skip to main content

MySQL performance on Lenovo ThinkSystem SR665 V3

Solution Brief

Home
Top
Published
15 Dec 2025
Form Number
LP2347
PDF size
7 pages, 211 KB

Abstract

Modern enterprises face growing challenges in managing data-intensive workloads as digital transformation accelerates. To address these demands, Lenovo ThinkSystem SR665 V3 servers, powered by AMD EPYC™ 9004 (Genoa) and 9005 (Turin) processors, deliver exceptional compute density, memory bandwidth, and I/O performance.

This document evaluates MySQL performance on the Lenovo ThinkSystem SR665 V3 platform, comparing two generations of AMD EPYC processors using industry-standard benchmarks. The goal is to demonstrate how architectural advancements in Zen 5 improve throughput, scalability, and efficiency. This solution brief highlights performance testing of MySQL 8.0.93 on Lenovo ThinkSystem SR665 V3 servers equipped with AMD EPYC Genoa and Turin CPUs. Using HammerDB TPROC-C benchmarks, we measured throughput across varying user loads to assess scalability and efficiency. The server architecture supports AMD EPYC™ 5th Gen 9005 and 4th Gen 9004 series processors, delivering up to 192 cores and 384 threads per socket for increased parallelism in compute-intensive workloads. Memory performance is optimized with DDR5 DIMMs operating at speeds up to 6400 MT/s, enabling higher bandwidth and lower latency for demanding applications. It incorporates Compute Express Link (CXL) v1.1. For storage, the platform offers high-performance NVMe drives and a broad range of configurations to meet diverse capacity and throughput requirements. I/O performance is maximized through PCIe 5.0 support. Additionally, the integration of the OCP 3.0 NIC adapter delivers modular networking options with enhanced scalability and flexibility for evolving data center architectures.

This document aims to highlight the generational advancements in AMD’s EPYC processor lineup by comparing two CPUs from the same family: one based on the Genoa architecture and the other on Turin. By analyzing performance metrics collected using hammerDB for MySQL TPROC-C, we demonstrate the increased computational power introduced with the new generation.

Related product families

Product families related to this document are the following: