KVBench

View as Markdown

KVBench is a benchmarking utility that generates and runs NIXLBench commands for profiling KV cache transfer performance across LLM architectures. The profile command invokes nixlbench as a subprocess, executing the transfer benchmarks that KVBench plans based on model architecture and access patterns.

Command Categories

KVBench Commands

  • plan — Display the recommended NIXLBench configuration for a given model architecture and access pattern
  • profile — Run NIXLBench with the planned configuration and collect performance results
  • kvcache — Display KV cache layout information for a model architecture

CTP Commands

  • ct-perftest — Run custom traffic performance tests using asymmetric transfer matrices
  • sequential-ct-perftest — Run sequential custom traffic performance tests for ordered matrix evaluation

Supported Models

KVBench includes model architecture definitions for several LLM families: DeepSeek R1, Llama 3.1, and more. See Commands and Examples for details on defining custom model architectures.

Next Steps