Introduction

Test Objectives

This chapter describes the Kunpeng Automatic Tuning tests in single-application exclusive scenarios. It summarizes the performance-sensitive parameters and recommended values for key components, including various database systems, big data compute engines, AI inference systems, and the Java platform. It covers the following parameters:

• Service parameters of the application
• Parameters of dependency components
• Key parameters of the OS and kernel

This chapter aims to provide R&D personnel, O&M engineers, and performance tuning practitioners with reference and guidance for performance tuning based on test data, helping them identify performance bottlenecks and improve system metrics and stability.

Statement and Applicability

The recommended parameter settings in this chapter are based on internal test results from a single-application exclusive test environment Perform full verification and necessary adjustment based on your specific service scenarios, hardware conditions, and load characteristics.

In multi-application collaborative or more complex runtime environments, this chapter still offers valuable reference and guidance. Its core purpose is to help readers identify performance-sensitive parameters at various levels using test data and highlight the key areas and breakthrough directions for tuning.

If environmental conditions allow, it is recommended to use the Kunpeng Automatic Tuning tool to tune performance, obtain a parameter configuration that better matches your actual runtime environment, and achieve improved results.