Benchmark Methodology

The Benchmark Process

A VoxelBench benchmark runs 14 automated tests on your Minecraft server, covering three areas:

1. Single-Core Performance — How fast your CPU handles Minecraft's main thread
2. Hardware — Raw memory, disk, and multi-core capabilities
3. Gameplay Stress — How the server performs under real Minecraft workloads

The plugin orchestrates all tests automatically. You start the benchmark, and the plugin handles the rest: spawning entities, loading chunks, activating redstone, triggering explosions, etc. No manual intervention needed.

Test Duration

A full benchmark takes 5 to 40 minutes depending on your server's performance. Faster servers complete tests quicker — this is itself a performance indicator and is factored into the score.

Preparing for a Reliable Benchmark

Getting consistent, comparable results requires controlling the test environment. Here's a complete checklist.

Server Configuration

Minecraft Version

• Use a stable release (not snapshots or pre-releases)
• The same Minecraft version should be used when comparing results
• Newer versions may have different performance characteristics

Server Software

• Paper is recommended for the most optimized results
• Spigot, Purpur, and Folia are supported
• Vanilla servers will score lower due to lack of optimizations — this is expected and reflects real-world performance

Java Version

• Use Java 21 (LTS) or newer
• Java 17 works but Java 21 has measurable performance improvements
• Avoid Java 8 — it lacks modern GC algorithms and optimizations

JVM Arguments (Critical)

The garbage collector configuration has a major impact on results:

• Aikar's flags are the recommended baseline for most servers
• ZGC (-XX:+UseZGC) produces very low GC pause times but may use more memory
• G1GC (default with Aikar's flags) is a good balance of throughput and latency
• Avoid default JVM flags without tuning — they lead to long GC pauses that directly penalize your score

Key JVM settings to verify:

• -Xms and -Xmx should be equal (prevents heap resizing during tests)
• Allocate enough RAM for the tests (8 GB minimum recommended, 10-12 GB ideal)
• Don't over-allocate (32 GB for a test server wastes memory and can increase GC pauses)

Environment

No Players

• Run benchmarks with zero players online
• Player activity creates unpredictable load that varies between runs
• Even AFK players generate chunk ticks and entity interactions

No Other Plugins (Ideal)

• For the most accurate hardware comparison, test with only the VoxelBench plugin
• If you need to test your production configuration, understand that other plugins add overhead
• Anticheat plugins especially can interfere with entity spawning and block physics tests

Dedicated Resources

• Close other applications on the machine (or VM/container)
• On shared hosting: other customers' servers affect your results
• CPU throttling (power saving modes, thermal throttling) will produce inconsistent results
• If possible, run the benchmark during low-activity hours on the host machine

World State

• Use a fresh world or one with minimal builds
• Large existing worlds add overhead to chunk ticking and world save tests
• The plugin creates its own test areas, but existing loaded chunks still consume resources
• Officials benchmark use a flat world with seed "benchmark"

Network

• Network latency does not affect the score (all measurements are server-side)
• However, the connection between plugin and VoxelBench servers must be stable for result submission

What Makes Results Comparable

Two benchmarks are directly comparable when:

When comparing hosting providers, VoxelBench controls for all these factors by using identical server configurations across providers.

Hosting Provider Comparison

When we benchmark hosting providers on the Hosting Comparison page:

1. Same server JAR — Identical Paper version and build
2. Same Java version — Same JDK distribution and version
3. Same JVM flags — Identical startup arguments
4. Same world — Fresh world, no builds, seed "benchmark"
5. Same plugins — VoxelBench plugin only
6. Same configuration — Default server.properties and Paper config
7. Multiple runs — Several benchmarks per provider to account for variance
8. Different times — Tests at various times of day to catch resource contention

This ensures the only variable is the hosting infrastructure itself: CPU, RAM speed, disk I/O, and how well resources are isolated from other customers.

Factors That Hurt Your Score

Factors That Don't Affect Your Score