How CPU and Memory Affect MySQL Performance and How to Choose Server Configurations

Spring Boot video tutorial by Song is now available; interested readers can follow the link.

Many factors affect MySQL performance; this article examines how CPU and memory impact MySQL and how to choose appropriate server configurations.

CPU

When selecting a database server, should you buy a multi‑core CPU or a high‑frequency CPU? Developers often face this question from procurement teams.

Budget constraints usually prevent having both, so understanding MySQL’s execution model is essential to make the right choice.

If the workload is CPU‑intensive (i.e., many complex queries), increasing the CPU clock speed matters more because MySQL processes each statement on a single core; adding more cores does not speed up a single query.

To improve overall concurrency, core count becomes important: each core can handle one query at a time, so ten cores can process ten queries simultaneously, raising QPS.

High‑traffic internet applications that rely on MySQL typically require high concurrency, making multi‑core CPUs the preferred option.

Older MySQL versions have limited multi‑core support, so using the latest MySQL release is advisable when opting for many cores.

Also avoid running a 32‑bit MySQL server on a 64‑bit CPU.

Memory

Memory access is much faster than disk I/O; caching data in RAM reduces disk reads and boosts database efficiency.

Memory influences database performance in two ways:

Read efficiency – retrieving data from RAM is faster than from disk.

Write efficiency – with ample memory, writes can be buffered and flushed in larger batches, reducing disk I/O.

However, adding memory blindly is not advisable; it should be based on MySQL’s operational characteristics.

For common storage engines:

MyISAM caches indexes in memory while the OS handles data storage.

InnoDB caches both data and indexes in memory, improving overall performance.

Thus, memory requirements differ between storage engines.

On production servers with large RAM (e.g., 128 GB) but relatively small data size (e.g., 50 GB), performance issues are unlikely to be solved by adding more memory; root cause analysis is needed.

When selecting memory modules, match the motherboard’s maximum supported frequency, avoid mixing brands, and ensure identical frequency, voltage, timing, and capacity across sticks; for cloud servers, these concerns are less relevant.

This concludes the brief discussion; other MySQL performance factors will be covered in future articles.

Recommended Reading:

How to Design Redis Caching for Billion‑Scale Systems

10 Design Principles to Move Closer to an Architect Role

Spring Boot Integration with Kafka

Welcome to follow the public WeChat account "Internet Full‑Stack Architecture" for more valuable information.