Benchmark of Top Ten Registry Cleaners
These tests were performed in October of 2016
In recent years, Macecraft Software has published a total of five registry cleaner product comparison studies which have attempted to answer questions about the quality of different registry cleaning products as well as what kind of effects they can have on your system.
The testing was performed using an Oracle VirtualBox (version: 5.0.26 r108824) virtual environment running Windows 10 with 4 CPU cores, 8 GB of RAM and without any internet connectivity. The virtual machine was run on a host computer running Ubuntu Linux 16.04 LTS (64 bit) with Intel Quad Core i7-6700K CPU @ 4 GHz, 32 GB RAM with a system disk of Samsung SSD 850 EVO.
The products tested were chosen simply with a Google search of Windows registry cleaners and the top ten results were picked.
All possible measures have been taken to ensure the verifiability and credibility of these these benchmark results. For instance, all the testing has been recorded in both video and screenshots, links to see these recordings for each tested software are provided below.
The following products were tested in this benchmark of registry cleaners. Each software was downloaded from their official website and installed using their default settings.
|Software Product||Version||Installer file size||Installation size||License||Comments|
|Auslogics Registry Cleaner||22.214.171.124||7.76 MB||27.2 MB||Freeware||Displays a browser popup. Bundled software: Auslogics BoostSpeed.|
|AVG PC TuneUp||126.96.36.199637||3 MB||254 MB||15 day trial||Online installer. Browser popup.|
|CCleaner||5.22.5724||8 MB||18.3 MB||Free for non-commercial use||Browser popup.|
|CleanMyPC Registry Cleaner||4.5||3.3 MB||7.5 MB||15 day trial||Seems rebranded version of Registry Clean Expert.|
|Comodo System Utilities||4.9||2.7 MB||4.8 MB||15 day trial|
|Eusing Free Registry Cleaner||4.0 (build: 20160918)||0.9 MB||1.9 MB||Freeware|
|jv16 PowerTools||188.8.131.528||9.9 MB||13.4 MB||60 day trial||Release Candidate level software.|
|Little Registry Cleaner||1.6.0||5.4 MB||6.4 MB||Freeware||Requires the download of .NET framework.|
|Registry Clean Expert||4.9||2.7 MB||4.8 MB||15 day trial||Seems rebranded version of CleanMyPC Registry Cleaner.|
|Wise Registry Cleaner||9.31.599||4.2 MB||8.0 MB||Freeware||Browser popup. Bundled software: SpyHunter|
Test 1: Number of errors found
Some popular software were installed to the Windows 10 test environment, such as 7-Zip, Audacity, Google Chrome, IrfanView, K-Lite Codec Pack, Norton Security, Notepad++, OpenOffice, Paint.net, SUPERAntiSpyware and VLC Media Player. After these software were installed, all the registry cleaner products were installed to the same system and ran using their default settings.
|Software Product||Errors found|
|Auslogics Registry Cleaner||353|
|AVG PC TuneUp||164|
|CleanMyPC Registry Cleaner||252|
|Comodo System Utilities||n/a|
|Eusing Free Registry Cleaner||143|
|Little Registry Cleaner||276|
|Registry Clean Expert||269|
|Wise Registry Cleaner||258|
It was not possible to test Comodo System Utilities, the button to start scanning remained disabled and could not be clicked. CCleaner’s results only include the number of registry errors it reports, the software’s file system cleaning module also found 633 temporary files. Wise Registry Cleaner was used using its deep scan feature.
If you wish to verify these results, you can view the video recording of this testing session here: (video no longer available)
Test 2: Windows startup time measured with a stopwatch
Using the same test environment with all the same software as in the previous test, the system’s startup times were manually measured using a stopwatch. This was done before and after the system was cleaned with the registry cleaner products.
To establish a baseline, the system was restarted four times and its startup time was recorded. The startup times were: 107, 104, 95 and 108 seconds, resulting a median startup time of 105.5 seconds.
Next, the system was cleaned by each registry cleaner product, the system was restarted one time to apply the changes and restarted two times with the startup time measured. After each software was used, the system was restored to its original state using the virtual machine’s snapshot restore feature.
Cleaning the Windows 10 system with registry cleaners had the following impact on the system’s startup time:
|Software Product||First startup after cleaning||Second startup after cleaning||Median startup time after cleaning|
|Auslogics Registry Cleaner||102 sec.||102 sec.||96.5 sec.|
|AVG PC TuneUp||94 sec.||105 sec.||99.5 sec.|
|CCleaner||109 sec.||96 sec.||102.5 sec.|
|Eusing Free Registry Cleaner||109 sec.||101 sec.||105 sec.|
|jv16 PowerTools||72 sec.||74 sec.||73 sec.|
|Wise Registry Cleaner||98 sec.||108 sec.||103 sec.|
It was not possible to test CleanMyPC Registry Cleaner or Registry Cleaner Expert, because these products do not allow cleaning of any found errors without buying the full product. It was not possible to test Comodo System Utilities, because its button to start the scan remained disabled. Trying to clean the found items with Little Registry Cleaner caused it to crash, therefore it could not to be tested.
As can be seen from the results above, most registry cleaners had a minimal, under 5% improvement to the system’s startup time. Since the startup time measurements were done by hand using a stopwatch, any improvement under 5% should be considered clearly within the margin of error. Three products provided a measurable performance improvement to Windows startup times, there products were Auslogics Registry Cleaner, AVG PC TuneUp and jv16 PowerTools, which improved the startup time by 8.53%, 5.69% and 30.81%, respectively.If you wish to verify these results, you can view the video recording of this testing session by clicking the video link from the results table above. The video recording of measuring the system’s baseline startup time, i.e. startup time before cleaning, can be seen here: (video links no longer available)
Test 3: Windows startup time measured with software
It must be noted that it is considerably challenging to accurately measure computer’s startup time manually by using a stopwatch, as what was done in the previous test.
