It’s become one of the commonest questions: should you pay extra for a larger SSD in your new M1 Mac, or will you get as good performance at a better price with a cheaper internal SSD and fast external storage? Now that manufacturers are claiming that their external SSDs are much faster than regular SATA/USB-C models, it’s tempting to believe them, and save some money on the cost of your Mac. In this article, I’ll show how you can use my free app Stibium to compare the two options a bit more objectively.

Unlike other storage benchmarking apps, Stibium provides a great deal of flexibility. For this workthrough, I’ll assume that the most performance-critical use of your Mac is going to be sustained high-speed reading, and the benchmark that you want must give a good indication of the read rate which can be sustained by regular apps over periods of many minutes.

What you’ll do is start by running a basic write benchmark, which is primarily intended to provide 160 test files for a more sustained read test, which will measure the rate at which 1600 files of different sizes, totalling over 320 GB, can be read. It’ll do that by reading those 160 test files, in randomised order, a total of ten times each. You can perform this test on as many different types of storage as you wish. In this example, you’ll run it on the internal SSD of your M1 Mac, where it requires around 35 GB of free space for the test files.

Start by creating a folder at the top level of your Home folder, with an appropriate name, to contain those test files. Then open Stibium, and set the following at the top of its window:

• the central popup in the top row to read 0x41, which results in all the bytes in its test files being 41 in hexadecimal;
• in the second row, set the number of repeats to 10;
• in the third row, tick No Cache, so that it prevents macOS from caching read files;
• to the right of that, uncheck Verbose, unless you really want to see every one of its 1600 measurements.

Click the Write Series… button at the left, select the test file folder which you just created, then click Write to start the write test. Leave your mouse/trackpad alone, and a minute or two later the tests should be complete. Sometimes you then need to force the window contents to update, when they’ll display the test report. The measured write speed will also be displayed as the write result above, in GB/s or MB/s as appropriate.

When I ran this on my M1 Macs, it gave a write speed of 2.7 GB/s for the internal storage.

It’s best if you can then quit Stibium, restart, let your Mac settle for a couple of minutes, and ensure that the next test isn’t going to be interrupted by Time Machine backing up or any other distractions.

Open Stibium again, noticing that the write result from the previous test is still displayed, and ensure that the number of repeats is still set at 10. Then click on the Read Folder… button in the middle, select the folder in which your test files are stored, and click on Read.

Leave your Mac alone for the next few minutes as it reads those files ten times and finally displays the results. In my test, the read speed came out to 2.9 GB/s.

I then repeated these tests with a test folder on my Samsung external X5 SSD. In the write test, it underwent thermal throttling almost immediately, resulting in an overall write speed of only 700 MB/s, instead of the 2 GB/s which I would have expected had it remained cool enough to write at its full speed. In the read test, there was no such throttling, and it achieved a read speed of 2.5 GB/s, that’s 86% of that of the internal storage.

Don’t forget to trash your test folders once you’ve finished.

Like my other free apps, Stibium has very detailed documentation, which is provided both as a separate PDF when you download it, and through the Help menu. This explains how it measures transfer speed, and even gives the Swift code used to read and write the test files.

Although I’m prepared to be surprised, as yet I have come across no other SSD which can match the performance of Apple’s internal SSDs fitted to the M1 MacBook Pro and Mac mini. If you find one, please let me know, now that you have the tool to verify the vendor’s claims.