Browsing by Author "Peng, Chao"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Screen Partitioning Load Balancing for Parallel Rendering on a Multi-GPU Multi-Display Workstation
    (The Eurographics Association, 2019) Dong, Yangzi; Peng, Chao; Childs, Hank and Frey, Steffen
    Commodity workstations with multiple GPUs have been built by engineers and scientists for real-time rendering applications. As a result, a high display resolution can be achieved by connecting each GPU to a display monitor (resulting in a tiled large display). Using a multi-GPU workstation may not always produce a highly interactive rendering rate due to imbalanced rendering workloads among GPUs. In this work, we propose a parallel load balancing algorithm based on a screen partitioning strategy to dynamically balance the amount of vertices and triangles rendered by each GPU. Each GPU renders a screen region whose size may be different from the screen regions of other GPUs, but the amounts of vertices and triangles in those screen regions are balanced. It is possible that a screen region rendered by a GPU has to be displayed by another GPU. We propose a frame exchanging algorithm that allows GPUs to exchange screen regions efficiently. The inter-GPU communication overhead is very small since the data transferred between GPUs are a small amount of image pixels.
  • No Thumbnail Available
    Item
    Spherical Parametric Measurement for Continuous and Balanced Mesh Segmentation
    (The Eurographics Association, 2023) Zhang, Huadong; Cao, Lizhou; Peng, Chao; Bikker, Jacco; Gribble, Christiaan
    Mesh segmentation is an important process for building the discrete mesh structure used on the GPU to accelerate geometry processing applications. In this paper, we introduce a novel mesh segmentation method that creates balanced sub-meshes for high-performance geometry processing. The method ensures topological continuity within sub-meshes (segments) and evenly distributes the number of triangles across all sub-meshes. A new cohesion algorithm computes the chord distances between triangles in the spherical domain and re-groups the triangles into the sub-meshes based on a distance-based measurement condition. A new refinement algorithm between the neighboring sub-meshes is conducted to resolve the non-manifold issue and improve the boundary smoothness. Both algorithms are executed in a parallel fashion. In advancing the state-of-the-art, our approach achieves exactly balanced triangle counts and mitigates the non-manifold issue significantly. The algorithms require the input meshes to have a closed-manifold genus of zero, which is a constraint that is commonly associated with the concept of sphere-based parameterization. We evaluated the effectiveness of our approach in supporting two geometry processing applications. The results show that the performance is enhanced by leveraging the structure of the balanced sub-meshes from our approach.