CrystalDiskMark is an open-source Windows storage benchmark that measures sequential and random read/write performance, helping users compare drives, verify upgrades, and investigate transfer speeds under configurable test conditions.
CrystalDiskMark is a Windows benchmarking utility that measures how quickly a storage device reads and writes data under selected workloads. It helps PC builders check a newly installed SSD, technicians investigate unexpectedly slow transfers, and owners compare storage configurations on the same computer. Sequential and random results describe different access patterns, so the interface provides more context than a single headline speed. You select a target drive and test settings, run the benchmark, and review the resulting measurements. The program is designed for controlled performance checks rather than continuous monitoring, and its scores describe the chosen test rather than every application workload.
CrystalDiskMark results depend on the selected workload and the system surrounding the drive, so consistent settings are essential for comparisons. Test file size, controller behavior, and competing disk activity can affect the numbers you see. Pairing a speed test with CrystalDiskInfo helps distinguish measured storage performance from the drive health information used during troubleshooting. Keep the same major benchmark version when comparing results because the developer warns that different major versions are not directly compatible. Network drives require running without administrator privileges. Write tests also consume flash write endurance, making a planned comparison more useful than repeatedly rerunning tests without a specific question.
CrystalDiskMark helps you verify whether a storage change produced a measurable improvement before you spend more time or money on further upgrades. A baseline taken before replacing a drive, changing its connection, or adjusting a configuration gives you something concrete to compare afterward. Looking at separate access patterns also helps avoid choosing storage solely by its largest advertised transfer figure. For support conversations, recording the workload alongside the result makes a report more useful than saying that a computer feels slow. Technicians can repeat the same configuration to investigate whether a difference persists, then decide which component or setting needs closer examination. Saved results create a practical record of tested conditions across maintenance sessions. These benefits depend on disciplined comparisons: hold relevant settings steady, account for other activity, and use the benchmark to answer a defined performance question rather than treating a higher score as proof of better performance in every task.
Key capabilities and highlights.
CrystalDiskMark measures sequential reads and writes to show how a drive handles contiguous data transfers. These results are useful when investigating workloads dominated by large files. Read and write speeds are reported separately, helping you identify whether retrieving existing data or writing new data is the more limited operation.
Random tests measure access to data at scattered locations rather than one continuous stream. They provide a different perspective from sequential throughput, particularly when examining small-block storage activity. Comparing random results under matching settings helps reveal differences that a large-file transfer figure alone can conceal during a storage evaluation.
Mixed profiles exercise reads and writes together instead of measuring each direction only in isolation. This adds context for situations where storage must service both kinds of requests. The Standard Edition includes profiles with mixed testing, allowing you to examine another workload without assembling a separate benchmark procedure yourself.
CrystalDiskMark offers profiles for different measurement goals, including Peak Performance and Real World Performance. Version 9 also adds a Flash Memory profile. Selecting a profile provides a repeatable starting configuration, while recording its name helps prevent accidental comparisons between results produced by substantially different workloads or measurement goals.
Test controls let you choose the target drive, run count, and test file size, with further settings for block size, queue depth, and threads. These options let you define a workload deliberately. Keeping them unchanged across comparison runs makes it easier to attribute differences to the storage configuration being investigated.
Results can be displayed as transfer rates, input/output operations per second, or average latency. Each view emphasizes a different aspect of storage behavior: data volume, request processing, or response time. Selecting the appropriate unit helps you describe the tested workload more precisely than quoting a speed figure without its measurement context.
The test-data setting supports random data and zero-filled data, letting you control what the benchmark transfers. Some SSDs produce different results depending on the data pattern. Recording this setting alongside the workload makes comparisons more transparent and helps explain why otherwise similar tests can return noticeably different numbers.
CrystalDiskMark can copy benchmark results to the clipboard or save them as text and image files. This makes it easier to attach evidence to a support request, maintain an upgrade record, or compare tested configurations. Exporting the result also avoids transcription mistakes when preserving several read and write measurements.
Browse historical releases and previous builds.
Common questions about CrystalDiskMark.
Yes. The software uses the MIT license, which permits commercial use, modification, and redistribution subject to its notice requirements.
Select the intended drive, choose the test size and run count, then click All to run the selected set of tests. Close unnecessary applications that actively use the target drive, and record the profile and settings alongside the results.
Advertised figures may use different workloads, queue depths, test sizes, or hardware. Your connection, controller, CPU, and other disk activity can also affect results. Compare equivalent test conditions before concluding that the drive is underperforming.
No. Throughput and latency measurements assess performance under a test workload. They do not establish physical condition, remaining endurance, or the absence of storage errors.
The developer warns against direct comparisons across major versions. For a controlled comparison, use the same major version, profile, test size, data pattern, and workload settings on each drive or configuration.
CrystalDiskMark is an open-source Windows storage benchmark that measures sequential and random read/write performance, helping users compare drives, verify upgrades, and investigate transfer speeds under configurable test conditions.
TreeSize is a disk space analyzer that scans drives, folders, and network shares, then shows exactly which subfolders and files are consuming storage so you can decide what to delete, archive, or move.
Microsoft .NET Windows Desktop Runtime 8.0 provides the shared components required to run compatible Windows Forms and WPF desktop applications without installing the complete .NET software development kit.
Microsoft .NET Runtime 8.0 provides the execution engine and core libraries needed to run compatible .NET applications, especially framework-dependent console programs, without installing the complete software development kit.
Adobe Acrobat Reader (64-bit) is a free Windows PDF viewer for opening, printing, searching, commenting on, filling, signing, sharing, and reviewing PDF documents without a paid Acrobat subscription.
Comments