TY - GEN
T1 - Efficient CPU-GPU work sharing for data-parallel javascript workloads
AU - Piao, Xianglan
AU - Kim, Channoh
AU - Oh, Younghwan
AU - Kim, Hanjun
AU - Lee, Jae W.
N1 - Publisher Copyright:
© Copyright 2014 by the International World Wide Web Conferences Steering Committee.
Copyright:
Copyright 2016 Elsevier B.V., All rights reserved.
PY - 2014/4/7
Y1 - 2014/4/7
N2 - Modern web browsers are required to execute many complex, compute-intensive applications, mostly written in JavaScript. With widespread adoption of heterogeneous processors, re- cent JavaScript-based data-parallel programming models, such as River Trail and WebCL, support multiple types of processing elements including CPUs and GPUS. However, significant performance gains are still left on the table since the program kernel runs on only one compute device, typ- ically selected at kernel invocation. This paper proposes a new framework for efficient work sharing between CPU and GPU for data-parallel JavaScript workloads. The work shar- ing scheduler partitions the input data into smaller chunks and dynamically dispatches them to both CPU and GPU for concurrent execution. For four data-parallel programs, our framework improves performance by up to 65% with a geometric mean speedup of 33% over GPU-only execution.
AB - Modern web browsers are required to execute many complex, compute-intensive applications, mostly written in JavaScript. With widespread adoption of heterogeneous processors, re- cent JavaScript-based data-parallel programming models, such as River Trail and WebCL, support multiple types of processing elements including CPUs and GPUS. However, significant performance gains are still left on the table since the program kernel runs on only one compute device, typ- ically selected at kernel invocation. This paper proposes a new framework for efficient work sharing between CPU and GPU for data-parallel JavaScript workloads. The work shar- ing scheduler partitions the input data into smaller chunks and dynamically dispatches them to both CPU and GPU for concurrent execution. For four data-parallel programs, our framework improves performance by up to 65% with a geometric mean speedup of 33% over GPU-only execution.
UR - http://www.scopus.com/inward/record.url?scp=84990946857&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84990946857&partnerID=8YFLogxK
U2 - 10.1145/2567948.2577338
DO - 10.1145/2567948.2577338
M3 - Conference contribution
AN - SCOPUS:84990946857
T3 - WWW 2014 Companion - Proceedings of the 23rd International Conference on World Wide Web
SP - 357
EP - 358
BT - WWW 2014 Companion - Proceedings of the 23rd International Conference on World Wide Web
PB - Association for Computing Machinery, Inc
T2 - 23rd International Conference on World Wide Web, WWW 2014
Y2 - 7 April 2014 through 11 April 2014
ER -