DySel: Lightweight dynamic selection for kernel-based data-parallel programming model

TY - GEN

T1 - DySel

T2 - 21st International Conference on Architectural Support for Programming Languages and Operating Systems, ASPLOS 2016

AU - Chang, Li Wen

AU - Kim, Hee Seok

AU - Hwu, Wen Mei

PY - 2016/3/25

Y1 - 2016/3/25

N2 - The rising pressure for simultaneously improving performance and reducing power is driving more diversity into all aspects of computing devices. An algorithm that is wellmatched to the target hardware can run multiple times faster and more energy efficiently than one that is not. The problem is complicated by the fact that a program's input also affects the appropriate choice of algorithm. As a result, software developers have been faced with the challenge of determining the appropriate algorithm for each potential combination of target device and data. This paper presents DySel, a novel runtime system for automating such determination for kernel-based data parallel programming models such as OpenCL, CUDA, OpenACC, and C++AMP. These programming models cover many applications that demand high performance in mobile, cloud and high-performance computing. DySel systematically deploys candidate kernels on a small portion of the actual data to determine which achieves the best performance for the hardware-data combination. The test-deployment, referred to as micro-profiling, contributes to the final execution result and incurs less than 8% of overhead in the worst observed case when compared to an oracle. We show four major use cases where DySel provides significantly more consistent performance without tedious effort from the developer.

AB - The rising pressure for simultaneously improving performance and reducing power is driving more diversity into all aspects of computing devices. An algorithm that is wellmatched to the target hardware can run multiple times faster and more energy efficiently than one that is not. The problem is complicated by the fact that a program's input also affects the appropriate choice of algorithm. As a result, software developers have been faced with the challenge of determining the appropriate algorithm for each potential combination of target device and data. This paper presents DySel, a novel runtime system for automating such determination for kernel-based data parallel programming models such as OpenCL, CUDA, OpenACC, and C++AMP. These programming models cover many applications that demand high performance in mobile, cloud and high-performance computing. DySel systematically deploys candidate kernels on a small portion of the actual data to determine which achieves the best performance for the hardware-data combination. The test-deployment, referred to as micro-profiling, contributes to the final execution result and incurs less than 8% of overhead in the worst observed case when compared to an oracle. We show four major use cases where DySel provides significantly more consistent performance without tedious effort from the developer.

KW - Dynamic Profiling

KW - Graphics Processing Unit

UR - http://www.scopus.com/inward/record.url?scp=84975230376&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84975230376&partnerID=8YFLogxK

U2 - 10.1145/2872362.2872373

DO - 10.1145/2872362.2872373

M3 - Conference contribution

AN - SCOPUS:84975230376

T3 - International Conference on Architectural Support for Programming Languages and Operating Systems - ASPLOS

SP - 667

EP - 680

BT - ASPLOS 2016 - 21st International Conference on Architectural Support for Programming Languages and Operating Systems

PB - Association for Computing Machinery

Y2 - 2 April 2016 through 6 April 2016

ER -