VPS - Second Benchmark Test

Recent benchmark tests of several virtual servers.

VPS - Second Benchmark Test-heroimage


Please note that initially, I wrote this blog post in German. This translation is for your convenience. Although every effort has been made to ensure accuracy, there may be translation errors. I apologize for any discrepancies or misunderstandings resulting from the translation. I am grateful for any corrections in the comments or via mail.

It’s been a while since my last benchmark test. I ran another one at the end of 2021 and would like to share my results (I somehow forgot to do it earlier). As I wrote last time, the performance of such a test is not the only criterion for evaluating a server or provider. It is also a snapshot in time.

I tested the following servers.

1bluVPS R84x EPYC 75028 GiB
ContaboVPS S SSD4x Xeon E5-26307.8 GiB
Hetzner*CX111x Xeon1.89 GB
Netcup*VPS 200 G81x QEMU Virtual CPU1.94 GB20 GB SSD
Netcup*RS Black2x AMD EPYC 7702P11.7 GB160 GB SSD
RackNerd*LEB Special 3.5 KVM3x Xeon E5-26903.35 GB45 GB SSD

* These are affiliate links. I receive a commission if a purchase is made via my link. There is no disadvantage for the visitor. At Hetzner, new customers even get 20€ cloud credit via the link. For Netcup, I have new customer vouchers here. The providers did not pay me to make this comparison.

The servers are mainly located in Germany. I rented the CX11 from Hetzner three times, once in Nuremberg, once in Falkenstein and once in Helsinki (Finland).

Of course, this does not compare completely similar servers, but the result may still be enjoyable. All servers in this list cost no more than €10 per month.

Test execution

I use Ubuntu 20.04 on the servers. As a test, I use Yet-Another-Bench-Script. This tests the write and read speed of the disks, the Internet speed, and CPU performance with the help of Geekbench.

Test result


The first test that is carried out is FIO. The write and read speed is tested with different block sizes.


Results in MB/s (more is better).

The Hetzner VPS in Helsinki was the fastest here. Strangely, the Netcup VPS was faster than the root server.


Results in GB/s (more is better).

This time the Hetzner VPS in Falkenstein achieved the highest speed.


Results in GB/s (more is better).

The Netcup servers perform best from a block size of 512k. The VPS R8 from 1blu and the VPS S SSD from Contabo are unfortunately far behind here.


Results in GB/s (more is better).

The results are similar here.

Result FIO


The next test is iperf, a TCP, UDP and SCTP speed test tool.

The servers were connected to the following providers during the test:

ClouviderLondon, UK
Online.netParis, FR
ClouviderNYC, NY, US
Velocity OnlineTallahassee, FL, US
ClouviderLos Angeles, CA, US
Iveloz TelecomSao Paulo, BR

Both download and upload speeds were measured with IPv4. Figures are in Gbit/s:

London, UK

Paris, FR


Tallahassee, FL, US

Los Angeles, CA, US

Sao Paulo, BR

Result iperf

The Hetzner VPSs are far ahead in this area. The disadvantage of the Hetzner cloud servers is that additional costs are incurred as soon as the monthly outgoing data exceeds 20TB. However, 20TB should be more than enough for many purposes.

The other servers are slower, but according to Speedtest, the average internet speed in Germany in January 2022 was 69.52 Mbit/s for downloads and 23.16 Mbit/s for uploads. Almost all results are higher than these values.


The latest test is Geekbench.

Geekbench 4

Geekbench 5

Geekbench result

Percentage comparison

Netcup’s root server is ahead in this test. In the single-core values, the Netcup VPS follows, followed by the Hetzner VPS and the VPS from 1blu. In the multi-core test, the VPSs with only one CPU core naturally perform worse. Here, the Contabo and 1blu servers with four virtual cores have the most significant advantage. However, it cannot quite match the dedicated CPU performance of the root server with 2 cores.

This website uses cookies. These are necessary for the functionality of the website. You can find more information in the privacy policy