Optimizing CPU Performance for Recommendation Systems At-Scale

Keyphrases

• Recommendation System 100%
• CPU Utilization 100%
• Stall 66%
• Recommendation Model 66%
• Caching 33%
• Popular 33%
• Major Contributor 33%
• Memory Access 33%
• Execution Time 33%
• High Latency 33%
• Integrated Technique 33%
• Personalized Recommendation 33%
• New Recommendations 33%
• Off-chip Memory Access 33%
• Memory Access Latency 33%
• Memory Bandwidth 33%
• Model Inference 33%
• Software Prefetching 33%
• Production Trace 33%
• Hyper-threading 33%
• Core Configuration 33%

Computer Science

• Deep Learning Method 100%
• Memory Access 66%
• Speed-up 33%
• Data Center 33%
• Execution Time 33%
• prefetching 33%
• Memory Bandwidth 33%
• Inference Model 33%
• Memory Access Latency 33%
• Chip Memory Access 33%
• Artificial Intelligence 33%
• Graphics Processing Unit 33%
• Parameter Size 33%