A Systolic Neural CPU Processor Combining Deep Learning and General-Purpose Computing with Enhanced Data Locality and End-to-End Performance

Keyphrases

• Deep Learning 100%
• General-purpose Computing 100%
• Heterogeneous Architectures 100%
• Data Locality 100%
• Enhanced Data 100%
• CPU Processor 100%
• Data Transfer 66%
• Neural Network Accelerator 66%
• Processing Element 66%
• RISC-V 66%
• Performance Improvement 33%
• Working Mode 33%
• Data Management 33%
• Power Efficiency 33%
• Image Classification 33%
• Processor Core 33%
• Area Overhead 33%
• Classification Task 33%
• Fifth Generation 33%
• Direct Memory Access 33%
• CPU Core 33%
• Reduced Instruction Set Computer 33%
• MNIST 33%
• Neural Processor 33%
• ImageNet Dataset 33%
• CIFAR-10 33%
• Deep Neural Network Accelerator 33%
• Deep Recurrent Neural Network 33%
• Machine Learning Tasks 33%
• Bi-directional Dataflow 33%
• Unified Architecture 33%

Computer Science

• Heterogeneous Architecture 100%
• Data Locality 100%
• Deep Learning Method 100%
• End Performance 100%
• Processing Element 66%
• Neural Network 66%
• Postprocessing 33%
• Image Classification 33%
• Power Efficiency 33%
• Direct Memory Access 33%
• Deep Neural Network 33%
• Reduced Instruction Set Computer 33%
• Data Management 33%
• Performance Improvement 33%
• Classification Task 33%
• Processor Core 33%
• Machine Learning 33%
• Learning System 33%