We’ve now reached the midpoint of week-long benchmarking series. Our first two articles discussed memory (RAMSpeed) and storage (Flexible IO Tester). Today, we’ll be focusing on a third element defining the performance of a cloud server: CPU.
The central processor is the brain of your server; most calculations take place in this component of the infrastructure. No wonder then that its efficiency is considered vital to a server’s overall performance.
What defines a good processor? GHz and core numbers, obviously. But it’s also defined by the amount of different cache levels and underlying infrastructure. A well-rounded CPU will perform best, even when certain specific tasks are more heavily impacted by other components of the server such as RAM.
To test overall CPU performance, you have to measure the speed at which it processes workloads. If a provider got good results in our two previous tests, and can get a good grade in this test too, then you’re probably looking at cloud servers that can guarantee decent performance for any workload.
Day 3 Benchmark:
The C-RAY benchmark is a simple raytracer that tests the floating point CPU performance. It is a multi-threaded test (16 threads per core). It’ll shoot 8 rays per pixel for anti-aliasing, generating a 1 600 x 1 200 image.
The test only requires a small amount of data, ensuring the system does not have to access RAM to complete the task. In doing so, only CPU is tested and the results are not influenced by other components of the server. This makes this benchmark an ideal one for testing processor speed.
In this benchmark, the results returned are the seconds it takes to process the workload: in other words, the smaller the number the better. In the graph above, you’ll see that SherWeb’s Performance Cloud Servers were the fastest, needing only 67 seconds to complete the test.
CPU speed will have a big impact on your overall performance.
The other providers needed 42% to 99% more time to complete the test. While Digital Ocean and Rackspace were roughly in the same ballpark, Microsoft— and Amazon in particular—seriously underperformed. In this test, Amazon was running AMD Opteron 4171 HE processors at 2.09 GHz. All the other providers were using Intel processors.
Having good CPU performance will mean your servers can juggle multiple tasks more seamlessly. For example, tasks can be interrupted, run simultaneously, and there can be numerous asynchronous events. And if the CPU’s performance is up to par, you won’t overload your system.
This level of CPU performance can come in handy if multiple users or applications are requesting resources from the server. The server will be better able to serve the numerous requests without delay. Keep in mind, though, that slow processing might not always be the CPU’s fault. In the case of a database server running MySQL, RAM will be more important than CPU speed. Also, if you want to add CPUs to your server, you need to also run the adequate amount of RAM. For example, SherWeb offers up 128 GB of RAM for 20 CPU cores.
I hope this third benchmarking article has shed some light on what goes into a high performance cloud server. Starting tomorrow, we’ll take a closer look at practical benchmarks to close off the week.
I also encourage you to test your workloads on SherWeb’s Performance Cloud Servers. Signing up takes only 2 minutes, there are no contracts, and you’ll only pay for what you use.
Sell all of our benchmarks!
Day 1: Cloud Servers Benchmarks – Day 1 (RAMSpeed)
Day 2: Cloud Servers Benchmarks – Day 2 (Flexible IO Tester)
Day 3: Cloud Servers Benchmarks – Day 3 (C-Ray)
Day 4: Cloud Servers Benchmarks – Day 4 (FFmpeg)
Day 5: Cloud Servers Benchmarks – Day 5 (Apache HTTP Server)