PERFORMANCE LAB
VPS performance
by configuration
Compare compute, NVMe storage and network performance for Linux VPS and Windows RDP.
Choose a tested plan
Each configuration includes its own report and run data.
Benchmark report
Download the individual runs, median values and test configuration.
- Report ID
- GA-20261112-LINUX-PRO
- Date
- Tested in
- Amsterdam
- Operating system
- Ubuntu 24.04 LTS
- Individual runs
- 5
Compare tested plans
Swipe the table to compare every configuration.
| Metric | Linux Plus | Linux Pro | Linux Max |
|---|---|---|---|
| Single-thread CPU | 3,030 MIPS | 3,240 MIPS | 3,380 MIPS |
| Multi-thread CPU | 11,130 MIPS | 16,930 MIPS | 42,510 MIPS |
| Sequential read | 5,360 MB/s | 6,770 MB/s | 7,600 MB/s |
| Sequential write | 4,150 MB/s | 5,200 MB/s | 6,440 MB/s |
| Random read | 221,600 IOPS | 292,200 IOPS | 432,200 IOPS |
| Mean random read latency | 0.095 ms | 0.080 ms | 0.063 ms |
| Download | 941 Mbps | 950 Mbps | 1,920 Mbps |
| Upload | 921 Mbps | 936 Mbps | 1,882 Mbps |
How the tests are run
Compare like for like. The plan, image, region and date are listed in each report.
CPU
The built-in 7-Zip benchmark uses a 32 MiB dictionary. It runs with one thread and with the full vCPU count. The score is the total rating in MIPS.
Storage
fio uses a dedicated 16 GiB test file, direct I/O and a queue depth of 32. Sequential transfers use 1 MiB blocks. Random reads use 4 KiB blocks. The latency figure is the mean total latency of the random read workload.
Network
iPerf2 uses four parallel TCP streams for 30 seconds in each direction against a test peer in the same region.
Reading the results
Each headline number is the median of five runs for that metric. The files contain all five values. Actual results vary by workload, location and operating system.