Testing External Drive Speed on a Mac, and Reading the Result

macOS leaves benchmarking to third parties

On Windows, CrystalDiskMark is the obvious starting point. macOS ships with nothing comparable, so measuring an external drive or a USB stick means installing a tool.

Blackmagic Disk Speed Test is the popular pick, and for good reason: free, one click, clear sequential read and write figures. It also has limits. Only large sequential transfers are measured, random 4K performance is ignored, and the numbers arrive without any interpretation. A figure that looks far too low leaves you guessing whether the drive, the cable or the port is at fault.

The four numbers a real benchmark reports

The CrystalDiskMark profiles have become a de facto standard, and they cover four situations:

• Sequential read and write, the figure quoted for large video and photo transfers;
• low-queue sequential, which is closer to how a normal single-task copy feels;
• random 4K at high queue depth, where small-file workloads live;
• random 4K at queue depth one, the responsiveness you notice when a folder full of small files opens.

What counts as normal

Compare your sequential read result against these bands:

• under 45 MB/s means the link is almost certainly USB 2.0, so try a USB 3.x or Thunderbolt cable plugged straight into the Mac;
• 80 to 160 MB/s is ordinary for a 2.5-inch mechanical drive;
• 200 to 470 MB/s fits an SSD on USB 3.x at 5Gbps;
• 470 to 950 MB/s is full-speed USB 3.1 Gen 2 at 10Gbps;
• above 950 MB/s puts you in Thunderbolt and USB 4 territory, close to the drive's own ceiling.

Three reasons a result comes out low

A low number is not automatically a dying drive. In order of likelihood:

• The link. A charge-only cable with no data lines, or a USB 2.0 port behind a hub, holds you near 40 MB/s.
• SSD cache exhaustion. Budget SSDs reach their headline speed only while writing inside an SLC cache. A long sustained write drains it and throughput can fall by 40% or more. That is normal behaviour, and it matters for sustained work such as video editing.
• A write figure below roughly a third of the read figure. Suspect cache exhaustion, thermal throttling or degrading health, and back up the important files soon.

DiskBee's built-in test

DiskBee includes a free speed test that follows the same four-profile methodology, measuring read and write separately and showing both on gauges with live curves and IOPS. Three sizes are offered: 512MiB for a quick check, 1GiB as the standard run, and 4GiB to expose cache exhaustion and thermal throttling.

Each run ends with an analysis rather than a bare score. It reports whether the drive is stuck on a USB 2.0 link, whether the SSD drained its cache mid-write, and whether the write speed is abnormally low, then says what to do next. Results can be copied out in one click. Because the test runs through the app's own NTFS read-write path, the figures match the speeds you actually get when copying files.

The test is free in both editions. In the Mac App Store build, pick the disk when prompted on first run so the sandbox grants access. It writes a temporary file and removes it afterwards, so leave the drive idle while it runs.

What to remember

Running a benchmark is easy. Reading it is the skill. Below 45 MB/s points at a USB 2.0 link. A sustained write that falls by half is the signature of SSD cache exhaustion. Write speeds that stay under a third of read speeds justify a health check.