VPS Benchmarking Guide: Test CPU, Disk I/O, Network Speed, and Latency

Benchmarking your VPS is the only reliable way to know if you’re getting the performance you pay for. Raw specs — “4 vCPUs, 8GB RAM” — tell you very little about real-world throughput. This guide covers practical benchmarking for CPU, memory, disk I/O, network, and application-level performance, with commands you can run on any Linux VPS.

Installing Benchmarking Tools

Most benchmarks use a handful of well-established tools. Install them all upfront:

# Debian/Ubuntu
sudo apt install sysbench fio htop iperf3 speedtest-cli
# For geekbench-style testing
curl -fsSL https://sh.geekbench.com | sh

CPU Benchmarking with sysbench

Sysbench calculates prime numbers to test raw CPU throughput:

# Single-threaded test
sysbench cpu --cpu-max-prime=20000 run

# Multi-threaded (use all vCPUs)
sysbench cpu --cpu-max-prime=20000 --threads=$(nproc) run

Compare your results against provider averages. A good VPS should score 800-1200 events per second per core in this test. For a comparison of VPS providers and their typical benchmark results, check out VPS providers on our site.

Memory Benchmarking

Test memory bandwidth and latency:

# Sequential memory write
sysbench memory --memory-block-size=1M --memory-total-size=10G run

# Random memory access
sysbench memory --memory-block-size=1M --memory-total-size=10G --memory-access-mode=rnd run

Disk I/O Benchmarking with fio

Fio is the gold standard for disk benchmarking. It tests sequential and random read/write patterns:

# Sequential read
fio --name=seqread --ioengine=libaio --direct=1 --bs=1m --size=2G --numjobs=1 --iodepth=64 --rw=read

# Random 4K writes (critical for database workloads)
fio --name=randwrite --ioengine=libaio --direct=1 --bs=4k --size=2G --numjobs=1 --iodepth=64 --rw=randwrite

# Random 4K reads (email/web server workloads)
fio --name=randread --ioengine=libaio --direct=1 --bs=4k --size=2G --numjobs=1 --iodepth=64 --rw=randread

Interpretation guide for disk benchmarks:

• NVMe SSD: 2000+ MB/s sequential read, 100K+ IOPS random 4K
• SATA SSD: 400-550 MB/s sequential, 40-80K IOPS random 4K
• HDD-backed VPS: 100-200 MB/s sequential, <5K IOPS random 4K

Network Performance Testing

Use iperf3 for throughput and speedtest-cli for real-world bandwidth:

# iperf3 (requires a server on the other end)
iperf3 -c iperf.he.net -t 30

# speedtest-cli (Ookla)
speedtest-cli --simple

# Check latency to key regions
ping -c 10 google.com
mtr google.com

Application-Level Benchmarking

Benchmarks that mirror your actual workload are most valuable:

# Web server (using ApacheBench)
sudo apt install apache2-utils
ab -n 10000 -c 100 http://your-server-ip/

# Database (using pgbench for PostgreSQL)
sudo apt install postgresql-contrib
pgbench -i -s 50
pgbench -c 10 -j 2 -t 1000

Recording and Comparing Results

Create a benchmark log to track performance over time:

#!/bin/bash
# save as ~/benchmark.sh
echo "=== CPU ===" >> ~/benchmark_log.txt
sysbench cpu --cpu-max-prime=20000 --threads=$(nproc) run | grep "events per second" >> ~/benchmark_log.txt
echo "=== DISK ===" >> ~/benchmark_log.txt
fio --name=randread --ioengine=libaio --direct=1 --bs=4k --size=1G --numjobs=1 --iodepth=64 --rw=randread --output-format=json 2>/dev/null | python3 -c "import json,sys; d=json.load(sys.stdin); print(f'Random Read IOPS: {d[\"jobs\"][0][\"read\"][\"iops\"]}')" >> ~/benchmark_log.txt
echo "---$(date)---" >> ~/benchmark_log.txt

Run benchmarks monthly to catch performance degradation. For provider-specific benchmark data to compare your results against, see the full specs on our VPS comparison platform.

Conclusion

Consistent benchmarking helps you validate that your VPS provider delivers the promised resources, diagnose performance issues, and make informed upgrade decisions. Save your results, run tests at different times of day, and always test with the same parameters for apples-to-apples comparisons.