The impact of the communication on the performance of numerical algorithms increases with the number of cores. In the context of sparse linear systems of equations, solving Ax = b on a very large computer with thousands of nodes requires the minimization of the communication to achieve very high efficiency as well as low energy cost. The high level of sequentiality in the Incomplete LU factorization (ILU) makes it difficult to parallelize. We first introduce in this manuscript a Communication-Avoiding ILU preconditioner, denoted CA-ILU(k), that factors A in parallel and then is applied at each iteration of a solver as GMRES, both steps without communication. Considering a row block of A, the key idea is to gather all the required dependenci...
Texte intégral accessible uniquement aux membres de l'Université de LorraineThis dissertation treats...
We present block LU factorization with panel rank revealing pivoting (block LU PRRP), a decompositio...
We present block LU factorization with panel rank revealing pivoting (block LU_PRRP), a decompositio...
The impact of the communication on the performance of numerical algorithms increases with the number...
The impact of the communication on the performance of numerical algorithms increases with the number...
This dissertation focuses on a widely used linear algebra kernel to solve linear systems, that is th...
In this paper we present an algorithm for computing a low rank approximation of a sparse matrix base...
This paper presents CALU, a Communication Avoiding algorithm for the LU factorization of dense matri...
Abstract. Since the cost of communication (moving data) greatly exceeds the cost of doing arithmetic...
International audienceSince the cost of communication (moving data) greatly exceeds the cost of doin...
International audiencen this paper we present an algorithm for computing a low rank approximation of...
We introduce a parallel algorithm for computing the low rank approximation $A_k$ of a large matrix $...
Cette thèse traite d une routine d algèbre linéaire largement utilisée pour la résolution des systèm...
The solution of dense systems of linear equations is at the heart of numerical computations. Such sy...
Texte intégral accessible uniquement aux membres de l'Université de LorraineThis dissertation treats...
We present block LU factorization with panel rank revealing pivoting (block LU PRRP), a decompositio...
We present block LU factorization with panel rank revealing pivoting (block LU_PRRP), a decompositio...
The impact of the communication on the performance of numerical algorithms increases with the number...
The impact of the communication on the performance of numerical algorithms increases with the number...
This dissertation focuses on a widely used linear algebra kernel to solve linear systems, that is th...
In this paper we present an algorithm for computing a low rank approximation of a sparse matrix base...
This paper presents CALU, a Communication Avoiding algorithm for the LU factorization of dense matri...
Abstract. Since the cost of communication (moving data) greatly exceeds the cost of doing arithmetic...
International audienceSince the cost of communication (moving data) greatly exceeds the cost of doin...
International audiencen this paper we present an algorithm for computing a low rank approximation of...
We introduce a parallel algorithm for computing the low rank approximation $A_k$ of a large matrix $...
Cette thèse traite d une routine d algèbre linéaire largement utilisée pour la résolution des systèm...
The solution of dense systems of linear equations is at the heart of numerical computations. Such sy...
Texte intégral accessible uniquement aux membres de l'Université de LorraineThis dissertation treats...
We present block LU factorization with panel rank revealing pivoting (block LU PRRP), a decompositio...
We present block LU factorization with panel rank revealing pivoting (block LU_PRRP), a decompositio...