AS SSD Compression-Benchmark Explained: Reading the Graph

Under the Tools menu of AS SSD Benchmark sits the Compression-Benchmark, a graph that confuses many first-time users. Instead of one speed number, it draws lines that rise, fall, or stay flat as the test data becomes less compressible. That shape is the result. This guide explains how to run the test, how to read its graph in plain language, and why the compressibility of data can change the speed of some controllers while leaving others almost untouched.

Start with the standard tool from the download page and a normal main-window run first, following the main how-to-use tutorial. The compression graph is easiest to interpret when you already know your sequential read and write figures, because the graph is essentially a deeper look at how those sequential transfers behave with different kinds of data.

What compressible data means

Some data can be squeezed into a smaller representation without losing anything. A text file full of repeated words compresses a lot. A ZIP archive, a JPEG photo, or an encrypted file is already squeezed or effectively random, so it barely compresses at all. The Compression-Benchmark fills its test data with a controlled mix, moving from highly compressible content at the left of the graph to incompressible content at the right, and measures transfer speed at each step.

Why would a storage drive care? Some SSD controllers have historically used compression inside the drive as part of how they write data: if incoming data can be compressed on the fly, the controller writes less physical data than the computer sent, which can make the transfer finish faster and reduce writing work. Other controllers do not rely on that technique and handle every gigabyte the same way regardless of its content. The graph reveals which behaviour your drive shows, without you needing to know its internal design in advance.

How to run the Compression-Benchmark

Select the target drive in the main window, open Tools, and choose Compression-Benchmark. Press Start in the new window and wait. The test works through its data patterns and draws read and write lines across the graph, with data compressibility along the horizontal axis and speed in MB/s on the vertical axis. It can take a while, because it is running many passes rather than one quick measurement.

Run it on a quiet system with enough free space, just as you would the main test. Close heavy background work, let any earlier benchmark finish and the drive settle, and do not compare a graph from a busy system with one from an idle system. Keep the resulting graph image with your main screenshot if you are documenting a drive for troubleshooting or for a forum question, because the shape often explains arguments that a single Seq number cannot settle.

Reading the graph: flat, sloping, and stepped shapes

A flat line means speed barely changes from compressible to incompressible data. The drive delivers roughly the same sequential performance whatever you feed it. For most users this is the most predictable behaviour: the speed you see with ordinary mixed files is the speed the graph shows everywhere.

A line that starts high and slopes down to the right means the drive is faster with highly compressible data and slower once the data can no longer be compressed. The left side of that graph flatters the drive if your real files are mostly already-compressed media, archives, and encrypted data. Read the right side of the graph as the more realistic figure for that kind of workload, and read the gap between the two sides as the size of the compression effect.

An uneven or stepped line can appear when the controller changes strategy at certain points, or when other factors such as heat, free space, or background activity intrude during a long test. Do not over-interpret a single jagged run. Re-run once on a quiet, settled system and look at the overall shape rather than every small bump.

• Compare read and write lines separately; write is where compression effects usually show most clearly.
• Note the speed at the far right (incompressible) as your conservative sequential figure.
• Note the speed at the far left, but do not quote it alone as the drive speed without saying the data was highly compressible.
• Compare shapes between drives, not just peak heights; a slightly lower flat line can be better in real use than a high line that collapses.

Why this matters when comparing benchmarks

Different benchmark tools use different test data. A tool that tests mainly with highly compressible data can make a compression-assisted controller look exceptionally fast, while a tool using incompressible data shows the same drive in a more modest light. Neither result is fabricated; they are measuring with different material. This is a classic reason two honest benchmarks of the same SSD disagree, and why a single headline number, quoted without its test conditions, is worth so little.

For your own comparisons, keep the material constant. Compare your drives with the same tool and same settings, and use the main AS SSD test size consistently too, as explained in test size 1 GB vs 10 GB. When a manufacturer’s advertised speed looks far above everything you measure, check what data pattern and workload produced that figure before you assume your drive is faulty.

What the graph does not tell you

The Compression-Benchmark is about sequential transfers with varying data. It does not replace the random 4K rows that describe everyday responsiveness, and it does not measure health, temperature, or endurance. A beautiful flat graph on a drive with weak 4K results will still feel ordinary in Windows. Read it alongside the main window and, for file workloads, the Copy-Benchmark, which varies file shape rather than data content.

In short, the graph is a honesty test for sequential speed. Flat means what you see is what you get with any data. Sloping down means early, compressible numbers flatter the drive and the right-hand side is your realistic planning figure. Once you can read that shape, manufacturer claims, review charts, and your own results all become much easier to reconcile.