10 Free Tools to Measure Hard Drive and SSD Performance
One of the most important hardware components inside your computer is obviously the drive that holds the operating system. If you have Windows installed on a slow hard drive, it doesn’t matter how powerful the other components like CPU and memory are, the system will boot slower, load programs more slowly, and lag quite badly while multitasking.
This is a reason why high performance hard drives and especially SSD’s are so popular these days. It’s because upgrading just that one part can breath new life into a computer and make it feel a lot more snappy to use. The actual hard disk or SSD performance under Windows is determined by several factors so how do you know if your drive is performing well, needs tweaking or is even holding the system back?
Thankfully, the internet is full of websites to show you how well nearly every drive around can perform. Also, loads of programs are available that can give your hardware a performance test to see how well it’s doing. There are many paid suites around like PCMark or Passmark that can test the whole system, but here we list for you 10 free tools that are specifically designed for testing the performance of hard drives and SSD drives.
All tools were tested on Windows 10 64-bit and are also compatible with Windows 7 and 8.
1. CrystalDiskMark
This is a very popular storage benchmarking tool because it’s versatile and can produce good results for just about everything from USB drives, to memory cards, RAMDisks, SSD drives and mechanical hard drives. CrystalDiskMark is very easy to operate too, just set the test size between 50MB and 4GB, the drive to test and the number of passes to run. More passes obviously should produce more accurate results.
Then click All to run a full test or any individual colored buttons below to run single tests. A useful extra option for SSD drives is whether to fill the test data randomly or with 0’s or 1’s. This will affect the results on drives with hardware compression such as those with Sandforce controllers. Portable, installer and custom skin versions are available.
2. ATTO Disk Benchmark
ATTO Disk Benchmark is a popular portable tool used by many hardware review websites and is also recommended by manufacturers such as Corsair to run speed tests on SSD drives. All the tests are sequential and are taken for read and write operations using block sizes of 512 bytes up to 64MB. A test file length of between 64KB and 32GB is also available, both are selectable from drop down menus.
Leaving Direct I/O enabled will rule out odd results due to any system caching. An I/O size of over a couple of Megabytes will produce virtually the same scores so probably isn’t needed for many users. Results can be saved out and loaded again at a later time. A look around the internet will likely find someone else who has posted ATTO results for similar hardware to your own.
Download ATTO Disk Benchmark (via Softpedia, the official site requires filling a form)
3. AS SSD Benchmark
AS SSD looks a bit like CrystalDiskMark and is designed primarily to benchmark SSD drives. It is another popular tool being used by hardware sites such as AnandTech’s SSD benchmark charts (which also includes ATTO) to show their results. The program uses incompressible data so some SSD’s will show much lower scores than usual if they compress their data.
Sequential and 4KB read and write scores will be shown along with access times and a final general overall score. The results can be changed to IOPS if you prefer from the View menu. A couple of useful additional benchmarks are available in the Tools menu like a Copy benchmark which simulates copying an ISO, game and a program, and also a read/write compression benchmark. AS SSD is also completely portable.
Download AS SSD Benchmark (website is in German, download link at bottom)
4. Anvil’s Storage Utilities
For an SSD or hard drive benchmark and test utility that really puts drives through their paces, they don’t come much more comprehensive than Anvil’s Storage Utilities. While being able to perform a complete read and write test using sequential and random operations, it displays a full set of results including response time, speed in MB/s and also IOPS (Input/Output Operations Per Second).
A total combined read/write score is shown in the yellow box. Also included are 3 extra IOPS tests, an endurance testing function and the ability in the settings to alter the amount of compression to use on the test file. More functions like a system information tab and a screenshot saver round off an impressive tool. Anvil’s Storage Utilities doesn’t seem to be developed anymore and the last version is from 2014.
Download Anvil’s Storage Utilities
5. HD Tune
HD Tune is probably the most well known hard disc drive benchmarking and diagnostic utility and will likely be in every tech users USB toolkit. The free version 2.55 is getting old now having not been updated since 2008 and might have minor issues with some of the latest hard drive models. Some functions like error checking and benchmarking should still work fine.
After a test, the benchmark result graph will show the minimum, maximum and average read speed along with the average access time in milliseconds and the burst rate. The block size can be changed in the options from 512 bytes up to 8MB and a slider can move between faster less accurate or slower more accurate test speeds.
