In the context of the resolution of Maxwell's equation on 3D surface meshes, it is necessary to use algebraic solvers that are costly in time and memory. To reduce these costs, an improvement has been to apply hierarchical compression techniques on matrices. This kind of matrix allows the required ressources to be drastically reduced, but induce an important imbalance between the load of the different processes involved in the computation. This imbalance is a limiting factor for the scalability of simulation codes with a large number of computational nodes.In this thesis, we have been looking for reducing this imbalance, by using both static and dynamic load balancing methods.The static load balancing methods have allowed us to change the ...
The Boundary Element Method (BEM) requires the solving of large, ill-conditionned, dense linear syst...
In heterogeneous distributed environments such as the Grid, the available resources, namely the netw...
This paper is aimed at designing efficient parallel matrix-product algorithms for heterogeneous mast...
In the context of the resolution of Maxwell's equation on 3D surface meshes, it is necessary to use ...
Dans le cadre de la résolution des équations de Maxwell sur des maillages 3D surfaciques, il est néc...
Dans le cadre de la résolution des équations de Maxwell sur des maillages 3D surfaciques, il est néc...
In this paper, we describe and evaluate an extension of the Chameleon library to operate with hierar...
In this paper, we describe and evaluate an extension of the Chameleon library to operate with hierar...
Matrices and tensors are amongst the most common tools to represent and exploit information. Some so...
Dans cette thèse, nous nous sommes intéressée à la mise en oeuvre d'algorithmes itératifs sur des gr...
In this paper, we describe and evaluate an extension of the Chameleon library to operate with hierar...
The Boundary Element Method (BEM) requires the solving of large, ill-conditionned, dense linear syst...
International audienceWe present the principles of a parallel approach suited to the simulation of a...
International audienceWe present the principles of a parallel approach suited to the simulation of a...
The Boundary Element Method (BEM) requires the solving of large, ill-conditionned, dense linear syst...
The Boundary Element Method (BEM) requires the solving of large, ill-conditionned, dense linear syst...
In heterogeneous distributed environments such as the Grid, the available resources, namely the netw...
This paper is aimed at designing efficient parallel matrix-product algorithms for heterogeneous mast...
In the context of the resolution of Maxwell's equation on 3D surface meshes, it is necessary to use ...
Dans le cadre de la résolution des équations de Maxwell sur des maillages 3D surfaciques, il est néc...
Dans le cadre de la résolution des équations de Maxwell sur des maillages 3D surfaciques, il est néc...
In this paper, we describe and evaluate an extension of the Chameleon library to operate with hierar...
In this paper, we describe and evaluate an extension of the Chameleon library to operate with hierar...
Matrices and tensors are amongst the most common tools to represent and exploit information. Some so...
Dans cette thèse, nous nous sommes intéressée à la mise en oeuvre d'algorithmes itératifs sur des gr...
In this paper, we describe and evaluate an extension of the Chameleon library to operate with hierar...
The Boundary Element Method (BEM) requires the solving of large, ill-conditionned, dense linear syst...
International audienceWe present the principles of a parallel approach suited to the simulation of a...
International audienceWe present the principles of a parallel approach suited to the simulation of a...
The Boundary Element Method (BEM) requires the solving of large, ill-conditionned, dense linear syst...
The Boundary Element Method (BEM) requires the solving of large, ill-conditionned, dense linear syst...
In heterogeneous distributed environments such as the Grid, the available resources, namely the netw...
This paper is aimed at designing efficient parallel matrix-product algorithms for heterogeneous mast...