Find answers to common questions about our VPS testing.
Feel free to contact us here.
Disk tests are run every 6 hours, CPU tests are run every 6 hours, and network tests are run every 12 hours. To decrease the network loads, we assign each VPS a random test time so that the network tests are not all carried out at the same time.
If you're a VPS provider, submitting your VPS allows us to test it and provide you with a performance score. This helps you showcase your VPS's capabilities to potential customers.
Submit your VPS using our form for evaluation; if it passes our review, we’ll add it to the listing. As another option, we can create an account on your platform and you can allocate a VPS to us post-registration.
We use the latest version of multiple linux distributions that support our testing software. Our preferred distribution is Ubuntu.
We do! (kind of) We've taken the best parts of YABS and other scripts to create our own custom script. Since we are benchmarking these servers multiple times a day we needed to make sure it was fast and efficient. So we cut out the fluff and kept the core elements to give you a quick and accurate benchmark.
We use a custom version of Taganaka's SpeedTest to test network performance. It features smart server discovery to automatically select the fastest nearby server, line profiling to choose the most suitable test based on connection speed, and aggressive multi-threading to fully saturate available bandwidth for accurate upload and download measurements.
We use fio, a powerful and flexible I/O benchmarking tool on Ubuntu, to measure disk performance, including read/write speeds, IOPS, and latency. It's well-suited for evaluating VPS storage by simulating real-world workloads and stress-testing drives through customizable scenarios.
To provide a single overall disk performance score, we weight the results from four block sizes: 40% 4K, 30% 64K, 20% 512K, and 10% 1M. This weighting reflects a balance between small random I/O and larger sequential transfers, offering a more representative measure of overall storage performance.
We use stress-ng to benchmark CPU performance for approximately 30 seconds every three hours. We record total CPU performance across all cores and then divide by the number of cores to produce an average per-core score.
Upon VPS submission, we change the root password, modify the SSH port, and then disable root access altogether for security. We also install Fail2Ban to help prevent brute-force attacks and may restrict access to our private VPN. These steps ensure consistent security and testing conditions.
We first average the results of all the VPSs listed on the site to create a baseline. Then we take the results of the VPS being tested and compare it to the baseline. The higher the score the better the performance.
Uptime is monitored every 60 seconds by Uptime Robot to quickly detect any downtime. This will be updated as we add more servers to our testing pool.
We do not accept payments in exchange for listings.
We only use the VPS for performance testing. We do not use them for any other purpose.
Yes! This is part of our testing process to ensure each provider is not trying to cheat the system! If a change is detected without prior notice from the provider we will remove the VPS from our site and blacklist them from future submissions.
MB (megabyte) measures data size (e.g., file size, storage). Mb (megabit) measures data transfer speed (e.g., network bandwidth).
Most VPS environments (KVM, QEMU, Xen, VMware, etc.) expose virtualized block devices (like vda, sda, etc.) to the guest OS. These are not the real physical drives. Most VPS see a "virtual disk" that behaves like a generic HDD or disk image. The hypervisor decides how to emulate the disk, and often sets ROTA=1 (rotational), even if the real disk is SSD or NVMe.
Similiar to the hard drive, most VPS environments (KVM, QEMU, Xen, VMware, etc.) expose virtualized RAM to the guest OS. These are not the real physical RAM. Most VPS see a "virtual RAM" that behaves like a generic RAM. The hypervisor decides how to emulate the RAM, and often sets the speed to 1000MHz.