6. HD Tach
HD Tach is a rather old hard drive benchmarking tool dating from 2004, but it still produces good results for today’s drives. However, there are a few minor compatibility issues and the program itself needs to be run in XP compatibility mode for Windows Vista or above. Long (32MB) or Short (8MB) block tests can be chosen from the drive selection window and the results will pop up within a minute or two.
The results window shows a read performance graph across the whole drive in addition to a burst speed chart and information for CPU utilization, average access time and also the average read speed. There is a drive comparison button where you can view benchmarks for other drives but they’re mostly old scores from early 2000’s era hardware. HD Tach can be made portable if you extract the setup executable with Universal Extractor.
7. Disk Thruput Tester
DiskTT is a small and portable testing tool that uses Windows core read and write functions to write a temporary file to the specified drive. Then it’s read back sequentially and randomly to get three resulting scores. The test file can be between 10MB and nearly 100GB with a block size to test of 1KB up to 64MB.
To test a different drive to the default of C, change the path of the test file to “User defined path” and select a drive. Windows file caching settings are available as an option but it will often skew the results if they are not set correctly, so are best left alone unless you have a specific reason. To only run a random read test click the button where it says read/write.
Download Disk Throughput Tester
8. Roadkil’s Disk Speed
Roadkil makes a number of small and useful utilities, Disk Speed is another one that simply does the job you expect with no frills or bloat (it’s only ~90KB). It’s a bit old dating from 2009 but the program is simple to run, just select the logical drive letter or physical drive number and press the Begin Test button.
Disk Speed only runs read tests but you will get a range of scores from 512 byte blocks up to 1MB blocks for both random reads and linear reads for each block size. The “Results in Brief” box will show the average access time, maximum read speed, cached (burst) speed and an overall score which can be compared online at Roadkil’s website with the same or similar model drives.
9. HD Speed
HD Speed is small and portable at around 90KB, but has more configuration options to suit different storage devices. It also has Read+Write and Read+Write+Verify modes, but be aware the data on any device you want to write test WILL BE DESTROYED. Make sure to right click and run this program as Administrator, even if you are an Administrator. If you don’t, all the drives won’t show up and there will a several second delay while starting the program.
Something useful not found on other testing tools is the option to test a specific position on the drive, 0% for the faster outer part of a mechanical hard drive, 100% for the slower inner. To run a benchmark and get the average speed on the graph, choose the drive, enter the time to run it in the box and press Start. Block size can be left at Auto or changed from 1KB up to 16MB, a log file can also be created to review the results.
10. DiskMark
DiskMark is a nice tool to run because it displays plenty of information about average, minimum, maximum and last read and write scores. All the data is laid out in both raw number format and a live graph. The one slight problem with the program is actually configuring the test to perform.
You can obviously leave it at the default of 64KB chunks, in an 8MB test file which runs 320 times. To get another score such as using 1MB or maybe 4KB blocks, you have to enter the Chunk Size, then enter the IOs (number of blocks to make up the test file), and finally choose the number of times the test file is read and written to. Incorrect setup will either produce a very short inaccurate test or a test which takes far too long. Separate 32-bit and 64-bit versions are available.
Benchmark Notes: For less experienced users, a lot of the numbers thrown up by any benchmarking tool can look a bit confusing because you might not know what you’re supposed to take notice of. A few simple things to look for are:
Sequential (also seq or linear) – This is the speed at which a drive can read or write continuous data in the form of large files. It’s better for data storage drives or games partitions to have a higher sequential speed where most files will be at least a Megabyte in size. Sequential tests will show the maximum possible read and write speeds a drive can achieve.
4K – The 4K results produced by benchmark tools show how fast a drive is at accessing small files, and are more important for a Windows boot drive. Most notably random 4K read and write speeds show how quickly the files are accessed when scattered across a partition. Higher 4K results will also produce a better and smoother Windows during multitasking. When looking at buying an SSD drive to install Windows and applications, compare the 4K scores first and not the headline grabbing sequential speeds.
AS SSD Benchmark is always in German language,no matter how often i click on the english version tab.
Is there a bootable type of SSD benchmark software? I find that windows based softwares are inconsistent.
Pretty much all drive reviewers I’ve seen use Windows based tools so they can’t be that inconsistent. I think it could be more to do with what’s running on your system, such as antivirus, Windows updates/tasks, or third party software doing things in the background while a disk test is running.
I don’t recall ever seeing a disk benchmark software that doesn’t require a full OS to run, whether it’s Windows or Linux. Obviously bootable Live CDs still run an OS although it will usually be stripped down to a large degree.
