Registry Cleaners

This is a question I am asked regularly – here is a valid test and it’s results. This is courtesy of – Link

Are system/Registry cleaners worthwhile?

Fred Langa

As Windows and third-party software have evolved over the years, Windows Secrets has periodically put various Registry- and system-cleanup products through their paces to examine the claims made by and about them.

For example, most cleanup software claims to streamline and shrink the Registry by removing obsolete, erroneous data and broken links stored there. Other products claim to do more, such as removing junk files, reducing boot times, and improving overall system performance.

Back in the days of Windows XP, third-party system cleaners were of real value in removing detritus from the operating system. But the current versions of Windows have many more built-in tuning and self-repair options. So do third-party cleaners still deliver any real-world benefit? Do they live up to their impressive claims?


Updating the WS cleaning-tools tests

Windows Secrets’ previous coverage of cleanup tools has laid an extensive foundation.

See, for example:

  • “The Windows Maintenance Challenge: Part 1″ – Aug. 14, 2014, Top Story
  • “The Windows Maintenance Challenge: Part 2″ – Aug. 28, 2014, Top Story
  • “Test-driving ‘free scan’ tune-up suites” – Aug. 9, 2012, Top Story
  • “Putting Registry- /system-cleanup apps to the test” – Nov. 10, 2011, Top Story

This article builds on that previous coverage; it was prompted by several factors.

Reader questions: A new release of Macecraft’s jv16 PowerTools X (free/paid; site) generated a flood of reader email asking about the product’s extraordinary claims.

For example, an ad mailing from a “Macecraft Software — PC optimization expert” carried the subject line “Make your computer at least 47 percent faster.”

I thought the “at least” might be a misprint, but the claim was repeated and expanded in the text of the advertisement, which stated:

“I personally ran jv16 PowerTools X in the Windows 7, Windows 8.1, and Windows 10 (Technical Preview) operating systems, and, on average, it made the PC system start 47 percent faster. This amazing tool also improved the system’s overall performance by 13 percent, as measured by the PCMark 7 benchmark software.”

Those claims prompted many Windows Secrets readers to write in, asking for some objective tests of the new version.

Major upgrades to an often-recommended tool: Heavy-duty cleaning tools can actually damage a PC with too-aggressive cleaning. They might remove data and Registry settings that appear to be unneeded but that actually serve a useful purpose.

Because of this, for several years now, I’ve been recommending — and using — safer, less-aggressive cleaning tools. Regular Windows Secrets readers will know that my primary choice has been CCleaner (free/paid; site). But the last version to undergo formal testing in Windows Secrets was Version 3.12; CCleaner is now at Version 5.07. It’s time for a fresh look.

Emergence of new cleaning tools: Over the years, Windows Secrets has also reported on a wide range of other cleaning tools, including Crossrider’s Reimage, Corel’s WinZip System Utilities Suite, Norton’s PC Checkup, AVG’s PC TuneUp, iolo’s System Checkup/System Mechanic, and PC Pitstop’s PC Matic.

But the popular — and heavily promoted — IObit Advanced SystemCare (free/paid; site) hasn’t been covered in Windows Secrets. So I have included it in this article.


Strengths and limitations of tests like these

It’s important to note that all reviews of PC-cleaning software share a major shortcoming. The results of performance tests principally apply only to the machine tested. Your system likely has a very different configuration and will produce significantly different results.

Consequently, published test results are best used only as a general guide; you can’t use them as the definitive word on what might work best with your specific combination of hardware, software, skills, and personal preferences.

To help you find the “best fit” utility for your unique circumstances, I’ll provide information on how you can adjust and adapt the data shown here to make it more representative of your setup.

But for the very best results possible, I recommend that you run similar tests on your own PC. For a safe test methodology, see the concepts and procedures described in the aforementioned Top Story, “The Windows Maintenance Challenge: Part 1″ and the follow-up, Part 2 Top Story.

