Question

CPU-optimized much slower than basic/shared CPU

I had some performance issues on a server where the CPU was over utilized. It had a 2 vCPU / 2GB RAM Basic - Premium Intel and the load was very CPU intensive. So I figured that a CPU-optimized droplet would be the best choice. I chose a CPU-optimized droplet with 2 dedicated vCPUs and 4GB RAM. After switching the performance problems did not get any better, and probably even worse.

So I figured that I should do a benchmark test om both machines (with minimal load of course). I used sysbench (https://github.com/akopytov/sysbench#linux) and here is the results:

Events per second:

  • Basic/shared CPU: about 4100
  • CPU-optimized/Dedicated CPU: about 1950

I have run the tests several times with different settings, but the results are pretty consistent. The basic droplet is also about twice as fast when it comes to the latency.

The basic droplets is more than twice as fast! For half the price! What is going on here? This can’t be normal…


Submit an answer

This textbox defaults to using Markdown to format your answer.

You can type !ref in this text area to quickly search our full set of tutorials, documentation & marketplace offerings and insert the link!

Sign In or Sign Up to Answer

These answers are provided by our Community. If you find them useful, show some love by clicking the heart. If you run into issues leave a comment, or add your own answer to help others.

Hi there,

When it comes to Basic Droplets, it’s a case of your mileage may vary depending on your neighbors. For example, if your Droplet was created on a new hypervisor that’s empty and there are just a couple of Droplets there, the performance may well match up with a dedicated server. But as traffic on the HV builds up over time, the performance will likely deteriorate based on the workloads of all tenants. A CPU-Optimized Droplet will likely have more reliable performance throughout as you would have dedicated CPUs no matter the other tenants on the same hypervisor.

That being said, you’ve mentioned that you ran multiple tests under different settings and still continued to see a significant difference, which is unusual behavior. What you could do is reach out to the DigitalOcean support team who might be able to take a look at your Droplets and see if they can look at the underlying hypervisor and any traffic to see if they have any insights.

https://www.digitalocean.com/support/

Hope that helps!

- Bobby.