In this work, we present Dido, an implicitly parallel domain-specific language (DSL) that captures high-level stencil abstractions and automatically generates high-performance parallel stencil code for distributed-memory architectures. The generated code uses ORWL as a communication and synchronization backend. We show that Dido achieves a huge progress in terms of programmer productivity without sacrificing the performance. Dido supports a wide range of stencil computations and real-world stencil-based applications. We show that the well-structured code generated by Dido lends itself to different possible optimizations and study the performance of two of them. We also combine Dido's code generation technique with the polyhedral loop optimi...
Communicated by Guest Editors Our aim is to apply program transformations to stencil codes in order ...
Code transformations, such as loop tiling and loop fusion, are of key importance for the efficient i...
Parallel architectures have now reached every computing device, but software developers generally la...
In this work, we present Dido, an implicitly parallel domain-specific language (DSL) that captures h...
International audienceWe present a source-to-source auto-generating framework that enables alarge pr...
Abstract Performance optimization of stencil computations has beenwidely studied in the literature, ...
Application codes reliably achieve performance far less than the advertised capabilities of existing...
The key common bottleneck in most stencil codes is data movement, and prior research has shown that ...
Application codes reliably achieve performance far less than the advertised capabilities of existing...
Communicated by Guest Editors The implementation of stencil computations on modern, massively parall...
Performance optimization of stencil computations has been widely studied in the literature, since th...
Les architectures parallèles sont désormais présentes dans tous les matériels informatiques, mais le...
The implementation of stencil computations on modern, massively parallel systems with GPUs and other...
The natural and the design limitations of the evolution of processors, e.g., frequency scaling and m...
This paper describes a new technique for optimizing serial and parallel stencil- and stencil-like op...
Communicated by Guest Editors Our aim is to apply program transformations to stencil codes in order ...
Code transformations, such as loop tiling and loop fusion, are of key importance for the efficient i...
Parallel architectures have now reached every computing device, but software developers generally la...
In this work, we present Dido, an implicitly parallel domain-specific language (DSL) that captures h...
International audienceWe present a source-to-source auto-generating framework that enables alarge pr...
Abstract Performance optimization of stencil computations has beenwidely studied in the literature, ...
Application codes reliably achieve performance far less than the advertised capabilities of existing...
The key common bottleneck in most stencil codes is data movement, and prior research has shown that ...
Application codes reliably achieve performance far less than the advertised capabilities of existing...
Communicated by Guest Editors The implementation of stencil computations on modern, massively parall...
Performance optimization of stencil computations has been widely studied in the literature, since th...
Les architectures parallèles sont désormais présentes dans tous les matériels informatiques, mais le...
The implementation of stencil computations on modern, massively parallel systems with GPUs and other...
The natural and the design limitations of the evolution of processors, e.g., frequency scaling and m...
This paper describes a new technique for optimizing serial and parallel stencil- and stencil-like op...
Communicated by Guest Editors Our aim is to apply program transformations to stencil codes in order ...
Code transformations, such as loop tiling and loop fusion, are of key importance for the efficient i...
Parallel architectures have now reached every computing device, but software developers generally la...