Performance modeling for CNN inference accelerators on FPGA

Computer Science

• Performance Model 100%
• Convolutional Neural Network 100%
• Network Inference 100%
• Field Programmable Gate Arrays 100%
• Parallelism 20%
• Data Communication 20%
• Performance Bottleneck 20%
• Design Phase 20%
• Memory Access 20%
• Energy Efficiency 20%
• Resource Utilisation 20%
• External Memory 20%
• Design Parameter 20%
• Implementation Phase 20%
• Artificial Neural Network 20%

Physics

• Convolutional Neural Network 100%
• Artificial Neural Network 33%
• Energy Efficiency 33%

Keyphrases

• On-chip Buffer 33%