Automatic Tuning of a Parallel Pattern Library for Heterogeneous Systems with Intel Xeon Phi

Abstract

Pattern libraries are important tools for high productivity application development. Their struggle for best performance is complicated by the fact that they are used to execute user-provided code, which is not known during their creation. This makes pattern libraries good candidate for automatic software tuning. In this paper, we deal with automatic online parameter tuning of the HyPHI hybrid pattern library for heterogeneous systems equipped with the Intel Xeon Phi coprocessors. We propose a framework that can be used to combine a pattern library with an existing tuning library in a practical and efficient way. Our experiments show that tuning can noticeably improve the performance of the library and it introduces very little overhead.

Hybrid small size high performance computing resource for medical image analysis

“…To implement the above method to optimize performance, different parts of the corresponding codes run on different PUs to be computed at the same time. Namely, an image is processed using a CPU including Intel graphics cards (GFX) if exists, a GPU by Nvidia CUDA libraries, and an Intel MIC PU architecture, typically on Xeon Phi processors by Intel Parallel Libraries [9]. As Gaussian blur and minimum/maximum pixel intensity extraction run sufficiently fast on CPU, it is worthless uploading and downloading data to/from co-processor cards.…”