EGPGV20: Eurographics Symposium on Parallel Graphics and Visualization

Permanent URI for this collection

https://diglib.eg.org/handle/10.2312/2632899

Browse

Browsing EGPGV20: Eurographics Symposium on Parallel Graphics and Visualization by Subject "Computing methodologies"

Now showing 1 - 5 of 5
  • Thumbnail Image
    Item
    Effective Parallelization Strategies for Scalable, High-Performance Iterative Reconstruction
    (The Eurographics Association, 2020) Gribble, Christiaan Paul; Frey, Steffen and Huang, Jian and Sadlo, Filip
    Iterative reconstruction techniques in X-ray computed tomography converge to a result by successively refining increasingly accurate estimates. Compared to alternative approaches, iterative reconstruction imposes significant computational demand but generally leads to higher reconstruction quality and is more robust to inherently imperfect scan data. We explore several strategies for exploiting parallelism in iterative reconstruction and evaluate their scalability and performance on modern workstation-class systems. Results show that scalable, high performance iterative reconstruction is possible with careful attention to the expression of parallelism in both the projection and backprojection phases of computation.
  • Thumbnail Image
    Item
    Fast Multi-View Rendering for Real-Time Applications
    (The Eurographics Association, 2020) Unterguggenberger, Johannes; Kerbl, Bernhard; Steinberger, Markus; Schmalstieg, Dieter; Wimmer, Michael; Frey, Steffen and Huang, Jian and Sadlo, Filip
    Efficient rendering of multiple views can be a critical performance factor for real-time rendering applications. Generating more than one view multiplies the amount of rendered geometry, which can cause a huge performance impact. Minimizing that impact has been a target of previous research and GPU manufacturers, who have started to equip devices with dedicated acceleration units. However, vendor-specific acceleration is not the only option to increase multi-view rendering (MVR) performance. Available graphics API features, shader stages and optimizations can be exploited for improved MVR performance, while generally offering more versatile pipeline configurations, including the preservation of custom tessellation and geometry shaders. In this paper, we present an exhaustive evaluation of MVR pipelines available on modern GPUs. We provide a detailed analysis of previous techniques, hardware-accelerated MVR and propose a novel method, leading to the creation of an MVR catalogue. Our analyses cover three distinct applications to help gain clarity on overall MVR performance characteristics. Our interpretation of the observed results provides a guideline for selecting the most appropriate one for various use cases on different GPU architectures.
  • Thumbnail Image
    Item
    Finding Efficient Spatial Distributions for Massively Instanced 3-d Models
    (The Eurographics Association, 2020) Zellmann, Stefan; Morrical, Nate; Wald, Ingo; Pascucci, Valerio; Frey, Steffen and Huang, Jian and Sadlo, Filip
    Instancing is commonly used to reduce the memory footprint of massive 3-d models. Nevertheless, large production assets often do not fit into the memory allocated to a single rendering node or into the video memory of a single GPU. For memory intensive scenes like these, distributed rendering can be helpful. However, finding efficient data distributions for these instanced 3-d models is challenging, since a memory-efficient data distribution often results in an inefficient spatial distribution, and vice versa. Therefore, we propose a k-d tree construction algorithm that balances these two opposing goals and evaluate our scene distribution approach using publicly available instanced 3-d models like Disney's Moana Island Scene.
  • Thumbnail Image
    Item
    High-Quality Rendering of Glyphs Using Hardware-Accelerated Ray Tracing
    (The Eurographics Association, 2020) Zellmann, Stefan; Aumüller, Martin; Marshak, Nathan; Wald, Ingo; Frey, Steffen and Huang, Jian and Sadlo, Filip
    Glyph rendering is an important scientific visualization technique for 3D, time-varying simulation data and for higherdimensional data in general. Though conceptually simple, there are several different challenges when realizing glyph rendering on top of triangle rasterization APIs, such as possibly prohibitive polygon counts, limitations of what shapes can be used for the glyphs, issues with visual clutter, etc. In this paper, we investigate the use of hardware ray tracing for high-quality, highperformance glyph rendering, and show that this not only leads to a more flexible and often more elegant solution for dealing with number and shape of glyphs, but that this can also help address visual clutter, and even provide additional visual cues that can enhance understanding of the dataset.
  • Thumbnail Image
    Item
    Improving Performance of M-to-N Processing and Data Redistribution in In Transit Analysis and Visualization
    (The Eurographics Association, 2020) Loring, Burlen; Wolf, Mathew; Kress, James; Shudler, Sergei; Gu, Junmin; Rizzi, Silvio; Logan, Jeremy; Ferrier, Nicola; Bethel, E. Wes; Frey, Steffen and Huang, Jian and Sadlo, Filip
    In an in transit setting, a parallel data producer, such as a numerical simulation, runs on one set of ranks M, while a data consumer, such as a parallel visualization application, runs on a different set of ranks N: One of the central challenges in this in transit setting is to determine the mapping of data from the set of M producer ranks to the set of N consumer ranks. This is a challenging problem for several reasons, such as the producer and consumer codes potentially having different scaling characteristics and different data models. The resulting mapping from M to N ranks can have a significant impact on aggregate application performance. In this work, we present an approach for performing this M-to-N mapping in a way that has broad applicability across a diversity of data producer and consumer applications. We evaluate its design and performance with a study that runs at high concurrency on a modern HPC platform. By leveraging design characteristics, which facilitate an ''intelligent'' mapping from M-to-N, we observe significant performance gains are possible in terms of several different metrics, including time-to-solution and amount of data moved.