An adaptive performance modeling tool for GPU architectures

Keyphrases

• GPU Kernel 100%
• Modeling Tools 100%
• Adaptive Performance 100%
• GPU Architecture 100%
• Performance Modeling 100%
• Microarchitecture 66%
• Performance Model 66%
• Performance Bottleneck 66%
• Analytical Model 33%
• Programmer 33%
• Data Parallelism 33%
• Execution Time 33%
• Compute Unified Device Architecture 33%
• Full System 33%
• System Complexity 33%
• Compiler 33%
• Memory Access 33%
• Control Divergence 33%
• Graph-based 33%
• Abstract Interpretation 33%
• Scratchpad Memory 33%
• Performance Information 33%
• NVIDIA GPU 33%
• Memory Bank Conflicts 33%
• Analytical Performance Models 33%
• Workflow Graph 33%
• Bank Control 33%
• Performance Trends 33%
• Conflict Flows 33%
• Parallel Benchmarks 33%
• Kernel Implementation 33%

Computer Science

• Adaptive Performance 100%
• Performance Model 100%
• Graphics Processing Unit 100%
• Microarchitecture 25%
• Performance Bottleneck 25%
• Execution Time 12%
• compute unified device architecture 12%
• Workflow 12%
• Analytical Model 12%
• Control Flow 12%
• Abstract Interpretation 12%
• Scratchpad Memory 12%
• System Complexity 12%
• Performance Information 12%
• Uncoalesced Memory Access 12%