Therefore, the test was repeated using a software to measure the computer’s startup speed. The new jv16 PowerTools features a tool called the Startup Timer, which, when enabled, measures accurately the time it takes Windows to fully start.
Furthermore, it is safe to assume that one can see the results of running a registry cleaner more clearly the more there is something to clean in computer. That is, using a registry cleaner in a system that has very little software and nothing to clean is most likely unneeded.
This is why the Windows 10 test environment of the previous test was changed in a way more software were installed to the system.
Examples of software installed: Adobe AIR, Adobe Shockwave Player, AIM 7, BitTorrent Sync, Classic Shell, CutePDF Writer, eMule, Everything, FastStone Image Viewer, FileZilla Client, Foobar2000, Greenshot, ImgBurn, InfraRecorder, Java, Launchy, MediaMonkey, Microsoft Silverlight, Mozilla Firefox, Opera, Putty, Python, qBitTorrent, Steam, SumatraPDF, TeamViewer 11, TeraCopy, Trillian, VNC Server, Winamp, WinDirStat, WinMerge, WinRar, WinSCP and XnView.
To establish a baseline, the system was restarted multiple times and its startup time was recorded by jv16 PowerTools’ Startup Timer. The last four startup times were: 139, 110, 151 and 163 seconds, resulting a median startup time of 145 seconds. As expected, installing more software to a Windows system caused it to work slower.
Next, the system was cleaned by each registry cleaner product, the system was restarted one time to apply the changes and then restarted four more times and the startup time was recorded using jv16 PowerTools’ Startup Timer feature. After each software was used, the system was restored to its original state using the virtual machine’s snapshot restore feature.
Cleaning the Windows 10 system with registry cleaners had the following impact on the system’s startup time:
|Software Product||Errors cleaned||Median startup time after cleaning|
|Auslogics Registry Cleaner||579||144 sec.|
|AVG PC TuneUp||348||147 sec.|
|CCleaner||213 (plus 331 temp files)||133 sec.|
|Eusing Free Registry Cleaner||173||135 sec.|
|jv16 PowerTools||4299||128 sec.|
|Wise Registry Cleaner||525||141 sec.|
It was not possible to test CleanMyPC Registry Cleaner or Registry Cleaner Expert, because these products do not allow cleaning of any found errors without buying the full product. It was not possible to test Comodo System Utilities, because its button to start the scan remained disabled. For AVG PC TuneUp, CCleaner and Wise Registry Cleaner, all of their automated registry and system cleaning as well as system optimization features were used with their default settings. Of jv16 PowerTools, its Clean And SpeedUp My Computer feature was used with its default settings. Trying to clean the found items with Little Registry Cleaner caused it to crash, therefore it could not be tested.
This test seemed to provide quite a challenge to the registry cleaner products with all products finding well over a hundred issues to fix from the test environment.
As can be seen from the results table above, most products were able to provide an improvement to the system’s startup speed. With three products, namely Auslogics Registry Cleaner, AVG PC TuneUp and Wise Registry Cleaner, the speed improvement was under 5%, which can be argued to be within the margin of error of this type of testing. With the other three products, namely CCleaner, Eusing Free Registry Cleaner and jv16 PowerTools the speed improvement was over 5%, to be exact the improvement in speed was: 8.28%, 6.90% and 11.72%, respectively.This test demonstrates that at least in a case of computer with many installed software, running a registry cleaner product can have a measurable impact on the system’s startup speed. In the case of the product that provided the biggest improvement, that is, jv16 PowerTools, the speed improvement on median startup time was 17 seconds.If you wish to verify these results, you can view the video recording of this testing session by clicking the video link from the results table above. The video recording of measuring the system’s baseline startup time, i.e. startup time before cleaning, can be seen here: (video link no longer available)
The conducted benchmark tests seem to indicate that overall, many registry cleaner products had an effect to the average system startup time.
The first study of this kind dates back to 2009 and was entitled “Registry Cleaner Review and Performance Comparison Study”. A follow-up to this research was released later in the same year. It was Macecraft Software’s “In-Depth Review of 31 The Most Popular Registry Cleaner Products” study.
In 2010, a Comparison of 176 Registry Cleaner Products study was published.
In September of 2013 we did a study from a different point of view: this time it focused on whether or not running a registry cleaner software results in changes to system performance. To measure this, we used globally recognized benchmark scoring software to measure the performance of the test system before and after using a registry cleaner. That study was entitled the Registry Cleaner Benchmark Study.
The most recent study was published in March of 2015 and it further continued to discuss actual, measurable effects that different products claiming to optimize Windows have on a computer’s performance. This study was entitled the Benchmark Report of 133 Windows Optimization Software Products.
This current case study aims to continue the tradition that Macecraft Software has established while releasing similar product comparison test reports alongside every recent major product release. The latest benchmark test report currently covers the following tests completed in October 2016:
- Test 1: Number of errors found
- Test 2: Windows startup time (measured with a stopwatch)
- Test 3: Windows startup time (measured with software)
Now that jv16 PowerTools has been released, this is a practice that we wish to expand over a follow-up, extended study that we will later do in January of 2017.
All of the test results are accompanied by screenshots and video recordings which have been saved during various testing phases in order to address some of the most common test result credibility concerns for this type of studies.
This case study is open to discussion and public review on our forum, in this thread.
We would strongly appreciate if you could join the discussion, let us know what you think about the test results and recommend us new test ideas that we should cover in future similar studies.