Hello, i am looking for a Hard Disk/Drive utility that can scan and “repair” weak sectors.
Also interested to know if there is a tool that scan a hard drive surface from back to front and also randomly perhaps?
Thanks
James
Bad sectors are usually a sign that your drive is starting to fail. If you have a high number of bad or pending sectors it is highly recommended that you replace the drive rather than attempt to repair it.
WD Data Lifeguard is a free disk diagnostic tool from Western Digital. It can run a “basic test” (checking SMART info), an “extended test” (full scan that locates bad sectors), or completely wipe the drive.
If bad sectors are found during an extended test, you will be given an option to attempt to repair these sectors. However, like I said above, it is recommended that you replace a failing drive.
You can also use CrystalDiskInfo (from the makers of CrystalDiskMark) to display SMART information and monitor drive health in real time. CDI once saved me from having to pay for expensive data recovery by showing an alarm for pending sectors before my drive died 2 days later.
HD Tach cannot run on my Windows 7 64bit system, it says that only could be run on Win2K or WinXP.
Read the article again and you will see that it does work and what you need to do to get it to work.
Here is a simple disk benchmark tool written in Java I’ve been working on. Would appreciate any feedback anyone is willing to provide. Best wishes,
sourceforge.net/projects/jdiskmark/
I’ve written a simplistic tool in Python to benchmark HDDs and SSDs. It is completely open-source and now in alpha stage, though you already could use it, and if interested, participate in development.
Here’s the link: github.com/thodnev/MonkeyTest
CrystalDiskMark installs a hidden program on your computer that forces your browser to use a particular search engine.
That’s because you are downloading the version with ads, get the one without ads to not have it install opencandy adware.
So how do you know which version your are downloading? With or without ads?
In fact, there is no version with ads any more, the author removed opencandy earlier in 2016.
Thanks for the heads up,
I was looking to download and use CrystalDiskMark, but not now!!
Why are you taking notice of a 3 year old comment when a newer comment is telling you that problem doesn’t exist anymore and hasn’t existed since 2016??
@ relcocq
SSD drives aren’t effected by fragmented disk, so no shouldn’t effect the results
You should never run de-frag programs on a SSD, ever.
Yes and no. Defrag on a HDD, from Win 7+, it run Optimize, which is the trim command. This is very important. Why? simple.
An HDD is seperate into block and then into pages. It can only do 3 operations on block: Read, Write and Erase. It cannot overwrite like a hard disk. But it function the same way: When you delete something, it marks the pages related to these data as available, but not empty. A problem though, on page, it can only do Read and Write, not erase.
When you have to erase a page because you run out of empty page, the ssd controller must copy all the good page in cache from the block (Read action + write action in the SSD cache), erase the block so all pages are empty and write back the good page from the cache. This operation take time and this is way your SSD speed degrate with usage.
Thus, in the early days of SSD, Trim appeared to fix this problem. When Trim is run, it does the operation above on all needed block and pages, thus leaving the SSD only with good and empty pages. This process does add some wear on the drive, but very little.
With time, OS and controller have implemented various way to bypass this problem, like doing the trim action on each erase, never leaving bad pages.
More and reference at: anandtech.com/show/2738/8
I did not see Hard Disk Sentential among your programs. I’ve been using it for years now.
3 years back it told me the bad sectors of the 750GB hard disk in my Dell 545 Inspiron had increased and its performance had dropped to 80%.
The system was still under warranty so I contacted Dell. After a lengthy phone call discussing the problem, they a sent a technician out. After 90 minutes, he agreed the drive was problematic, and gave me a 1TB replacement drive.
I installed the new drive, made it the boot and moved everything important over to it.
I made the 750GB drive a secondary just to see how long it might last. About one week later it failed.
I am sure those programs you list are great but I am thankful Hard Disk Sentential was recommended to me and have it on all of our computers.
Hard Disk Sentinel is a completely different type of software and is not a tool specifically for bench marking your drives.
Besides, it isn’t even free and this article is specifically about free software for the job.
Installed HD Tach on Win 8. It will not run and displays a message that it will only run 2000 or XP.
Go to the Simpli Software program files folder and run the hdtach.exe in Windows XP compatibility mode, works fine then.
Thanks Ray!!!
Has anyone considered whether or not a fragmented disk would effect the results?
I would think so.
Rob.