Comparison Report of Stress Testing on Different CPU Models

This topic has been translated from a Chinese forum by GPT and might contain errors.

Original topic: 不同型号cpu压测对比报告

| username: jingyesi3401

[TiDB Usage Environment] Production Environment, Testing

[TiDB Version] v6.1.5

[Encountered Problem: Phenomenon and Impact] Currently, the CPU model of the company’s operating server is as shown in Figure 1 (Intel Gold 6326 CPU (2.9GHz/16 cores)). Yesterday, we received a notification from the IT department stating that this server is currently out of stock and may take 3-5 months to restock. We need to evaluate the performance loss if we use the Intel Silver 4314 CPU (2.4GHz/16 cores). Does anyone have a sysbench benchmark comparison report for different CPU models for reference, so we can estimate a relatively accurate performance loss value?

Figure 1: Current operating server CPU model

Figure 2: Proposed replacement CPU model

| username: 像风一样的男子 | Original post link

Under the same number of cores, the CPU has little impact on database performance; the main factors are still disk and memory.

| username: Miracle | Original post link

TiDB seems to only have requirements for the number of cores, not the model. As long as there are enough cores, the model is not that important.

| username: xingzhenxiang | Original post link

The number of CPU cores doesn’t have much impact, but I noticed that the frequency has also decreased, which should have an effect.

| username: Fly-bird | Original post link

The higher the clock speed, the better the performance.

| username: 大飞哥online | Original post link

In terms of performance impact ranking, it should be disk, memory, CPU.

| username: cassblanca | Original post link

The CPU should no longer be a performance bottleneck now, right?

| username: GangShen | Original post link

You can refer to this article:

| username: system | Original post link

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.