1992
DOI: 10.1007/3-540-55895-0_431
|View full text |Cite
|
Sign up to set email alerts
|

Numerical performance of an asynchronous Jacobi iteration

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

1
8
0

Year Published

1996
1996
2023
2023

Publication Types

Select...
5
2
1

Relationship

0
8

Authors

Journals

citations
Cited by 11 publications
(9 citation statements)
references
References 3 publications
1
8
0
Order By: Relevance
“…It performs an asynchronous Jacobi with local Gauss-Seidel method as in [2] on the system of linear equations in (6). This method converges for any initial estimate of r when T W − I is strictly or irreducibly diagonally dominant [3], [4], [5] and in fact in some cases the asynchronous version has been shown to converge faster than the synchronous version [2]. Thus, to prove that the risk calculations will converge, it is enough to show strict diagonal dominance for T W − I.…”
Section: A Convergencementioning
confidence: 99%
“…It performs an asynchronous Jacobi with local Gauss-Seidel method as in [2] on the system of linear equations in (6). This method converges for any initial estimate of r when T W − I is strictly or irreducibly diagonally dominant [3], [4], [5] and in fact in some cases the asynchronous version has been shown to converge faster than the synchronous version [2]. Thus, to prove that the risk calculations will converge, it is enough to show strict diagonal dominance for T W − I.…”
Section: A Convergencementioning
confidence: 99%
“…In [3,5,7] the authors study the performance of APJ in terms of time to converge. In this paper we will make a similar study except that we record a complete log of all of the iterations and communications that take place in the implementation.…”
Section: X[i](k)=m[ii]*x[i](k-1)+c[i] For All Inputs J X[i](k)=x[i](kmentioning
confidence: 99%
“…The entire system of nodes is thus collectively solving a set of linear equations, in which each row corresponds to one node, in a distributed fashion by performing an asynchronous Jacobi with local Gauss-Seidel method, as in [6].…”
Section: Coupled Risk Estimation Algorithmmentioning
confidence: 99%