AuroraFlow, an Easy-to-Use, Low-Latency FPGA Communication Solution Demonstrated on Multi-FPGA Neural Network Inference
Open Access
- 25 May 2025
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 35-48
Abstract
No abstract availableFunding Information
- German Federal Ministry for Environment, Nature Conservation, Nuclear Safety and Consumer Protection (67KI32004A)
This publication has 32 references indexed in Scilit:
- Optimizing Communication for Latency Sensitive HPC Applications on up to 48 FPGAs Using ACCL2024
- Extending Data Flow Architectures for Convolutional Neural Networks to Multiple FPGAs2023
- Trends in AI inference energy consumption: Beyond the performance-vs-parameter laws of deep learningSustainable Computing: Informatics and Systems, 2023
- In-depth FPGA accelerator performance evaluation with single node benchmarks from the HPC challenge benchmark suite for Intel and Xilinx FPGAs using OpenCLJournal of Parallel and Distributed Computing, 2022
- ACCL: FPGA-Accelerated Collectives over 100 Gbps TCP-IP2021
- EasyNet: 100 Gbps Network for HLS2021
- Fast inference of deep neural networks in FPGAs for particle physicsJournal of Instrumentation, 2018
- fpgaConvNet: A Framework for Mapping Convolutional Neural Networks on FPGAs2016
- Open MPI: A Flexible High Performance MPI2006
- SLURM: Simple Linux Utility for Resource Management2003