Reactive NUCA: Near-optimal block placement and replication in distributed caches

Keyphrases

• Near-optimal 100%
• Cache Design 100%
• Block Placement 100%
• Distributed Cache 100%
• Optimal Block Size 100%
• On chip 50%
• Caching 50%
• Chip-last 50%
• On-chip Communication 50%
• Scientific Workloads 50%
• Last-level Cache 50%
• Server Load 50%
• Latency 25%
• Performance Improvement 25%
• Lower Average 25%
• Conventional Design 25%
• Operating System 25%
• On-chip Memory 25%
• Core Count 25%
• Cache Capacity 25%
• Intelligent Placement 25%
• Multi-programmed Workloads 25%
• Private Cache 25%
• Shared Cache 25%
• Average Latency 25%
• Block Placement Policy 25%
• Multicore Processors 25%
• Cache Access Pattern 25%

Computer Science

• Working Set 100%
• Communication Delay 100%
• Access Pattern 50%
• Average Latency 50%
• Multicore Processor 50%
• Competing Design 50%
• Operating Systems 50%