How to Benchmark SSD Performance on Windows Using CrystalDiskMark

Solid‑state drives (SSDs) have become the default storage solution for most PCs, but not all SSDs are created equal. Whether you’ve just installed a new drive, upgraded to a faster model, or simply want to verify that your system is delivering the speeds it promises, benchmarking is the most reliable way to get concrete numbers. In this guide we’ll walk you through the entire process of testing SSD performance on a Windows machine using the popular, free utility CrystalDiskMark. No prior experience with benchmarking tools is required—just a willingness to follow a few straightforward steps.

What You’ll Need

• Windows 10 or Windows 11 (64‑bit recommended)
• A solid‑state drive you want to test (internal or external)
• CrystalDiskMark installer (latest version from the official site)
• Administrator rights on the PC
• At least 5 GB of free disk space on the test drive

Step 1: Download and Install CrystalDiskMark

Head over to the official CrystalDiskMark page at crystalmark.info. Click the “Download” button for the Windows version and choose the portable zip file if you prefer not to install, or the installer for a permanent setup. Run the installer (or unzip the portable version) and accept the default installation path, typically C:Program FilesCrystalDiskMark. The program requires no special configuration at this stage; just make sure you launch it with administrator privileges so it can access low‑level disk information.

Step 2: Prepare Your SSD

Before you start measuring, give the SSD a clean slate. If the drive is brand new, you can skip this step. For an existing drive, delete any large files you don’t need, and run a quick defragmentation (Windows’ built‑in “Optimize Drives” tool) to ensure the file system isn’t fragmented. Also, make sure the drive is connected directly to a SATA III or NVMe slot—not through a USB‑2.0 hub—because the interface will heavily influence the results.

Step 3: Close Background Applications

Benchmarking is a CPU‑ and I/O‑intensive operation. To avoid skewed numbers, close any unnecessary programs: web browsers, file sync services (OneDrive, Dropbox), antivirus real‑time scanning, and especially any software that writes to disk (e.g., video editors). You can open Task Manager (Ctrl+Shift+Esc), go to the “Processes” tab, and end tasks that are not essential. This step ensures that CrystalDiskMark has exclusive access to the drive during the test.

Step 4: Configure Test Settings

Launch CrystalDiskMark. You’ll see a simple interface with several dropdowns:

• Test drive: Select the letter of the SSD you want to benchmark (e.g., D:).
• Number of tests: The default “5” is a good balance between accuracy and time.
• Test size: Choose “1 GB” for a quick test or “4 GB” for a more thorough assessment. Larger sizes reduce the impact of cache effects.
• Queue depth: Keep the default “1” for sequential tests; set it to “32” for random I/O tests if you want to simulate heavy workloads.

For beginners, we recommend the following preset:

• Test size: 1 GB
• Number of tests: 5
• All four checkboxes (Seq Q32T1, Seq Q1T1, Rnd Q32T1, Rnd Q1T1) enabled

This configuration covers both sequential and random read/write performance, which are the two most common metrics people look at when evaluating SSDs.

Step 5: Run the Benchmark

Click the large “All” button to start the full suite of tests. CrystalDiskMark will display a progress bar for each test and update the results in real time. Expect the entire run to take anywhere from 2 to 5 minutes depending on your drive’s speed and the test size you selected. Do not interfere with the process—avoid moving files, plugging/unplugging USB devices, or putting the computer to sleep.

Step 6: Analyze the Results

When the benchmark finishes, you’ll see a table with four rows:

• Sequential Read (Seq Q1T1)
• Sequential Write (Seq Q1T1)
• Random Read (Rnd Q1T1)
• Random Write (Rnd Q1T1)

Each value is presented in MB/s (megabytes per second). Compare these numbers against the specifications listed by the SSD manufacturer. Small deviations (5‑10 %) are normal due to system overhead, but large gaps may indicate a configuration issue, outdated firmware, or a throttling problem.

For a deeper dive, you can also look at the “Q32” results, which simulate multiple outstanding I/O requests—useful for evaluating performance in professional workloads like video editing or database servers.

Step 7: Save and Share Your Results

CrystalDiskMark lets you export the results as a CSV file via the “Save” button. Store the file in a dedicated “Benchmark” folder for future reference, especially if you plan to compare multiple drives or track performance over time. You can also take a screenshot of the result table and embed it in a blog post, forum thread, or support ticket when seeking help.

Common Mistakes to Avoid

Even beginners can unintentionally invalidate their results. Here are the most frequent pitfalls and how to sidestep them:

• Testing on a partition, not the whole drive: Some SSDs have different performance zones. Always benchmark the root of the drive (e.g., D:) rather than a small sub‑folder.
• Running the test on a drive that’s nearly full: SSDs lose speed when they’re above 80 % capacity. Free up space before testing.
• Leaving Windows Update or other background services running: They can trigger disk writes mid‑test, skewing results.
• Using the wrong interface: Benchmarking a NVMe SSD through a USB‑C adapter that only supports USB 3.0 will cap speeds at ~5 Gb/s, far below the drive’s capability.
• Not disabling write caching: In some rare cases, Windows write cache can mask true write speeds. You can temporarily disable it in Device Manager → Disk drives → Properties → Policies, but remember to re‑enable it afterward.

Tips and Tricks

Want to get even more out of your benchmarking session? Try these pro‑level adjustments:

• Update SSD firmware: Manufacturers often release firmware that improves performance or stability. Check the vendor’s website before testing.
• Run the test multiple times: Take the average of three runs to smooth out occasional spikes.
• Use the “Custom” mode: Adjust queue depth and thread count to mimic specific workloads (e.g., gaming vs. workstation).
• Combine with Windows Performance Monitor: Open perfmon.exe and add counters for “Physical Disk → Disk Write Bytes/sec” to see real‑time throughput alongside CrystalDiskMark.
• Temperature check: High SSD temps (>70 °C) can trigger thermal throttling. Use a tool like HWMonitor to ensure the drive stays cool during the test.

Frequently Asked Questions

Do I need to run CrystalDiskMark as administrator?

While the program can launch without admin rights, running it as an administrator ensures it can access low‑level disk interfaces and eliminates permission‑related errors, especially on drives with BitLocker or other encryption.

Why are my sequential write numbers lower than the advertised speed?

Manufacturers usually quote peak speeds achieved under ideal conditions (e.g., large, continuous writes to a freshly formatted drive). Real‑world factors—such as existing data, drive wear, or a SATA III bottleneck—can lower the observed speed. Verify that the drive is connected to a SATA III (6 Gb/s) port and that the test file size is large enough to bypass cache effects.

Can I benchmark an external SSD?

Yes, but the results will reflect the bandwidth of the external interface (USB 3.1 Gen 2, Thunderbolt 3, etc.). For a fair comparison with internal drives, use the fastest connection your external enclosure supports, and note the interface in your report.

Conclusion

Benchmarking your SSD with CrystalDiskMark is a quick, inexpensive way to verify that the drive is performing as expected and to catch any hidden issues before they affect daily use. By following the steps outlined above—downloading the tool, preparing the drive, configuring sensible test parameters, and interpreting the results—you’ll gain confidence in your storage subsystem and be equipped to troubleshoot performance problems. Remember to avoid common mistakes, keep your firmware up to date, and revisit the benchmark after major system changes. Happy testing, and may your read/write speeds stay blazing fast!

Photo by Samsung Memory on Unsplash