Run Your First AS SSD Benchmark: Step-by-Step First Test

Your first AS SSD Benchmark run should be boring in the best possible way: a quiet computer, the correct drive selected, the default 1 GB test size, and a saved screenshot at the end. This hands-on guide tells you exactly what to close before you start, how long to expect the test to take, what to watch while it runs, and how to keep the result so it is still useful months later when you compare a new drive, a firmware change, or a Windows reinstall.

You do not need to tune Windows or change any advanced settings for a first run. In fact, changing settings before you have a baseline makes the result harder to interpret. Get one clean baseline first, following the general workflow in how to use AS SSD Benchmark, and only then experiment. If the tool is not on your computer yet, get it from the download page and extract the whole ZIP folder before you continue.

Before you start: a five-minute preparation checklist

Preparation matters because AS SSD Benchmark measures the drive while Windows is live. Anything else using that drive at the same time competes for its attention and lowers or scatters the numbers, especially in the random 4K tests. A few minutes of tidying up gives you a result you can trust and repeat.

• Close heavy programs: pause game downloads and updates, video renders, large file copies, cloud sync uploads, and virtual machines stored on the target drive.
• Let Windows settle: after booting, wait a few minutes for startup programs, indexing, and update checks to finish before you run the test.
• Check free space: make sure the drive has comfortably more free space than the test file. At the default 1 GB size, several gigabytes free is a sensible minimum; a nearly full SSD is slower and less consistent.
• Plug in laptops: run on mains power with a balanced or high-performance power plan, so power saving does not throttle the system mid-test.
• Note the drive temperature context: a drive that has just finished a huge copy job may be warm. Let the system idle briefly if you have just hammered it.

You do not need to disconnect from the internet or disable security software for a normal first run. Keep the setup close to how you actually use the computer, just without active heavy work in the background. That balance gives a realistic baseline rather than an artificial best case that you will never reproduce in daily use.

Extract and open the tool correctly

Extract the ZIP you downloaded into its own folder and keep the executable and the en-US language folder together. Double-click the executable; the tool is portable and does not install anything. If the window opens in German, do not panic and do not reinstall. Use the Language menu to switch to English, and check that the en-US folder sits next to the executable. Our companion guide to the German interface and how to switch it to English explains why that happens.

At the top of the window, select your SSD from the drive drop-down. Verify the model name and capacity shown under the drop-down. Many computers list several drives, including USB sticks and external drives, and testing the wrong one is the most common first-run mistake. It wastes time and produces a confusingly slow or fast result that has nothing to do with your main SSD.

Settings for a first run: change nothing

Leave the test size at 1 GB. Leave all the synthetic test checkboxes selected. Leave the display in MB/s. Those defaults are what most published screenshots use, which makes your result easy to compare. The main window covers sequential speed, random 4K speed at one thread, random 4K speed across 64 threads, and access time. Together they give a balanced picture: sequential for large transfers, 4K for everyday responsiveness, 4K-64Thrd for heavy multitasking loads, and access time for latency.

Do not start with the 10 GB size to be thorough. A bigger file takes longer, writes ten times as much data in that test, and answers a different question about sustained behaviour. It has its place, and we explain the trade-offs in test size 1 GB vs 10 GB, but a baseline should use the size everyone else uses for a quick comparison: 1 GB.

Press Start: what happens and how long it takes

Click Start. The program creates its test file on the selected drive and works through the read tests and write tests in order. The result table fills in row by row, and the running score appears at the end. On a typical SATA or NVMe SSD at 1 GB, expect a few minutes from start to finish. Older or nearly full drives take longer. A mechanical hard drive can take far longer, sometimes approaching an hour for the full set, because random access and access time tests are painfully slow on spinning disks.

While it runs, leave the mouse and keyboard alone as much as you can. Do not browse a drive-heavy photo library, do not start a game, and do not run Windows Update. If you notice the system was busy anyway, for example a sync client started uploading, it is better to abort, let the system settle, and restart the test than to keep a polluted result and wonder later why it looks odd.

Watch the access time and 4K rows with realistic expectations. New users often focus only on the big sequential number because it looks impressive and is the number used in advertising. In everyday Windows use, the 4K row and access time often explain how snappy a system feels far better. A first run is as much about learning which row matters for which task as it is about the score itself.

When it finishes: check the result before you save it

First, confirm the run completed every row and shows a total score. A partial run after an abort is not a baseline. Second, glance at the top information again: does the controller line and alignment status look normal, and was the correct drive tested? Third, ask whether the pattern makes sense. Sequential much higher than 4K is normal. Access time in milliseconds should be a small fraction for an SSD and much larger for a hard drive. If one number looks wildly out of line compared with the others, run the test once more on a quiet system before you draw conclusions.

Small run-to-run variation is normal. A few percent difference in sequential speed or 4K speed between two runs on different days is not a problem. What matters is a consistent pattern or a large drop. That is why a saved baseline is valuable: without it, you have only a feeling that the drive used to be faster, and feelings do not diagnose anything.

Save the screenshot and the XML the right way

From the File menu, save a screenshot of the finished window and export the XML result. Give the files useful names that include the drive and date, for example the model plus the year and month, and store them together in a benchmarks folder. Keep the context notes with them: test size, free space, and anything unusual about that day, such as a new driver or a fresh Windows install.

A screenshot is best for sharing in a forum post or comparing side by side. The XML export keeps the structured numbers. A dedicated guide to exporting AS SSD results shows both options and what to include when you ask others for help. When you do share, include the full window, not a cropped score alone, because helpers need to see the drive model, controller, and alignment lines at the top to give useful advice.

What to do next

Keep your baseline and use the computer normally. Re-run the same 1 GB test after a major change, such as moving the drive to a new computer, changing controller mode in the BIOS, or freeing a large amount of space, and compare like with like. If the new total is far lower and repeats across runs, work through alignment and controller checks next rather than re-running the benchmark ten times hoping for a better number. A good first run is not the highest run; it is a clean, documented run you can repeat.