Skip to Main content Skip to Navigation
Conference papers

Towards Chapel-based Exascale Tree Search Algorithms: dealing with multiple GPU accelerators

Tiago Carneiro 1 Nouredine Melab 1 Akihiro Hayashi 2 Vivek Sarkar 2
1 BONUS - Optimisation de grande taille et calcul large échelle
Inria Lille - Nord Europe, CRIStAL - Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189
Abstract : Tree-based search algorithms applied to combinatorial optimization problems are highly irregular and time consuming when it comes to solving big instances. Solving such instances efficiently requires the use of massively parallel distributed-memory supercomputers. According to recent Top 500 trends, the degree of parallelism in these supercomputers continues to increase in size and complexity, with millions of heterogeneous (mainly CPU-GPU) cores. Harnessing this scale of computing resources raises at least three challenging issues which are described and addressed in this paper. Indeed, as a step towards exascale computing, we revisit the design and implementation of tree search algorithms dealing with multiple GPUs, in addition to scalability and productivity-awareness using Chapel. The proposed algorithm exploits Chapel's distributed iterators by combining a partial search strategy with pre-compiled CUDA kernels for more efficient exploitation of the intra-node parallelism. Extensive experimentation on big N-Queens problem instances using 24 GPUs shows that up to 90% of the linear speedup can be achieved.
Document type :
Conference papers
Complete list of metadata

https://hal.archives-ouvertes.fr/hal-03149394
Contributor : Tiago Carneiro Pessoa <>
Submitted on : Tuesday, February 23, 2021 - 8:14:46 AM
Last modification on : Tuesday, March 16, 2021 - 3:04:05 PM

File

final_hpcs2020.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-03149394, version 1

Collections

Citation

Tiago Carneiro, Nouredine Melab, Akihiro Hayashi, Vivek Sarkar. Towards Chapel-based Exascale Tree Search Algorithms: dealing with multiple GPU accelerators. HPCS 2020 - The 18th International Conference on High Performance Computing & Simulation, Mar 2021, Barcelona / Virtual, Spain. ⟨hal-03149394⟩

Share

Metrics

Record views

67

Files downloads

70