Why Benchmarking Matters
Most people benchmark either because something feels slow or because they’re verifying a new purchase. Both are valid. The important thing is understanding what a given tool is actually measuring, because sequential read/write speeds look great on paper but real-world performance is usually about random 4K read/write and queue depth behavior.
The Tools Worth Knowing
CrystalDiskMark – The standard. Free, clear UI, measures sequential and random speeds across multiple configurations. If you only use one tool, use this one. It’s the benchmark most reviewers publish.
AS SSD Benchmark – Good for SSDs specifically. Gives you a score and breaks down multiple test scenarios. Useful for comparing SSDs against each other.
ATTO Disk Benchmark – Measures across varying file sizes. Good for getting a more complete performance picture than a single sequential test.
HDTune – Older but still solid for HDDs. Has a transfer rate test and a health scan. Useful if you’re checking an older mechanical drive.
Parkdale – Lighter weight, less known. Good if you want a quick read/write without installing a heavier tool.
UserBenchmark – Controversial. Some people like the relative scoring, but the benchmark methodology has been criticised in hardware communities. I’d use it for comparative context only, not as a primary measure.
What to Actually Look At
For SSDs: random 4K read speeds are more representative of day-to-day performance than sequential. For HDDs: average read speed and access time matter more than peak sequential.
If you’re buying secondhand storage, run both a benchmark and a health check (CrystalDiskInfo is good for the latter). A drive can benchmark fine but show concerning SMART data.
What’s everyone using?