Keynote
Massive Parallelism in Intel's Graphics Processors.
The performance improvement in graphics processors over the past decade has been tremendous, and it is intriguing to see that the graphics part of the chip die now is larger than for the CPU part. In this presentation, I will explain how this kind of parallelism can be achieved by describing the graphics processor in Intel's Broadwell architecture. In addition, I will show how a ray tracer can be mapped, using OpenCL, to such an architecture with shared memory, and will demonstrate interactive performance.
|
Tomas Akenine Möller
is a professor in computer science with specialization in computer graphics and image processing at the Department of Computer Science, Lund University, Sweden. Over the past years, he has built the computer graphics group, LUGG (Lund University Graphics Group), there. He received an MSc in Computer Science and Engineering from Lund in 1995, and went on to Chalmers University of Technology, where he got his PhD in computer graphics in 1998. During 2000, he was a post doc at UC Berkeley, and also spent some time at UC San Diego (2004/2005) as a visiting researcher. Tomas has been co-papers-chairing Graphics Hardware (2004), Eurographics Symposium on Rendering (2006), and Eurographics 2010. He currently works part time at Intel in Lund as a tech lead with a fantastic team of graphics researchers. In December 2008, Intel acquired Swiftfoot Graphics, which is a company he co-founded with three of his PhD students. Tomas is also a co-author of the book Real-Time Rendering. He has written 80+ papers, and still counting.
|
|
Program
Monday, May 25
|
|
|
14.40
|
Efficient Representations, session chair: Timo Ropinski, University of Ulm
|
|
|
Large-Scale Parallel Visualization of Particle-Based Simulations using Point Sprites and Level-Of-Detail
|
Silvio Rizzi, Mark Hereld, Joseph Insley, Michael E. Papka, Thomas Uram, Venkatram Vishwanath
|
|
Memory-Efficient On-The-Fly Voxelization of Particle Data
|
Tobias Zirr, Carsten Dachsbacher
|
|
Visualization of 2D Wave Propagation by Huygens' Principle
|
Stefan Heßel, Olliver Fernandes, Sebastian Boblest, Malte Hoffmann, Andrea Beck, Philipp Offenhäuser, Claus-Dieter Munz, Colin Glass, Thomas Ertl, Filip Sadlo
|
|
Visualization Showcase: General-Relativistic Black Hole Visualization
|
Thomas Müller, Sebastian Boblest, Daniel Weiskopf
|
|
|
16.30
|
Parallel Rendering, session chair: Marco Ament, Karlsruhe Institute of Technology
|
|
|
SIMD Parallel Ray Tracing of Polyhedral Grids
|
Brad Rathke, Ingo Wald, Ingo; Kenneth Chiu, Carson Brownlee
|
|
Packet-Oriented Streamline Tracing on Modern SIMD Architectures
|
Bernd Hentschel, Jens Henrik Göbbert, Paul Springer, Michael Klemm, Andrea Schnorr, Torsten Kuhlen
|
|
Volume Rendering Via Data-Parallel Primitives
|
Matthew Larsen, Stephanie Labasan, Paul Navrátil, Jeremy Meredith, Hank Childs
|
|
Visualization Showcase: Visualization of High-Resolution Weather Model Data
|
Si Liu, Greg Foss, Greg Abram, Anne Bowen
|
|
|
18.30
|
Welcome Reception at T-Hotel
|
|
Tuesday, May 26
|
|
9.00
|
Improved Algorithms, session chair: Filip Sadlo, University of Heidelberg
|
|
|
TOD-Tree: Task-Overlapped Direct send Tree Image Compositing for Hybrid MPI Parallelism
|
Pascal Grosset, Manasa Prasad, Cameron Christensen, Aaron Knoll, Charles Hansen
|
|
Contour Tree Depth Images For Large Data Visualization
|
Tim Biedert, Christoph Garth
|
|
Out-of-Core Framework for QEM-based Mesh Simplification
|
Hiromu Ozaki, Fumihito Kyota, Takashi Kanai
|
|
Visualization Showcase: Visualizing Groundwater Flow Through Karst Limestone
|
Carson Brownlee, Aaron Knoll, Paul Navrátil, Kevin J. Cunningham, Michael C. Sukop, and Sadé Garcia
|
|
|
11.00
|
Keynote Lecture by Tomas Akenine Moller: Massive Parallelism in Intel's Graphics Processors.
|
|
|
|