LogP Performance Assessment of Fast Network Interfaces

نویسندگان

  • David Culler
  • Lok Tin Liu
  • Richard P. Martin
  • Chad Yoshikawa
چکیده

We present a systematic performance assessment of the hardware and software that provides the interface between applications and emerging high-speed networks. Using LogP as a conceptual framework and Active Messages as the communication layer, we devise a set of communication microbenchmarks. These generate a graphical signature from which we extract the LogP performance parameters of latency, overhead, and bandwidth. The method is illustrated on three diverse platforms: Intel Paragon, Meiko CS-2, and a cluster of SparcStations with Myrinet. The study provides a detailed breakdown of the differences in communication performance among the platforms. While the details of our microbenchmark depend on Active Messages, the methodology can be applied to conventional communication layers.

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

LogP performance characterization of fast messages atop Myrinet

The Fast Messages (FM) project is a highly optimized , low latency messaging layer providing a virtual interface to the communication hardware. The purpose of this paper is to present a detailed characterization of the FM communication messaging layer using the LogP model. We have used a series of communication microbenchmarks to measure the parameters of the LogP model. These specialized bench...

متن کامل

Cross-Platform Analysis of Fast Messages for Myrinet

Designing high performance communication software requires an in-depth understanding of the role of the components of the underlying machine/network architecture and their reciprocal interactions. The task is complicated by the continuing technological advances in the machine architectures. In this paper we analyze the performance of the Fast Messages (FM) high performance messaging layer on th...

متن کامل

Fast Parallel Sorting under Logp : from Theory to Practice 3

1.1 ABSTRACT The LogP model characterizes the performance of modern parallel machines with a small set of parameters: the communication latency (L), overhead (o), bandwidth (g), and the number of processors (P). In this paper, we analyze four parallel sorting algorithms (bitonic, column, radix, and sample sort) under LogP. We develop implementations of these algorithms in a parallel extension t...

متن کامل

Efficient Parallel FFTs for Different Computational Models

We select the Fast Fourier Transfrom (FFT) to demonstrate a methodology for deriving the optimal parallel algorithm according to predetermined performance metrics, within a computational model. Following the vector space framework for parallel permutations, we provide a speci cation language to capture the algorithm, derive the optimal parallel FFT speci cation, compute the arithmetic, memory, ...

متن کامل

LogGP in theory and practice - An in-depth analysis of modern interconnection networks and benchmarking methods for collective operations

Accurate measurement and modeling of network performance is important for predicting and optimizing the running time of high-performance computing applications. Although the LogP family of models has proven to be a valuable tool for assessing the communication performance of parallel architectures, non-intrusive LogP parameter assessment of real systems remains a difficult task. Based on an ana...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 1996