DwarfCode: A Performance Prediction Tool for Parallel Applications

Abstract

We present DwarfCode, a performance prediction tool for MPI applications on diverse computing platforms. The goal is to accurately predict the running time of applications for task scheduling and job migration. First, DwarfCode collects the execution traces to record the computing and communication events. Then, it merges the traces from different processes into a single trace. After that, DwarfCode identifies and compresses the repeating patterns in the final trace to shrink the size of the events. Finally, a dwarf code is generated to mimic the original program behavior. This smaller running benchmark is replayed in the target platform to predict the performance of the original application. In order to generate such a benchmark, two major challenges are to reduce the time complexity of trace merging and repeat compression algorithms. We propose an O(mpn) trace merging algorithm to combine the traces generated by separate MPI processes , where m denotes the upper bound of tracing distance, p denotes the number of processes, and n denotes the maximum of event numbers of all the traces. More importantly, we put forward a novel repeat compression algorithm, whose time complexity is O(nlogn). Experimental results show that DwarfCode can accurately predict the running time of MPI applications. The error rate is below 10% for compute and communication intensive applications. This toolkit has been released for free download as a GNU General Public License v3 software.

Optimizing GPU Memory Transactions for Convolution Operations

“…A wide range of techniques have been proposed to accelerate convolution operations [1], [2], [3], [4], [5], [6], [7], [8]. Among these methods, general matrix multiplication (GEMM) [6], [7], fast fourier transform (FFT) [2] and winograd [3] methods are the broadly adopted ones.…”

Automatic translation of data parallel programs for heterogeneous parallelism through OpenMP offloading

“…It seems that these applications are not suitable for heterogeneous platforms. To solve this problem, a promising approach is to make the translator can predict appli-cation performances on different platforms [35,36,32,6], and automatically decide whether to offload or not.…”

Multi-Parameter Performance Modeling Based on Machine Learning with Basic Block Features

“…Trace-based methods: Trace-driven methods, which are frequently used in simulators [8], [7] and benchmark generation tools [24], [25], can capture detailed performance behavior and model the performance of parallel programs automatically. However, trace-driven methods have some limitations.…”