A disk performance test measures HDD and SSD read/write speeds. This guide shows how to use MiniTool Partition Wizard to measure that and read the results using IOPS, transfer rate (throughput), sequential and random read/write speed, and queue number.

What Is a Disk Performance Test

A disk performance test measures how quickly an HDD or SSD reads and writes data under a selected workload. You can use it to establish a baseline for a new drive, compare storage devices, or investigate a change in read/write speeds.

Two useful metrics are throughput and IOPS. Throughput describes the amount of data transferred per second, while IOPS stands for input/output operations per second. Both metrics need to be interpreted alongside the test settings.

The following guide shows how to measure disk performance with MiniTool Partition Wizard and explains how to choose settings and interpret the results.

How to Measure Disk Performance with a Disk Benchmark Tool

MiniTool Partition Wizard includes a Disk Benchmark feature for testing storage performance. It can be used to measure the read and write speeds of HDDs and SSDs under different test settings.

MiniTool Partition Wizard FreeClick to Download100%Clean & Safe

Here is step-by-step guide to disk benchmark in MiniTool Partition Wizard.

Tips:
Before testing, pause file transfers and other disk-intensive tasks. To compare results, record the drive model, connection type, and settings. Disk Benchmark creates a temporary test file and deletes it when done; test time depends on the settings.

Step 1: Download and launch MiniTool Partition Wizard to get its main interface. And then click Disk Benchmark on the toolbar.

launch MiniTool Partition Wizard and click Disk Benchmark on the toolbar

Step 2: Select a drive and set parameters to test the disk performance.

Select a drive and set parameters to test the disk performance

Parameters:

  • Select a drive: Computer users select a local disk or a volume to test its performance. The physical disk on where the local disk is located will be listed afterwards.
  • Transfer Size: It ranges from 1 KB to 2048 KB.
  • Total Length: It ranges from 100 MB to 4096 MB.
  • Queue Number: It ranges from 1 to 512.
  • Thread Number: It ranges from 1 to 64.
  • Test Mode: There are three modes for users to choose: Sequential, Random, and Sequential & Random.
  • Cool Down Time: It is used to reduce the hard drive temperature before starting the next test.

Recommended settings for beginners: set Transfer Size to 4 KB, keep Queue Number at 8 or below (use 1 as the main reference), and set Thread Number according to your CPU cores. Run both Sequential and Random modes for a complete disk speed test.

When the above parameters are set, the disk benchmark will show you the read and write speed of the hard drive. The result is shown as follows.

Step 3: Click Start after the parameters are set. Then, wait seconds to get a disk performance test result.

wait seconds to get a disk performance test result

How to Set Parameters in Disk Benchmark

The whole operation process is very simple. But what’s difficult for common users is to decide the parameter values and interpret the result.

Transfer Size

As we all know, data package will be divided into multiple blocks in input or output transmission process. And then these blocks are transferred. So, computer users can determine the block size in data transmission process by setting transfer size value.

The transfer size will affect the IOPS and Throughput. Generally speaking, there are more IOPS when transfer size is smaller, and there are fewer IOPS if transfer size is larger. Besides, there is a formula about IOPS, Throughput, and transfer size: Throughput = IOPS * Transfer Size.

Generally, bigger IOPS value and big transfer size all can bring in bigger Throughput value to some extent. However, there is an inverse relation between IOPS and transfer size.

For example, as the result picture shows, if the hard drive sequentially reads 200 MB data (total length) with 4 KB transfer size, the disk has a sequential reading speed at 33 MB/S. Then the corresponding sequential read IOPS = 33MB / 4KB.

If calculating IOPS one-by-one, it’s found that the IOPS values are about 8000, 7000, 5000, respectively when the transfer size are 4 KB, 8 KB, 16 KB, respectively. However, it is also found that the disk read/write speed (Throughput) is lowest instead when the IOPS value is biggest.

the relationship between IOPS, Throughput, and I/O size

So, computer users should know what value is important.

Actually, when testing the performance of HDD, computer users should pay attention to Throughput value in sequential reading/writing mode with big transfer size. Otherwise, when testing the performance of SSD, computer users should pay attention to IOPS value in random reading/writing mode with small transfer size.

However, due to advanced format technology, 4K alignment can improve disk performance greatly. So, the transfer size value is usually a multiple of 4 KB. So, when running the disk performance test, it is suggested for common computer users to select 4 KB as the minimum value.

Besides, if computer users often run applications with frequent random reading and writing requests like OLTP (Online Transaction Processing), IOPS is an important index. If computer users often run applications with frequent sequential reading and writing requests like VOD (Video On Demand), they should pay more attention to Throughput value.

Then, what’s the difference between sequential reading/writing mode and the random reading/writing mode?

Sequential Read/Write Mode

It is usually applied in reading/writing big files. These big files are stored on disk sequentially. And they are also read sequentially by read-write heads. Thus, the seek time and rotational delay time are reduced.

Random Read/Write Mode

It is usually applied in reading/writing small files. These small files are usually stored on disk randomly and discretely. When they are read/written by heads, the seek time and rotational delay time are greatly increased. This performance usually affects computer boot speed, large program booting efficiency, file search efficiency and so on.

