Dissecting GPU Memory Hierarchy Through Microbenchmarking

Keyphrases

• Access Efficiency 10%
• Access Latency 10%
• Bank Conflict 10%
• Cache System 10%
• Caching 10%
• Computational Power 10%
• Data Cache 10%
• Global Memory 10%
• Graphics Processing Unit 100%
• Johannes Kepler 20%
• Maxwell 30%
• Memory Access 10%
• Memory Hierarchy 100%
• Memory Performance 10%
• Micro-benchmarks 10%
• Microbenchmarking 100%
• NVIDIA 10%
• Shared Memory 20%
• Software Modeling 10%
• Software Optimization 10%
• Texture Memory 10%
• Three-generation 10%
• Translation Lookaside Buffer 10%
• Unknown Characteristics 10%

Computer Science

• Computational Power 10%
• Graphics Processing Unit 100%
• Memory Access 10%
• Memory Hierarchy 100%
• Memory Performance 10%
• Microbenchmarks 10%
• Optimization Software 10%
• Shared Memory 20%
• Translation Lookaside Buffer 10%