A Semi-Automated Tool Flow for Roofline Anaylsis of OpenCL Kernels on Accelerators
ICHEC, Dublin, Ireland
First International Workshop on Heterogeneous High-performance Reconfigurable Computing (H2RC’15), 2015
@article{muralidharan2015semi,
title={A Semi-Automated Tool Flow for Roofline Anaylsis of OpenCL Kernels on Accelerators},
author={Muralidharan, Servesh and O’Brien, Kenneth and Lalanne, Christian},
year={2015}
}
We propose a tool-flow methodology that can be applied to analyze and track the performance of OpenCL applications on heterogeneous platforms. Using a case study on a datacenter representative workload, we evaluate our tool flow on three distinct heterogeneous platforms and demonstrate how it can be employed more widely to provide insight and track attainable performance of OpenCL applications. Our methodology is motivated by the need for a common set of metrics that can characterize the performance and power efficiency of OpenCL applications on the increasingly diverse range of emerging heterogeneous platforms now relevant to both HPC and the datacenter market.
December 8, 2015 by hgpu
