On this website we provide additional and updated material to the course on Advanced Real-Time Volume Graphics that was originally held at SIGGRAPH'26.

Google slides, version submitted to SIGGRAPH'26
Google slides, updated version

CUDA/C++ sample code framework on GitHub.com
Data sets

Extended course abstract (pdf, 2.3 MB)
Course notes (pdf, 20 MB)

On Windows, for some of the dependencies we provide installers for convenience. On Unix-like operating systems we recommend to build the dependencies on the command line using the CMake and Ninja or Makefile toolchain (cmake $(build_dir); make && make install).

On NVIDIA systems with OptiX support (the OptiX library comes with the driver), it is recommended to use the NVIDIA OWL toolchain. On Windows, it is recommended to either build OWL from source, or to install including dependencies (CUDA v12.6, Visual Studio C++ x64 redistributable package). Binaries were created with Visual Studio 2019 (msvc 16):

OWL-1.2.4-win64.msi (msi, 1.7 MB)
NVIDIA CUDA 12.6: https://developer.nvidia.com/cuda-12-6-0-download-archive
64-bit redistributable package for Visual C++ 14 and above: https://aka.ms/vc14/vc_redist.x64.exe

On systems without NVIDIA GPU you can use fakeOWL for CPU emulation. On Unix-like systems, it is recommended to build the dependency from source. On Windows, we provide an installer built with Visual Studio 2019 (msvc 16):

fakeOWL-0.2.0-win64.msi (msi, 620 KB)
64-bit redistributable package for Visual C++ 14 and above: https://aka.ms/vc14/vc_redist.x64.exe

Responsibility for the content of this webpage is with the authors who are at this time researchers at various institutes, including the University of Cologne, the University of Utah (SCI), and NVIDIA. Please send correspondences to Stefan Zellmann.