If an SSD's box quotes speeds that seem too good to be true, there's a fair chance they came from ATTO. ATTO Disk Benchmark is a legitimate tool, but its default conditions are unusually favourable to SSDs, producing peak numbers that real-world mixed workloads rarely reach. Manufacturers love it for exactly that reason. Understanding why ATTO inflates — and what to look at instead — keeps you from buying on best-case marketing figures. This article explains the ATTO gap.
It pairs with the CrystalDiskMark explainer, our NVMe SSD guide, and the DRAM-less SSD trap.
Why ATTO Inflates
- Favourable data patterns: ATTO's defaults use data and access patterns that are easy for an SSD's controller to handle quickly, producing near-best-case throughput rather than the messy mixed data of real use.
- Best-case block sizes: it reports peak figures at the block sizes the drive handles best, which look great on a box but don't reflect a typical workload's mix.
- Short, ideal conditions: like many benchmarks, it doesn't capture what happens once a drive's fast cache fills under sustained writing.
The Cache and Sustained-Write Reality
The biggest gap ATTO hides is sustained write performance. Many SSDs use a fast SLC cache that delivers high speeds briefly, then slow dramatically once it's exhausted — a real problem on large transfers and especially on DRAM-less drives. ATTO's short, favourable test rarely reveals this cliff, so a drive that looks fast in ATTO can crawl when you actually copy a large folder. Cheaper QLC drives are especially prone to it.
What to Look at Instead
For a realistic picture, look at mixed/random performance (the low-queue random row in CrystalDiskMark), sustained write behaviour after the cache fills, and independent reviews that test real workloads — not the box's ATTO headline. The number on the packaging is a best case; your experience depends on the worst cases, especially under sustained load. See our benchmark-reading hub for the general principle.
Frequently Asked Questions
Why are SSD box speeds so high? They're often best-case figures from ATTO or sequential tests under favourable conditions — easy data patterns and ideal block sizes that real mixed workloads don't match. The marketing number is a peak, not a typical result.
What does ATTO not show? Sustained write performance after the SSD's fast SLC cache fills, where many drives — especially DRAM-less and QLC ones — slow dramatically. ATTO's short, favourable test rarely reveals this cliff, so a high ATTO number can hide poor large-transfer performance.
What should I look at instead of ATTO? Mixed/random performance (low-queue random in CrystalDiskMark), sustained write behaviour once the cache fills, and independent reviews testing real workloads. Your experience depends on worst cases under sustained load, not the box's best-case headline.
The One Thing to Remember
ATTO produces best-case SSD numbers — favourable data, ideal block sizes, short test — which is why they end up on the box and why real workloads rarely match them. It hides the sustained-write cliff that hits DRAM-less and QLC drives once their cache fills. Judge an SSD on mixed/random performance and sustained write from independent reviews, not the ATTO headline.
Want storage that's fast in real use, not just on the box? Configure a PC online → or talk to our team → and we'll pick an SSD that holds up under sustained, real-world load.