OSPRay - A CPU Ray Tracing Framework for Scientific Visualization

Abstract

Scientific data is continually increasing in complexity, variety and size, making efficient visualization and specifically rendering an ongoing challenge. Traditional rasterization-based visualization approaches encounter performance and quality limitations, particularly in HPC environments without dedicated rendering hardware. In this paper, we present OSPRay, a turn-key CPU ray tracing framework oriented towards production-use scientific visualization which can utilize varying SIMD widths and multiple device backends found across diverse HPC resources. This framework provides a high-quality, efficient CPU-based solution for typical visualization workloads, which has already been integrated into several prevalent visualization packages. We show that this system delivers the performance, high-level API simplicity, and modular device support needed to provide a compelling new rendering framework for implementing efficient scientific visualization workflows.

Screen-Space Normal Distribution Function Caching for Consistent Multi-Resolution Rendering of Large Particle Data

“…However, in principle all of these approaches suffer from aliasing due to severe under-sampling of very large scenes, which can still be explored interactively. The aliasing problem also applies to recent CPU-based methods [33,34], which scale significantly better in terms of performance than in terms of visual accuracy. Complementary techniques have also been developed to reduce the resulting noise [12].…”

“…That is, for a desired output resolution, we perform multiple iterations where each one computes one normal vector per pixel, and enters it into the pixel's S-NDF. Our implementation uses GPU ray casting [12] to generate normal vector samples, but any other method could be used as well, e.g., [33,34]. Fig.…”

“…However, at these scales the major factor that limits the effectiveness of visualizations is the insufficient sampling rate that is caused by the mismatch between data size and the resolution of today's output devices. In general, techniques based on ray casting [13,18,25,33,34] are prone to aliasing due to under-sampling, which leads to noisy renderings that are inconsistent across scales. The importance of this effect is shown in Fig.…”

“…However, our entire S-NDF pipeline could equivalently be implemented with state-of-the-art CPU ray tracers, e.g., [33,34], with no conceptual differences, while obtaining the same benefits. The most important property of our method is the computation and use of S-NDFs, which makes it inherently consistent across scales and output resolutions.…”

Performance Evaluation of Runtime Data Exploration Framework based on In-Situ Particle Based Volume Rendering

“…Embree [21] is a photorealistic ray-tracer, which consists of a set of low-level kernels for multiple platforms, and has a simple API to port its kernels. OSPRay [20] is a multi-platform ray-tracing framework for scientific visualization on GPUs and multiple CPU architectures with varying SIMD widths, and is integrated into VisIT and ParaView. BnsView [8] is a molecular visualization framework which delivers fast volume rendering and ball-and-stick ray casting on Xeon Phi and is implemented in a SPMD language.…”