AS SSD Test Size: 1 GB vs 3 GB vs 5 GB vs 10 GB Explained

In AS SSD Benchmark you can choose a test file size of 1, 3, 5, or 10 GB, and the default is 1 GB. New users often assume bigger must be more accurate, or worry that 1 GB is somehow cheating. Neither is quite right. The test size changes how long the test runs, how much data is written, and how much of a drive’s sustained behaviour you see. This guide explains what actually changes with each size, when the default is the right choice, and when a larger file tells you something the default cannot.

Whichever size you pick, get the tool from the download page and compare results only against results taken at the same size. A 1 GB result and a 10 GB result from the same drive are not interchangeable, and mixing sizes is one of the easiest ways to create a fake problem or hide a real one. If you are new to the workflow itself, start with how to use AS SSD Benchmark first.

What the test size actually controls

The synthetic tests create a test file of the chosen size on the selected drive and measure reading and writing against it. Sequential speed is measured over that file, the random 4K tests work within it, and the write access time uses the test file as well. Choosing 10 GB does not change what Seq, 4K, or 4K-64Thrd mean, and it does not unlock extra tests. It gives the drive a larger workload, which takes longer and can expose behaviour that a short test finishes before revealing.

Think of it like timing a car. A short sprint tells you how quickly it accelerates. A longer run tells you whether it can hold that pace. Both are real measurements, but they answer different questions. Most people asking is my SSD working properly need the sprint and the standard comparison numbers. People investigating slowdowns during long copies, game installs, or video work may need the longer run as a second step.

The 1 GB default: fast, standard, and comparable

Use 1 GB for your baseline, for quick health checks after a change, and for any comparison with screenshots you find in reviews or forums. It finishes in a few minutes on most SSDs, writes the least data, and is the size used in the overwhelming majority of shared results. If a review does not state its test size, it almost certainly used the default, although you should still treat unstated settings with caution.

The limitation of 1 GB is that it is short. Some drives can deliver very fast bursts for a short workload and then settle to a lower sustained speed during a long write, as internal buffers fill and the drive has to work harder to keep up. A 1 GB test may finish inside that fast opening phase. That does not make the result false; it makes it a short-workload result. It tells you little about a 40 GB game install.

3 GB and 5 GB: the middle ground

The 3 GB and 5 GB options are a compromise. They take noticeably longer than 1 GB and push past the shortest burst behaviour, without the time and write volume of a full 10 GB run. They are a reasonable second test when your 1 GB sequential write looks fine but real-world large copies feel slow, or when you want a more demanding check on a drive with plenty of free space.

You still cannot compare a 5 GB score directly with a 1 GB score as if the settings matched. Keep a separate record for each size, label screenshots clearly, and re-run at 1 GB when you need the standard comparison number. Middle sizes are diagnostic tools, not replacements for the baseline.

10 GB: sustained writes and drives that slow down

Choose 10 GB when your question is specifically about sustained performance: large writes that start fast and then drop, drives that feel fine in short tasks but crawl during installs, or checking a drive you plan to use for video capture or large project files. The longer write gives buffers time to fill and the drive time to settle into its sustained pace, so the sequential write row in particular may come out lower than at 1 GB. That lower number is information, not an error.

The costs are time and writes. A 10 GB run takes much longer, needs at least 10 GB of free space plus working room, and writes considerably more data. On an old drive, a nearly full drive, or a laptop running hot, those costs matter. Do not run 10 GB repeatedly in a loop. Once or twice to answer a sustained-performance question is enough, then return to 1 GB for routine checks.

How size affects each row differently

Sequential write is the row most likely to change with size, for the sustained-write reasons above. Sequential read usually changes less, because reading does not fill write buffers in the same way. 4K and 4K-64Thrd are about random access patterns rather than long continuous transfers, so their results tend to move less with file size, although a larger file and a fuller drive can still add some variation. Access time for writes uses the test file, while read access time is measured differently, so do not expect every row to shift by the same proportion.

The Compression-Benchmark is a different lens on a related question: it varies how compressible the data is rather than how large the file is. If your concern is that results change with the type of data, read the Compression-Benchmark explained alongside this guide. Size and compressibility are separate variables; change one at a time when you are diagnosing.

A practical testing routine

• Baseline every drive at 1 GB on a quiet system and save the screenshot.
• If large real-world writes feel slow despite a good baseline, run once at 5 GB or 10 GB and compare mainly the sequential write pattern.
• Free space, close background work, and let the drive cool between large runs so heat does not confuse the comparison.
• Record size, free space, and date with every saved result; never compare totals across different sizes.
• Return to 1 GB for follow-up checks so you can track change over time against your original baseline.

In short: 1 GB is not a toy setting and 10 GB is not the real test. The default is the standard for quick, comparable benchmarks. Larger sizes are for a specific sustained-write question, they cost time and writes, and a lower sequential write at 10 GB usually reveals how the drive behaves over a long workload rather than proving it is broken. Choose the size that matches your question, label it, and compare like with like.