USENIX ATC '23 - VectorVisor: A Binary Translation Scheme for Throughput-Oriented GPU Acceleration

USENIX ATC '23 - VectorVisor: A Binary Translation Scheme for Throughput-Oriented GPU Acceleration

Samuel Ginzburg, Princeton University, Mohammad Shahrad, University of British Columbia, Michael J. Freedman, Princeton University

Beyond conventional graphics applications, general-purpose GPU acceleration has had significant impact on machine learning and scientific computing workloads. Yet, it has failed to see widespread use for server-side applications, which we argue is because GPU programming models offer a level of abstraction that is either too low-level (e.g., OpenCL, CUDA) or too …

applications beyond binary british british columbia columbia computing general gpu graphics impact machine machine learning michael purpose translation university usenix workloads