Parallel GEMM-based convolutions for deep learning on multicore ARM and RISC-V architectures
Impacto
![Google Scholar](/xmlui/themes/Mirage2/images/uji/logo_google.png)
![Microsoft Academico](/xmlui/themes/Mirage2/images/uji/logo_microsoft.png)
Metadatos
Mostrar el registro completo del ítemcomunitat-uji-handle:10234/9
comunitat-uji-handle2:10234/7036
comunitat-uji-handle3:10234/8620
comunitat-uji-handle4:
INVESTIGACIONMetadatos
Título
Parallel GEMM-based convolutions for deep learning on multicore ARM and RISC-V architecturesFecha de publicación
2024-05-24Editor
ElsevierISSN
1383-7621Cita bibliográfica
Martínez, Héctor, et al. "Parallel GEMM-based convolutions for deep learning on multicore ARM and RISC-V architectures." Journal of Systems Architecture (2024): 103186.Tipo de documento
info:eu-repo/semantics/articleVersión
info:eu-repo/semantics/acceptedVersionPalabras clave / Materias
Resumen
We present high performance, multi-threaded implementations of three GEMM-based convolution algorithms for multicore processors with ARM and RISC-V architectures. The codes are integrated into CONVLIB, a library that ... [+]
We present high performance, multi-threaded implementations of three GEMM-based convolution algorithms for multicore processors with ARM and RISC-V architectures. The codes are integrated into CONVLIB, a library that has the following unique features: (1) scripts to automatically generate a key component of GEMM, known as the micro-kernel, which is typically written in assembly language; (2) a modified analytical model to automatically tune the algorithms to the underlying cache architecture; (3) the ability to select four hyper-parameters: micro-kernel, cache parameters, parallel loop, and GEMM algorithm dynamically between calls to the library, without recompiling it; and (4) a driver to identify the best hyper-parameters. In addition, we provide a detailed performance evaluation of the convolution algorithms, on five ARM and RISC-V processors, and we publicly release the codes. [-]
Publicado en
Journal of Systems Architecture, 153 (2024) 103186Datos relacionados
Data will be made available on request.Entidad financiadora
MCIN/AEI/10.13039/501100011033 | Generalitat Valenciana | Junta de Andalucía | European Union ‘‘NextGenerationEU’’/PRTR | Universitat Jaume I
Código del proyecto o subvención
PID2020-113 656RB-C22 | PROMETEO 2023-CIPROM/2022/20 | POSTDOC_21_00025 | RYC2021-033973-I | UJI-2023-04
Derechos de acceso
1383-7621/© 2024 Published by Elsevier B.V.
info:eu-repo/semantics/embargoedAccess
info:eu-repo/semantics/embargoedAccess
Aparece en las colecciones
- ICC_Articles [427]