Surely, the disk speed test in sequential & random mode is closer to life than other modes. This indicator also should not be ignored.

Thread Number

This value is usually related to CPU. It is related to the kernel number of the computer and Hyper-Threading technology.

Generally speaking, the thread number is determined by the core number. If your computer has two kernels, then the thread number is two. You can check it in your computer. Right click the taskbar, and select Task Manager. Then navigate to Performance to check Cores number.

Tip: In Windows 7, the cores number won’t be displayed directly, but you can get it by counting how many CPU graphs are there.

check processor number or thread number

However, Hyper-Threading technology (it is a string of codes) can divide a kernel into two threads. And computer operating system will regard one thread as one kernel. Thus, there are more threads to run more programs. Then you can check the real kernel number and thread number through the following way.

Press “Windows + R” keys and type “cmd” to open the command window. Then type “wmic” and press “Enter” key. And then type “cpu get” and press “Enter” key again. Then you can get the number of cores and number of logical processors (thread).

check number of cores and number of processors

The hyper-Threading technology can only be applied on Intel CPU. And it can be disabled in BIOS interface.

The thread number can be changed by the following steps. Press “Windows + R” keys and type “msconfig”. Navigate to Boot page and click Advanced Options to change the number of processors.

modify processor number

Queue Number

It is the number of I/O requests that are capable of waiting for service in port queue. It touches upon TCQ (Tagged Command Queuing) technology and NCQ (Native Command Queuing) technology. NCQ technology is specially applied in SATA disk. However, TCQ technology is applied more widely.

NCQ and ATA TCQ (PATA, SATA) can support at most 32-class queue length, while SCSI TCQ can support at most 256-class queue length. However, in actual computer operation, the queue length will not exceed 8. (High queue length will increase the I/O response time.)

So, common computer users are suitable to choose a queue length that equals to or is less than 8. Besides, computer users should mainly refer to the disk performance result with queue length that equals to 1. (The commonly defined queue length is 1.)

How to Optimize Disk Performance

Although a hard drive’s theoretical performance is defined, computer users can do many things to optimize the disk performance in actual operation. For example, computer users can defrag Windows and set disk cache.

As for disk defragmentation, please click How to Defrag Windows 7 to Improve Computer Performance to get a detailed guide.

As for disk cache, follow the following tutorial to set up. Press “Windows + R” keys and type “regedit” to open Registry Editor. Navigate to HKEY_LOCAL_MACHINE > SYSTEM > CurrentControlSet > Control > Session Manager > Memory Management to seek the IoPageLimit value.

Tip: If there is no IoPageLimit value, computer users can create one. Right click Memory Management and choose New > DWORD value. And then type its name as IoPageLimit.

And then, right click IoPageLimit to modify it. Select Decimal and type cache size in KB. 16384 (16 MB) and 32768 (32 MB) are recommended. Then, save the configuration and restart computer.

modify cache size

Surely, there are other methods that can optimize disk performance, please click Free Hard Disk Optimization Software Helps Optimize Disk Performance to know more.

If the disk has very poor performance and computer users can’t improve it, they should consider buying a new one.

When computer users buy a new disk online, the disk product names are like the following one.

WD Blue 1TB PC Hard Drive – 7200 RPM Class, SATA 6 Gb/s, 64 MB Cache, 3.5. —- A disk product name from Amazon

Hard drive vendors know well that their customers usually take disk performance into consideration when buying a hard drive. So, they will post the basic information about their hard drives like the above one.

We can know the following information from the name. This is a WD Blue PC hard drive. Its rotational speed is 7200 RPM, which will affect disk performance through affecting data access time. Its capacity is 1 TB and its port is SATA 6 Gb/s. It has 64 MB cache and the number 3.5 indicates its physical size.

Here is a guide to disk performance test. If you want to benchmark your disk, this article can help you.Click to Tweet

Conclusion

A disk performance test helps you check the real speed of your HDD or SSD and decide whether it needs optimization or replacement.

With the free Disk Benchmark in MiniTool Partition Wizard, you can measure disk performance in minutes and compare sequential and random results. If you have questions, leave a comment below or contact us via [email protected].

Disk Performance Test FAQ

What is a good SSD or HDD speed test result?
There is no single speed threshold that applies to every drive. A reasonable result depends on whether the device is an HDD, SATA SSD, or NVMe SSD, as well as its model, interface, workload, and test settings. Compare benchmark results with manufacturer specifications and measurements from comparable devices.
Does a hard drive performance test check drive health?
Not by itself. A hard drive test focused on speed measures performance under the selected workload. Health checks examine different information, such as supported health attributes and error indicators.
Why is my SSD or HDD benchmark slower than expected?
Unexpectedly low results may be caused by background activity, interface limitations, thermal conditions, drive configuration, or hardware problems. Repeat the test under consistent conditions, check the connection and drive health, and compare the results with the manufacturer's specifications.
Does a disk benchmark affect my data?
A read/write benchmark can perform write operations and may create temporary test files, depending on the tool and test mode. Use a reputable tool, read its instructions before testing, and back up important data before using unfamiliar disk utilities. Do not assume that every benchmark is entirely read-only or risk-free.
  • linkedin
  • reddit