The Community for Technology Leaders
2014 23rd International Conference on Parallel Architecture and Compilation (PACT) (2014)
Edmonton, Canada
Aug. 23, 2014 to Aug. 27, 2014
ISBN: 978-1-5090-6607-0
pp: 497-498
Bo Wu , The College of William and Mary, Virginia, USA
Guoyang Chen , The College of William and Mary, Virginia, USA
Dong Li , Oak Ridge National Laboratory, Tennessee, USA
Xipeng Shen , The College of William and Mary, Virginia, USA
Jeffrey S. Vetter , Oak Ridge National Laboratory, Tennessee, USA
ABSTRACT
To circumvent the limitation from the hardware scheduler on GPU, we create an SM-centric transformation technique. This technique enables complete control of the mapping between tasks and streaming multi-processors (SMs), and enables controlling the number of active thread blocks on each SM. Results show that our approach achieves better speedup than previous ones with kernel co-run cases.
INDEX TERMS
Graphics processing units, Hardware, Kernel, Instruction sets, Parallel processing, Runtime, Optimization
CITATION
Bo Wu, Guoyang Chen, Dong Li, Xipeng Shen, Jeffrey S. Vetter, "SM-centric transformation: Circumventing hardware restrictions for flexible GPU scheduling", 2014 23rd International Conference on Parallel Architecture and Compilation (PACT), vol. 00, no. , pp. 497-498, 2014, doi:10.1145/2628071.2628130
80 ms
(Ver 3.3 (11022016))