A Parallel Matrix Inversion Algorithm on Torus with Adaptive Pivoting

نویسندگان

  • Javed I. Khan
  • Woei Lin
  • David Y. Y. Yun
چکیده

This paper presents a parallel algorithm for matrix inversion on a torus interconnected MIMD-MC multi-processor. This method is faster than the parallel implementations of other widely used methods namely Gauss-Jordan, Gauss-Seidal or LU decomposition based inversion. This new algorithm also introduces a novel technique, called adaptive pivoting, for solving the zero pivot problem at no cost. Our method eliminates the costly row interchange used by the existing elimination based parallel algorithms.This paper presents the design, analysisand simulation results (on a 32 Node Meiko Transputer) of this new and efficient matrix inversion algorithm.

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

A Parallel Method for Matrix Inversion Based on Gauss-jordan Algorithm

The inversion of matrix is widely used in many scientific applications and engineering calculations and this operation requires a large number of computational efforts and storage space. In order to reduce the consumed time and to increase efficiency, a parallel algorithm for matrix inversion based on classic Gauss-Jordan elimination with pivoting is proposed. Upon the multi-core DSPs platform,...

متن کامل

Adaptive attitude controller of a reentry vehicles based on Back-stepping Dynamic inversion method

This paper presents an attitude control algorithm for a Reusable Launch Vehicle (RLV) with a low lift/drag ratio (L/D < 0.5), in the presence of external disturbances, model uncertainties, control output constraints and the thruster model. The main novelty of proposed control strategy is a new combination of the attitude control methods included backstepping, dynamic inversion and adaptive cont...

متن کامل

Scalable Stable Solvers for Non-symmetric Narrow-Banded Linear Systems

Banded linear systems with large bandwidths can be solved by similar methods as full linear systems. In particular, parallel algorithms based on torus-wrap mapping and Gaussian elimination with partial pivoting have been used with success. These algorithms are not suitable, however , if the bandwidth is small, say, between 1 and 100. As the bandwidth limits the amount of parallelism available a...

متن کامل

Adaptive Algorithm-based Fault Tolerance for Parallel Computations in Linear Systems

This paper presents a dynamically adaptive stabilization scheme for parallel matrix computation. The scheme performs automatic error detection and correction through inserting redundant, but concurrent tracer computations within the folds of the regular computation. It also eliminates thecostly rowinterchangeused in classical pivoting. A fault-tolerant double wavefront matrix algorithm for a MI...

متن کامل

Minimizing communication overhead for matrix inversion algorithms on hypercubes

The mirin contribution of this report is the development of novel algorithms {that make efficient use of the communication system in distributed memory architectures with plrocessing elements interconnected by a hypercube network. These algorithms achieve almost optirr~al overlap of communicatior~ delays by computation, leading to a minimization of communicatioi~ overhead. Rigorous ana1yt:ical ...

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 1992