Parallel S<i>P<sub>N</sub></i>on Multi-Core CPUS and Many-Core GPUS

Kirschenmann, Wilfried; Plagne, Laurent; Ponçot, Angélique; Vialle, Stéphane

doi:10.1080/00411450.2010.533741

Cited by 7 publications

(10 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The implementation of a neutron transport solver [6,8] with this version of Legolas++ will automaticaly provide a multi-target version of this solver. Efforts will be made to keep the portability of the performances currently available with MTPS.…”

Section: Discussionmentioning

confidence: 99%

“…The example presented in this section corresponds to a basic operation that represents the major part of the execution time of a neutron transport code [6,8]. Let A be a block-diagonal matrix with TDSM blocks.…”

Section: Implementation Of a Linear System Resolutionmentioning

confidence: 99%

“…The explicit GPU parallelization of a neutron transport code resulted in speed-ups around 30 over the sequential Legolas++ CPU implementation [6,8]. To generalize this gain of performances to other Legolas++ based applications, a parallel and multi-target version of Legolas++ is to be developed.…”

Section: Towards a Multi-target Linear Algebra Librarymentioning

confidence: 99%

“…For instance, Fig. 2 shows the sparsity pattern of a 5-level Legolas++ matrix block that belongs to the 7-level matrix of our neutron transport code [11,6,8].…”

Section: Towards a Multi-target Linear Algebra Librarymentioning

confidence: 99%

See 3 more Smart Citations

Multi-Target Vectorization with MTPS C++ Generic Library

Kirschenmann

Plagne

Vialle

2012

Applied Parallel and Scientific Computing

View full text Add to dashboard Cite

Abstract. This article introduces a C++ template library dedicated at vectorizing algorithms for different target architectures: Multi-Target Parallel Skeleton (MTPS). Skeletons describing the data structures and algorithms are provided and allow MTPS to generate a code with optimized memory access patterns for the choosen architecture. MTPS currently supports x86-64 multicore CPUs and CUDA enabled GPUs. On these architectures, performances close to hardware limits are observed.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Implementation Of a Linear System Resolutionmentioning

confidence: 99%

Section: Towards a Multi-target Linear Algebra Librarymentioning

confidence: 99%

“…For instance, Fig. 2 shows the sparsity pattern of a 5-level Legolas++ matrix block that belongs to the 7-level matrix of our neutron transport code [11,6,8].…”

Section: Towards a Multi-target Linear Algebra Librarymentioning

confidence: 99%

See 2 more Smart Citations

Multi-Target Vectorization with MTPS C++ Generic Library

Kirschenmann

Plagne

Vialle

2012

Applied Parallel and Scientific Computing

View full text Add to dashboard Cite

show abstract

“…This code relies on DIABOLO, a Simplified PN (SPN) method to compute the neutron flux inside the core for eigenvalue calculations. (1) In order to assess the accuracy of SPN results, a 3D Cartesian model of PWR nuclear cores has been designed and a reference neutron flux inside this core has been computed with the MCNP5 Monte Carlo transport code. (2) This kind of 3D whole core probabilistic evaluation of the flux is computationally very demanding.…”

Section: Introduction Industrial Contextmentioning

confidence: 99%

Shared memory parallelism for 3D Cartesian discrete ordinates solver

Moustafa

Dutka-Malen

Plagne

et al. 2015

Annals of Nuclear Energy

View full text Add to dashboard Cite

This paper describes the design and the performance of DOMINO, a 3D Cartesian SN solver that implements two nested levels of parallelism (multicore+SIMD) on shared memory computation nodes. DOMINO is written in C++,a multi-paradigm programming language that enables the use of powerful and generic parallel programming tools such as Intel TBB and Eigen. These two libraries allow us to combine multi-thread parallelism with vector operations in an efficient and yet portable way. As a result, DOMINO can exploit the full power of modern multi-core processors and is able to tackle very large simulations, that usually require large HPC clusters, using a single computing node. For example, DOMINO solves a 3D full core PWR eigenvalue problem involving 26 energy groups, 288 angular directions (S16), 46 ⇥ 10 6 spatial cells and 1 ⇥ 10 12 DoFs within 11 hours on a single 32-core SMP node. This represents a sustained performance of 235 GFlops and 40.74% of the SMP node peak perfomance for the DOMINO sweep implementation. The very high Flops/Watt ratio of DOMINO makes it a very interesting building block for a future many-nodes nuclear simulation tool.

show abstract