Título:
|
Reducing data movement on large shared memory systems by exploiting computation dependencies
|
Autor/a:
|
Barrera, I.S.; Ayguadé Parra, Eduard; Valero Cortés, Mateo; Moreto Planas, Miquel; Labarta Mancho, Jesús José; Casas Guix, Marc
|
Otros autores:
|
Universitat Politècnica de Catalunya. Departament d'Arquitectura de Computadors; Universitat Politècnica de Catalunya. CAP - Grup de Computació d'Altes Prestacions |
Abstract:
|
Shared memory systems are becoming increasingly complex as they typically integrate several storage devices. That brings different access latencies or bandwidth rates depending on the proximity between the cores where memory accesses are issued and the storage devices containing the requested data. In this context, techniques to manage and mitigate non-uniform memory access (NUMA) effects consist in migrating threads, memory pages or both and are generally applied by the system software. We propose techniques at the runtime system level to further mitigate the impact of NUMA effects on parallel applications' performance. We leverage runtime system metadata expressed in terms of a task dependency graph, where nodes are pieces of serial code and edges are control or data dependencies between them, to efficiently reduce data transfers. Our approach, based on graph partitioning, adds negligible overhead and is able to provide performance improvements up to 1.52× and average improvements of 1.12× with respect to the best state-of-the-art approach when deployed on a 288-core shared-memory system. Our approach reduces the coherence traffic by 2.28× on average with respect to the state-of-the-art. |
Abstract:
|
Peer Reviewed |
Materia(s):
|
-Àrees temàtiques de la UPC::Informàtica::Sistemes d'informació::Emmagatzematge i recuperació de la informació -Parallel programming (Computer science) -NUMA -Scheduling -Shared memory -Task-based programming model
Data transfer -Graph theory -Intelligent control -Memory architecture -Scheduling -Virtual storage -Graph Partitioning -Non uniform memory access -NUMA -Parallel application -Performance improvements -Shared memory -Shared memory system -Task-based programming -Data reduction -Programació en paral·lel (Informàtica) |
Derechos:
|
|
Tipo de documento:
|
Artículo - Versión publicada Objeto de conferencia |
Editor:
|
Association for Computing Machinery (ACM)
|
Compartir:
|
|