TOD-Tree: Task-Overlapped Direct send Tree Image Compositing for Hybrid MPI Parallelism

dc.contributor.authorGrosset, A. V. Pascalen_US
dc.contributor.authorPrasad, Manasaen_US
dc.contributor.authorChristensen, Cameronen_US
dc.contributor.authorKnoll, Aaronen_US
dc.contributor.authorHansen, Charlesen_US
dc.contributor.editorC. Dachsbacher and P. Navrátilen_US
dc.description.abstractModern supercomputers have very powerful multi-core CPUs. The programming model on these supercomputer is switching from pure MPI to MPI for inter-node communication, and shared memory and threads for intra-node communication. Consequently the bottleneck in most systems is no longer computation but communication between nodes. In this paper, we present a new compositing algorithm for hybrid MPI parallelism that focuses on communication avoidance and overlapping communication with computation at the expense of evenly balancing the workload. The algorithm has three stages: a direct send stage where nodes are arranged in groups and exchange regions of an image, followed by a tree compositing stage and a gather stage. We compare our algorithm with radix-k and binary-swap from the IceT library in a hybrid OpenMP/MPI setting, show strong scaling results and explain how we generally achieve better performance than these two algorithms.en_US
dc.description.sectionheadersImproved Algorithmsen_US
dc.description.seriesinformationEurographics Symposium on Parallel Graphics and Visualizationen_US
dc.publisherThe Eurographics Associationen_US
dc.subjectI.3.1 [Computer Graphics]en_US
dc.subjectHardware Architectureen_US
dc.subjectParallel processingen_US
dc.subjectI.3.2 [Computer Graphics]en_US
dc.subjectGraphics Systemsen_US
dc.subjectDistributed/network graphicsen_US
dc.titleTOD-Tree: Task-Overlapped Direct send Tree Image Compositing for Hybrid MPI Parallelismen_US
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
5.18 MB
Adobe Portable Document Format