This paper by Purcell et al. presents a photon mapping algorithm that runs entirely on the GPU. The paper presents details for tracing photons, building the photon map, and computing the radiance estimate at each pixel using a k-nearest neighbor search.(Photon Mapping on Programmable Graphics Hardware. Timothy J. Purcell, Craig Donner, Mike Cammarano, Henrik Wann Jensen, and Pat Hanrahan. Proceedings of Graphics Hardware 2003, July 2003.)

This paper by Carr et al. describes a method for computing subsurface scattering on the GPU. They use a multi-resolution meshed atlas and modern GPU programmability to devise a real-time GPU algorithm that can render semi-transparent objects with diffuse subsurface-scattered illumination under dynamic lighting and viewing conditions. (GPU Algorithms for Radiosity and Subsurface Scattering. Nathan A. Carr, Jesse D. Hall, and John C. Hart. Proceedings of Graphics Hardware 2003, July 2003.)

This paper by Bill Mark et. al describes the Cg high-level graphics hardware programming language and the associated run-time system. The paper details the rationale behind the decisions made in the design of the language. (Cg: A system for programming graphics hardware in a C-like language. Bill Mark, R. Steven Glanville, Kurt Akeley, and Mark J. Kilgard. To appear in the proceedings of SIGGRAPH 2003.)

This paper by Harris et al. at UNC Chapel Hill describes a physically-based, visually-realistic interactive cloud simulation implemented on programmable GPUs. The simulation uses “flat 3D textures” – 3D data laid out as slices tiled in a 2D texture – to implement efficient 3D simulations on the GPU. The work required to simulate a single time step is automatically spread over many frames while the user views the results of the previous time step at high frame rates. (Simulation of Cloud Dynamics on Graphics Hardware. Mark J. Harris, William V. Baxter III, Thorsten Scheuermann and Anselmo Lastra. To appear in Proceedings of Graphics Hardware 2003.)

This paper by Theodore Kim and Ming Lin describes a physically-based technique for simulating ice crystal growth. The most expensive part of the simulation pipeline, the phase and temperature field calculations, can be mapped to graphics hardware. The resulting implementation experiences up to a 9x speedup. (Visual Simulation of Ice Crystal Growth. Theodore Kim and Ming Lin. To appear in the proceedings of ACM SIGGRAPH / Eurographics Symposium on Computer Animation 2003.)

This SIGGRAPH 2003 paper by Karl Hillesland, Sergey Molinov, and Radek Grzeszczuk casts nonlinear optimization as a data streaming process for computation on GPUs. The authors apply this approach to two distinct image-based modeling problems: light field mapping approximation and fitting the Lafortune model to spatial BRDFs. (Nonlinear Optimization Framework for Image-Based Modeling on Programmable Graphics Hardware. Karl E. Hillesland, Sergey Molinov, and Radek Grzeszczuk. To appear in the proceedings of SIGGRAPH 2003.)

This paper by Bolz et al. of Cal Tech shows two basic, broadly useful, computational kernels implemented on GPUs: a sparse matrix conjugate gradient solver, and a regular-grid multigrid solver. The paper demonstrates a prototype implementation on NVIDIA’s GeForce FX, using geometric flow (cube smoothing movie, 3D photography scan denoising movie) and fluid simulation (particle advection movie) as application examples. (Sparse Matrix Solvers on the GPU: Conjugate Gradients and Multigrid. Jeff Bolz, Ian Farmer, Eitan Grinspun and Peter Schröder. To appear in the proceedings of SIGGRAPH 2003.)

Abstract: The Sequenced Convex Subtraction (SCS) algorithm for Constructive Solid Geometry (CSG) sequentially subtracts convex volumes from the z-buffer. This paper presents an improvement to subtraction sequence generation which uses object space overlap information to give O(n) length sequences in the best case and (unchanged) O(n²) sequences in the worst case. (Improved CSG Rendering using Overlap Graph Subtraction Sequences. N. Stewart, G. Leach, S. John. International Conference on Computer Graphics and Interactive Techniques in Australasia and South East Asia – GRAPHITE 2003, pp. 47-53)

This paper by Goodnight et al. at the University of Virginia demonstrates an implementation on two modern graphics architectures of the multigrid method for solving boundary value problems, such as the systems of partial differential equations that arise in physical simulation problems like fluid flow and heat transfer. (A Multigrid Solver for Boundary Value Problems Using Graphics Hardware. Nolan Goodnight, Gregory Lewin, David Luebke, and Kevin Skadron, University of Virginia Technical Report CS-2003-03 (January 2003).)

This paper by Lefohn et. al. at the University of Utah demonstrates a 3D level-set PDE solver implemented entirely on an ATI Radeon 8500 GPU. It shows that in addition to the basic level-set equations, the second-order mean curvature speed term can be successfully evaluated with only 8-bit textures. The paper demonstrates the solver segmenting the cortical surface from a 256 x 256 x 175 MRI volume and compares the results to a CPU implementation. The object oriented framework with which the solver was built, “Glift,” is also discussed. (A GPU-Based, Three-Dimensional Level Set Solver with Curvature Flow. Aaron Lefohn and Ross Whitaker. University of Utah tech report UUCS-02-017, December, 2002.)