As graphics processing units (GPUs) are broadly adopted, running multiple applications on a GPU at the same time is beginning to attract wide attention. Recent proposals on multitasking GPUs have focused on either spatial multitasking, which partitions GPU resource at a streaming multiprocessor (SM) granularity, or simultaneous multikernel (SMK), which runs multiple kernels on the same SM. However, multitasking performance varies heavily depending on the resource partitions within each scheme, and the application mixes. In this paper, we propose GPUMaestro that performs dynamic resource management for efficient utilization of multitasking GPUs. GPU Maestro can discover the best performing GPU resource partition exploiting both spatial multitasking and SMK. Furthermore, dynamism within a kernel and interference between the kernels are automatically considered because GPU Maestro finds the best performing partition through direct measurements. Evaluations show that GPU Maestro can improve average system throughput by 20.2% and 13.9% over the baseline spatial multitasking and SMK, respectively.
|Title of host publication||ASPLOS 2017 - 22nd International Conference on Architectural Support for Programming Languages and Operating Systems|
|Publisher||Association for Computing Machinery|
|Number of pages||14|
|Publication status||Published - 2017 Apr 4|
|Event||22nd International Conference on Architectural Support for Programming Languages and Operating Systems, ASPLOS 2017 - Xi'an, China|
Duration: 2017 Apr 8 → 2017 Apr 12
|Name||International Conference on Architectural Support for Programming Languages and Operating Systems - ASPLOS|
|Other||22nd International Conference on Architectural Support for Programming Languages and Operating Systems, ASPLOS 2017|
|Period||17/4/8 → 17/4/12|
Bibliographical notePublisher Copyright:
© 2017 ACM.
All Science Journal Classification (ASJC) codes
- Information Systems
- Hardware and Architecture