaray for my new Raspberry PI 3: https://youtu.be/7bJTEmdlT2Y I haven't done anything special but applying some tiny compile fixes for this architecture, so there's no optimizations yet. I'm just porting the SIMD math lib to ARM NEON....
by szellmann
Mon Dec 12, 2016 3:46 pm
Forum: General Development
Topic: Multiple Textures CUDA/OpenCL
Replies: 8
Views: 9008

Re: Multiple Textures CUDA/OpenCL

I used tiles/bricks, used morton curves and my texturing functions also support SoA access w/ SSE and AVX. Of course the outcome of the tests depends on overall coherence. Baseline was that linear filtering had a significant impact when texturing was emulated, while there was virtually no difference...
by szellmann
Mon Dec 12, 2016 8:17 am
Forum: General Development
Topic: Multiple Textures CUDA/OpenCL
Replies: 8
Views: 9008

Re: Multiple Textures CUDA/OpenCL

In my personal experience the __global char* approach is very fast even if it's not hardware accelerated. I'm curious, what difference in latency do you actually observe when comparing nearest neighbor vs. linear filtering. My rt lib has a dedicated CPU API where you can call texture access "i...
by szellmann
Thu Dec 08, 2016 10:37 am
Forum: General Development
Topic: Multiple Textures CUDA/OpenCL
Replies: 8
Views: 9008

Re: Multiple Textures CUDA/OpenCL

From CUDA 6.0 or so onwards, and on Kepler and newer, you can use texture objects. I'd guess that they are quite efficient, just as bindless textures in OpenGL are. You can have a dynamic number of texture objects / can pass a list (e.g. thrust::device_vector) of texture objects to the kernel. For l...
by szellmann
Wed Aug 31, 2016 10:54 am
Forum: Visuals
Topic: GPU VCM
Replies: 30
Views: 13672

Re: GPU VCM

Edit: seems like there is no problem with the handedness after all, but rather with me not properly understanding tangent space :) Sorry for capturing the thread, this errant is unrelated.. @atlas: The method in the paper is especially nice because it is SIMD/SoA vector friendly. With conventional m...
by szellmann
Mon Jul 11, 2016 9:12 am
Forum: General Development
Topic: Map OpenGL depth buffer in CUDA kernel
Replies: 0
Views: 6074

Map OpenGL depth buffer in CUDA kernel

I'm cross-posting a question I already asked on the Nvidia developer forums https://devtalk.nvidia.com/default/topic/948593/cuda-programming-and-performance/map-opengl-depth-buffer-in-cuda-kernel/. I thought that maybe someone here might have had similar problems in the past and would share his or h...

Go to advanced search