High Performance Ray Tracing With Embree On Intel Xeon Phi

Ray tracing is a technique for generating images of synthetic scenes. Because ray tracing simulates the physics of light transport in the real world, it can be used to achieve high quality and even photorealistic results. The chapter authors in High Performance Parallelism Pearls describe how the Intel Embree ray tracing kernel library can be used to achieve high performance and high image quality on Intel Xeon Phi coprocessors. They demonstrate that this hardware / software combination can reach levels of performance on non-trivial workloads with performance comparable to – and often better – than any other CPU or GPU known today. Moreover, the ray tracing computation itself and related vectorization strategies are conceptually straightforward for non-experts, and a simple example is given that illustrates the key elements of an application which uses Embree on Xeon Phi coprocessors.

For more information about why I think augmented reality is the “gold rush” of the twenty-teens and why animation tools like Embree are a key technology, please see:

• Augmented Reality Company Magic Leap Bought by Google for $500M
• Farber HPCwire interview (Oct. 14. 2014).
• Scientific Computing blog (Oct. 13, 2014).
• TechEnablement: Compare Augmented Reality Displays from NVIDIA, Oculus Rift, Meta, and Others

Rendered by Embree, the Imperial Crown of Austria

Chapter Authors

Gregory Johnson

Gregory S. Johnson is a computer graphics researcher at Intel. His areas of interest include real-­‐time and photorealistic rendering, visibility algorithms, spatial data structures, and graphics hardware architectures. Greg received his Ph.D. in 2008 in Computer Science at the University of Texas under Bill Mark, for his work on the Irregular Z-­‐Buffer, a hardware architecture for accelerating irregular data structures useful in visibility computations. Greg was previously a Research Associate with the Texas Advanced Computing Center at UT, where he developed new methods for distributed and large-­‐data visualization, and led the Scientific Visualization Group. Before heading to UT, Greg served as an Associate Staff Scientist at the San Diego Supercomputer Center, and in the distant past spent a summer working at NeXT. 

Ingo Wald

Ingo Wald is a Research Scientist at Intel. After receiving his Ph.D. in engineering from Saarland University, Germany, he first did a one-­‐year Post-­‐Doc at the Max-­Planck Institute for Informatics in Saarbruecken, Germany. He then joined the Scientific Computing and Imaging Institute at the University of Utah, where he eventually became a Research Assistant Professor at the School of omputing. In late 2007, Ingo left the University of Utah to join Intel Corp as a Research Scientist. Ingo’s research interest revolve around all aspects of ray tracing and lighting simulation, realtime graphics, parallel computing, and general visual / high-­‐performance computing.

Sven Woop

Sven Woop is a Graphics Research Scientist at Intel Corporation, where he develops the Embree Ray Tracing Kernels. He received his masters degree in Computer Science in 2004 and his Ph.D. on a hardware architecture for Real­‐time Ray Tracing in 2007 at Saarland University in Germany. His research interests include Computer Graphics, Hardware Design, and Parallel Programming.

Carsten Benthin

Carsten Benthin is a Graphics Research Scientist at Intel Corporation. His research interests include all aspects of ray tracing and high performance rendering, throughput and high performance computing, low­‐level code optimization, and massively parallel hardware architectures. Carsten received his Ph.D. in 2006 from Saarland University in Germany.

Manfred Ernst

Manfred Ernst is a member of the Chromium team at Google. From 2009 to 2013, Manfred was a Research Scientist at Intel Labs, where he worked on interactive ray tracing. Prior to joining Intel, he co-­‐founded Bytes+Lights, which developed software for CAD data preparation and visualization. From 2003-­2007 Manfred worked as a researcher for the Computer Graphics Group at the University of Erlangen-­‐Nuremberg, where he received his Ph.D. in Computer Science. His primary research interests are photorealistic rendering, acceleration structures for ray tracing, sampling and data compression.

Click to see the overview article “Teaching The World About Intel Xeon Phi” that contains a list of TechEnablement links about why each chapter is considered a “Parallelism Pearl” plus information about James Reinders and Jim Jeffers, the editors of High Performance Parallelism Pearls.