Measuring Performance of Transformer Inference

This chapter is divided into eight parts; they are: • Metrics for LLM Inference • Measuring a Single Request • Warmup and Synchronization • Measuring GPU Work with CUDA Events • Measuring Memory Usage • Measuring Concurrent Requests • Multiple GPUs and Multiple Machines • Cost per Token The most common inference metrics are: • Latency:  How long a request takes from start to finish.