Instruction-Based Coordination of Heterogeneous Processing Units for Acceleration of DNN Inference

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Petropoulos, Anastasios, Antonakopoulos, Theodore
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866915706225819648
author Petropoulos, Anastasios
Antonakopoulos, Theodore
author_facet Petropoulos, Anastasios
Antonakopoulos, Theodore
contents This paper presents an instruction-based coordination architecture for Field-Programmable Gate Array (FPGA)-based systems with multiple high-performance Processing Units (PUs) for accelerating Deep Neural Network (DNN) inference. This architecture enables programmable multi-PU synchronization through instruction controller units coupled with peer-to-peer instruction synchronization units, utilizing instruction types organized into load, compute, and store functional groups. A compilation framework is presented that transforms DNN models into executable instruction programs, enabling flexible partitioning of DNN models into topologically contiguous subgraphs mapped to available PUs. Multiple deployment strategies are supported, enabling pipeline parallelism among PUs and batch-level parallelism across different PU subsets, with runtime switching among them without FPGA reconfiguration. The proposed approach enables design space exploration, supporting dynamic trade-offs between single-batch and multi-batch performance. Experimental results on ResNet-50 demonstrate notable compute efficiency, up to $98\%$, and throughput efficiency gains, up to $2.7\times$, over prior works across different configurations.
format Preprint
id arxiv_https___arxiv_org_abs_2511_15505
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Instruction-Based Coordination of Heterogeneous Processing Units for Acceleration of DNN Inference
Petropoulos, Anastasios
Antonakopoulos, Theodore
Hardware Architecture
This paper presents an instruction-based coordination architecture for Field-Programmable Gate Array (FPGA)-based systems with multiple high-performance Processing Units (PUs) for accelerating Deep Neural Network (DNN) inference. This architecture enables programmable multi-PU synchronization through instruction controller units coupled with peer-to-peer instruction synchronization units, utilizing instruction types organized into load, compute, and store functional groups. A compilation framework is presented that transforms DNN models into executable instruction programs, enabling flexible partitioning of DNN models into topologically contiguous subgraphs mapped to available PUs. Multiple deployment strategies are supported, enabling pipeline parallelism among PUs and batch-level parallelism across different PU subsets, with runtime switching among them without FPGA reconfiguration. The proposed approach enables design space exploration, supporting dynamic trade-offs between single-batch and multi-batch performance. Experimental results on ResNet-50 demonstrate notable compute efficiency, up to $98\%$, and throughput efficiency gains, up to $2.7\times$, over prior works across different configurations.
title Instruction-Based Coordination of Heterogeneous Processing Units for Acceleration of DNN Inference
topic Hardware Architecture
url https://arxiv.org/abs/2511.15505