The hardware that your website is hosted on is super important to how fast your web pages load. If you’re someone who is as obsessed with web performance as I am and are considering cloud VPS hosting, this blog post is perfect for you.
My goal here is to compare the performance differences between the Vultr Cloud Compute and Vultr High Frequency Compute plans. Specifically, I compared the difference between the $5/month Cloud Compute plan and the $6/month High Frequency plan to see if the extra $1/month is worth it.
By looking at the hardware and running my own site speed tests with a basic WordPress website, I was able to determine the performance difference between these two VPS hosting plans.
Vultr Hardware Differences
After deploying a $5 Cloud Compute and $6 High Frequency server in Atlanta with identical WordPress websites, I ran some basic Linux commands to see what kind of hardware was under the hood.
|Cloud Compute||High Frequency|
|CPU Clock Rate||2.399 GHz||3.792 GHz|
|SSD Cached Reads||9.258 GB/s||11.806 GB/s|
|SSD Buffered Disk Reads||386 MB/s||913 MB/s|
|Geekbench Single-Core Memory Score||2734||5622|
The CPU clock rate was determined with the
lscpu | grep MHz command. As you can see, the clock rate for the High Frequency plan is over 50% faster than the Cloud Compute plan.
To get the SSD metrics in the table above, I averaged out the results from three runs of the
hdparm -tT /dev/vda command. Again, the High Frequency plan yielded faster results than the Cloud Compute plan.
As for the RAM memory score, any commands I ran on the servers came back with unknown, so these Geekbench scores are from Vultr’s website.
Real World Performance Speed Tests
Next up, I deployed another Vultr test server in Atlanta and installed a set of open source tools to measure the performance of web sites called Sitespeed.io.
I then proceeded to run a series of 100 back-to-back page speed tests to both the Cloud Compute server and the High Frequency server. The speed tests were throttled to simulate a 3G connection. This smooths out any network hiccups.
Keep in mind that these two servers and WordPress websites are completely identical except for the hardware they run on. Both servers were running:
- WordPress 5.4.1
- PHP 7.2.24
- No SSL certificate
- IP address (no domain name)
- No CDN
So with a level playing field, these real world speed tests should give us a good idea about the performance difference between these two Vultr plans. Here are the speed test results.
|Cloud Compute||High Frequency|
|Backend Time||337 ms||331 ms|
|Server Response Time||371 ms||367 ms|
|First Paint||1.720 s||1.720 s|
|Page Load Time||3.074 s||3.070 s|
As you can see in the table above, the High Frequency plan beat out the Cloud Compute plan in all metrics except for the time to first paint which resulted in a tie. On average, the High Frequency server performed up to 2% faster than the Cloud Compute server.
Is High Frequency Worth It?
Now the question becomes, is a 2% performance difference worth paying 20% more per month?
For most websites, I would say yes.
In my opinion, every single millisecond counts when it comes to page speed. By having your website on a faster web server, you are not only giving your users a better experience, but you’re helping your website rank higher in Google search results.
Page speed is now a ranking factor for mobile search! ?
— Ilya Grigorik (@igrigorik) July 9, 2018
As always, if you have any questions about these two Vultr plans or web hosting in general, let me know in the comments below. Also, get up to $100 of free Vultr credit with my referral link! Note that the credit is applied after signup.