Matches in SemOpenAlex for { <https://semopenalex.org/work/W1613527773> ?p ?o ?g. }
Showing items 1 to 77 of
77
with 100 items per page.
- W1613527773 abstract "As sequential computers seem to be approaching their limits in CPU speed there is increasing interest in parallel computers. This development calls for parallel algorithms for problems that may already have efficient sequential algorithms.The problem of solving a linear system of equations arises in many areas of science and engineering. Quite often each equation only involves a small number of the variables. In that case the linear system is sparse. If we can take advantage of the sparsity we can solve much larger systems.Consider the linear system Ax = $b$, where $A$ is a sparse symmetric positive definite matrix. A common approach to solving this system is to use Cholesky factorization. Algorithms that solve sparse linear systems using the Cholesky factorization usually consist of the following steps.First the matrix $A$ is permuted in order to get a sparse Cholesky factor $L$. Then there is the symbolic factorization of $A$ to determine the nonzero structure of $L$. Finally the value of $L$ is calculated in the numeric factorization phase and $x$ is computed by solving the triangular systems Ly = $b$ and $Lsp{T}x$ = $y$.In this thesis we present parallel algorithms for all the above steps except for the permutation step.Before the symbolic factorization we compute the elimination tree of $A$. Elimination trees have many applications in sparse matrix computations. Therefore our parallel algorithm to find elimination trees is important in its own right. The algorithm we present for symbolic factorization then uses the elimination tree to compute the nonzero structure of $L$.We next present a parallel algorithm to compute the numeric factorization of $A$. It runs in time proportional to the height of $A$'s elimination tree times a log factor. We also show how that algorithm can be converted into an NC algorithm (i.e., an algorithm that runs in polylogarithmic time) by the use of fast algorithms for dense matrices.Finally we demonstrate a parallel algorithm to solve sparse triangular systems of equations. There again we show a version that runs in time related to the height of the elimination tree and a version that is an NC algorithm." @default.
- W1613527773 created "2016-06-24" @default.
- W1613527773 creator A5090785363 @default.
- W1613527773 date "1989-01-03" @default.
- W1613527773 modified "2023-09-27" @default.
- W1613527773 title "Parallel sparse Cholesky factorization" @default.
- W1613527773 hasPublicationYear "1989" @default.
- W1613527773 type Work @default.
- W1613527773 sameAs 1613527773 @default.
- W1613527773 citedByCount "1" @default.
- W1613527773 crossrefType "journal-article" @default.
- W1613527773 hasAuthorship W1613527773A5090785363 @default.
- W1613527773 hasConcept C11413529 @default.
- W1613527773 hasConcept C121332964 @default.
- W1613527773 hasConcept C134306372 @default.
- W1613527773 hasConcept C134978465 @default.
- W1613527773 hasConcept C137559481 @default.
- W1613527773 hasConcept C158693339 @default.
- W1613527773 hasConcept C163716315 @default.
- W1613527773 hasConcept C187834632 @default.
- W1613527773 hasConcept C2524010 @default.
- W1613527773 hasConcept C33923547 @default.
- W1613527773 hasConcept C34727166 @default.
- W1613527773 hasConcept C41008148 @default.
- W1613527773 hasConcept C42355184 @default.
- W1613527773 hasConcept C44363057 @default.
- W1613527773 hasConcept C46085209 @default.
- W1613527773 hasConcept C56372850 @default.
- W1613527773 hasConcept C62520636 @default.
- W1613527773 hasConcept C6802819 @default.
- W1613527773 hasConcept C94523830 @default.
- W1613527773 hasConceptScore W1613527773C11413529 @default.
- W1613527773 hasConceptScore W1613527773C121332964 @default.
- W1613527773 hasConceptScore W1613527773C134306372 @default.
- W1613527773 hasConceptScore W1613527773C134978465 @default.
- W1613527773 hasConceptScore W1613527773C137559481 @default.
- W1613527773 hasConceptScore W1613527773C158693339 @default.
- W1613527773 hasConceptScore W1613527773C163716315 @default.
- W1613527773 hasConceptScore W1613527773C187834632 @default.
- W1613527773 hasConceptScore W1613527773C2524010 @default.
- W1613527773 hasConceptScore W1613527773C33923547 @default.
- W1613527773 hasConceptScore W1613527773C34727166 @default.
- W1613527773 hasConceptScore W1613527773C41008148 @default.
- W1613527773 hasConceptScore W1613527773C42355184 @default.
- W1613527773 hasConceptScore W1613527773C44363057 @default.
- W1613527773 hasConceptScore W1613527773C46085209 @default.
- W1613527773 hasConceptScore W1613527773C56372850 @default.
- W1613527773 hasConceptScore W1613527773C62520636 @default.
- W1613527773 hasConceptScore W1613527773C6802819 @default.
- W1613527773 hasConceptScore W1613527773C94523830 @default.
- W1613527773 hasLocation W16135277731 @default.
- W1613527773 hasOpenAccess W1613527773 @default.
- W1613527773 hasPrimaryLocation W16135277731 @default.
- W1613527773 hasRelatedWork W1521333973 @default.
- W1613527773 hasRelatedWork W187894598 @default.
- W1613527773 hasRelatedWork W1983836100 @default.
- W1613527773 hasRelatedWork W2005673224 @default.
- W1613527773 hasRelatedWork W2059454546 @default.
- W1613527773 hasRelatedWork W2068038816 @default.
- W1613527773 hasRelatedWork W2084098088 @default.
- W1613527773 hasRelatedWork W2099030560 @default.
- W1613527773 hasRelatedWork W2269629783 @default.
- W1613527773 hasRelatedWork W2299023367 @default.
- W1613527773 hasRelatedWork W2346928474 @default.
- W1613527773 hasRelatedWork W2493468169 @default.
- W1613527773 hasRelatedWork W2566608457 @default.
- W1613527773 hasRelatedWork W2621360859 @default.
- W1613527773 hasRelatedWork W2911184731 @default.
- W1613527773 hasRelatedWork W3039986651 @default.
- W1613527773 hasRelatedWork W3105227640 @default.
- W1613527773 hasRelatedWork W315355027 @default.
- W1613527773 hasRelatedWork W71300111 @default.
- W1613527773 hasRelatedWork W281564841 @default.
- W1613527773 isParatext "false" @default.
- W1613527773 isRetracted "false" @default.
- W1613527773 magId "1613527773" @default.
- W1613527773 workType "article" @default.