User Tools

Site Tools


start

ANR DEDALES (ANR‐14‐CE23‐0005)

Project DEDALES aims at developing high performance software for the simulation of two phase flow in porous media. The project will specifically target parallel computers where each node is itself composed of a large number of processing cores, such as are found in new generation many-core architectures.

The project will be driven by an application to radioactive waste deep geological disposal. Its main feature is phenomenological complexity: water-gas flow in highly heterogeneous medium, with widely varying space and time scales. The assessment of large scale model is of major importance and issue for this application, and realistic geological models have several million grid cells. Few, if at all, software codes provide the necessary physical features with massively parallel simulation capabilities. The aim of the DEDALES project is to study, and experiment with, new approaches to develop effective simulation tools with the capability to take advantage of modern computer architectures and their hierarchical structure.

To achieve this goal, we will explore two complementary software approaches that both match the hierarchical hardware architecture: on the one hand, we will integrate a hybrid parallel linear solver into an existing flow and transport code, and on the other hand, we will explore a two level approach with the outer level using (space time) domain decomposition, parallelized with a distributed memory approach, and the inner level as a subdomain solver that will exploit thread level parallelism.

Linear solvers have always been, and will continue to be, at the center of simulation codes. However, parallelizing implicit methods on unstructured meshes, such as are required to accurately represent the fine geological details of the heterogeneous media considered, is notoriously difficult. It has also been suggested that time level parallelism could be a useful avenue to provide an extra degree of parallelism, so as to exploit the very large number of computing elements that will be part of these next generation computers. Project DEDALES will show that space-time DD methods can provide this extra level, and can usefully be combined with parallel linear solvers at the subdomain level.

Work within the project is organized among three main tasks (in addition to a management task):

  1. Domain decomposition algorithms: here we will extend DD methods, already developed for one phase flow to the case of two phase flows. The main issue will be dealing with the non-linearity and the coupling between pressure and saturation. A prototype simulator, using a Matlab library, will be built.
  2. Development of parallel hybrid linear solvers: these are by nature hierarchical solvers, with a natural two level parallelism: a parallel direct factorization performed on each subproblem and the iterative method orchestrating the solution on the interfaces. This task will show that it is possible to combine the direct solver Pastix with a conjugate gradient method with a single programming model by exploiting the StarPU runtime system.
  3. Implementation on hierarchical architectures: here we will integrate the results of the first two tasks in order to build a two-phase flow code based on a domain decomposition method (and MPI) from Task 1, and on a version of the hybrid linear solver (and a thread model) from Task 2. This combination leads naturally to a method that can take advantage of two level parallelism.

For all tasks, realistic test cases aimed at showing the validity, and the parallel scalability of the chosen approach, will be used. The most demanding models will be at the frontier of what is currently feasible both for the size of models.

The goal of the DEDALES project is to run these examples one order of magnitude faster that what is currently feasible.

The consortium brings together research groups with expertise in numerical methods for porous media flow (Inria / Pomdapi), in domain decomposition (Pomdapi, Laga), in high performance linear algebra (Inria / Hiepacs,) in developing codes for various parallel platforms (Maison de la Simulation), together with a public agencies who has the need for, and the experience in running codes for demanding problems (Andra). Most partners in this project have a fruitful history of cooperation (through the previous ANR SHPCO2 project, the Inria-Andra agreement, or the Inria Project Lab on Exascale).

start.txt · Last modified: 2015/01/04 20:14 by mkern