Team Overview

While Moore’s Law continues, increases in clock speed have been stagnant due to physical obstacles. Thus, hardware and software have turned towards parallelism as an answer to keep up with the ever growing demand for more compute power. The probably highest degree of parallelism within a single chip is found on the modern graphics processing unit (GPU). However, traditional algorithms do not work well on massively parallel devices like the GPU; especially concerning the control over the execution itself. We investigate novel algorithms for scheduling dynamic workloads on the GPU, including algorithms for efficient task scheduling, memory management, and dynamic workload prioritization. Our research enables algorithms—which previously did not seem fit for GPU execution—to harness the full power of the GPU. Applying our insights to computer graphics, we are interested in highly efficient and on-the-fly generation of procedural content, prioritizing image synthesis for deadline-driven foveated rendering, and alternative rendering pipeline designs with new views on rasterization.

High performance graphics

Lead: Markus Steinberger, Dieter Schmalstieg

Summary: We have established a track record of advanced parallel programming techniques on the GPU. A particular area of interest lies on algorithms, which do not have parallel workloads immediately available up-front, but require dynamic scheduling. Examples for such algorithms are sparse matrix operations, deep tree traversals or adaptive rendering techniques. Dynamic scheduling also lends itself to enforcing priorities and carefully balancing producer-consumer problems.

Selected publications

Bernhard Kerbl, Michael Kenzel, Dieter Schmalstieg, Hans-Peter Seidel, and Markus Steinberger. Hierarchical Bucket Queuing for Fine-grained Priority Scheduling on the GPU. Computer Graphics Forum, 2016.

Rostislav Khlebnikov, Philip Voglreiter, Markus Steinberger, Bernhard Kainz, and Dieter Schmalstieg. Parallel Irradiance Caching for Interactive Monte-Carlo Direct Volume Rendering. Computer Graphics Forum, 33(3):61-70, June 2014.

Markus Steinberger, Michael Kenzel, Pedro Boechat, Bernhard Kerbl, Mark Dokter, and Dieter Schmalstieg. Whippletree: Task-based Scheduling of Dynamic Workloads on the GPU. ACM Transactions on Graphics (Proc. SIGGRAPH Asia 2014), December 2014.

Markus Steinberger, Bernhard Kainz, Bernhard Kerbl, Stefan Hauswiesner, Michael Kenzel, and Dieter Schmalstieg. Softshell: Dynamic Scheduling on GPUs. ACM Transactions on Graphics (Proc. SIGGRAPH Asia), 31(6), November 2012. Article 161.

Markus Steinberger, Bernhard Kainz, Stefan Hauswiesner, Rostislav Khlebnikov, Denis Kalkofen, and Dieter Schmalstieg. Ray Prioritization Using Stylization and Visual Saliency. Computers and Graphics, 36(6):673-684, October 2012.

Bernhard Kainz, Markus Grabner, Alexander Bornik, Stefan Hauswiesner, Judith Muehl, and Dieter Schmalstieg. Ray Casting of Multiple Volumetric Datasets with Polyhedral Boundaries on Manycore GPUs. Transactions on Graphics (Proc. SIGGRAPH Asia), 28(5), 2009.

Procedural modeling

Lead: Markus Steinberger, Dieter Schmalstieg 

Summary: Procedural modeling is essential in applications which require very large geometry sets at a level of detail which cannot be created by hand. For example, detailed city models can contain hundreds of thousands of buildings and billions of polygons. Such large models may not even fit into the available memory. We have developed the first approach to evaluate procedural models on the fly, thereby supporting literally infinitely detailed models.

Selected publications

Pedro Boechat, Mark Dokter, Michael Kenzel, Hans-Peter Seidey, Dieter Schmalstieg, and Markus Steinberger. Representing and Scheduling Procedural Generation using Operator Graphs. ACM Transactions on Graphics (Proc. SIGGRAPH Asia), December 2016.

Markus Steinberger, Michael Kenzel, Bernhard Kainz, Joerg Mueller, Peter Wonka, and Dieter Schmalstieg. Parallel Generation of Architecture on the GPU. Computer Graphics Forum, 33(2):73-82, April 2014. Best paper honorable mention award.

Markus Steinberger, Michael Kenzel, Bernhard Kainz, Peter Wonka, and Dieter Schmalstieg. On-the-fly Generation and Rendering of Infinite Cities on the GPU. Computer Graphics Forum, 33(2):105-114, April 2014.