Again, the results of my tests can serve as a useful, general guide. But only your own tests on your own PC will definitively show which tools are right for you — using both the utilities discussed below and any others you might run across.

(Warning: Before running a third-party cleaner on any PC used for live personal or work computing, always make sure you have a recent and full backup.)


Establishing the baseline setups and timings

To help ensure fairness and repeatability in my tests, I used two long-established, plain-vanilla, virtual PCs (VPCs) as baseline systems. One runs Win7 SP1 and the other runs Win8.1, upgraded from Win8.0. Both are real-life Windows setups I’ve used routinely for work for years.

To begin, I cloned their hard drives and setups to be certain I could restore the exact same VPC configurations for each test.

I then uninstalled all third-party software and performed thorough maintenance on both VPCs, using the techniques and tools documented in Windows Secrets articles such as the Jan. 16, 2014, Top Story, “Keep a healthy PC: A routine-maintenance guide,” and the Jan. 10, 2013, Top Story, “Let your PC start the new year right!”

If you’ve followed the recommendations in those and similar Windows Secrets articles, your PC will share at least some rough similarities with the test VPC setups.

Once the test VPCs were set up and running lean and clean, I took some baseline measurements of three of the main factors that Windows cleanup tools usually purport to improve: startup/shutdown times, disk clutter, and Registry bloat. Here’s how:

  • To accurately determine startup/shutdown times, I used the fully automated timing method described in the Aug. 28, 2014, Top Story. I performed each timing test a minimum of three times; I then averaged the results, rounding to the nearest whole second.
  • To determine disk bloat, I used File Explorer to measure the total amount of disk space in use, rounding to the nearest 0.1GB.
  • To gauge Registry bloat, I exported its contents and noted the size of the resulting .reg file, rounding to the nearest 1MB.

Figure 1 shows the baseline results.

Baseline results

Figure 1. Baseline measurements for startup/shutdown times, disk clutter, and Registry bloat for the two test VPCs

Note the quick startup/shutdown times. One way that my VPCs might differ from your setup is that they live on a speedy solid-state drive (SSD). One benefit of using an SSD is significantly faster startup and shutdown times. In a moment, I’ll provide information on how you can adapt my timing results to estimate performance on your system — though that’s a poor substitute for performing tests on your own system.


Adding software to create a used, “dirty” system

Next, to simulate a real-life system, I downloaded and installed 15 popular apps on both test VPCs. To select which apps to install, I used the then-current user-popularity ratings reported by

Where software conflicts were likely, I selected only one app for each category. For example, although AVG AntiVirus Free 2015 and Avast Free Antivirus 2015 were both extremely popular downloads, installing more than one full-time antivirus app on a given PC is known to cause system trouble — including excessive system slow-downs. Therefore, I selected just one AV app.

The selected apps include:

  • AVG AntiVirus Free 2015 (site)
  • IObit Malware Fighter Free (site)
  • IObit Smart Defrag (site)
  • betternet VPN (site)
  • Mask Surf Pro (site)
  • Google Chrome (site)
  • Mozilla Firefox (site)
  • Opera browser (site)
  • PrimoPDF (site)
  • CorelDRAW Graphics Suite Trial (site)
  • NexusFont (site)
  • Audacity (site)
  • Express Burn Disc Burning Software Free (site)
  • VLC Media Player (site)
  • Driver Booster 2 (site)

In each case, I allowed the apps to fully set themselves up the way the software publishers intended. For example, I accepted the default settings and options, and I didn’t meddle with the setup parameters. Both VPCs received exactly the same software selections, in the version appropriate to each VPC’s Windows version (7 or 8).

As expected, loading the system with this software changed all the baseline measures, as shown in Figure 2. In particular, the startup/shutdown times (relatively speaking) went through the roof — the Win7 boot times increased roughly sixfold!

Baseline dirty system

Figure 2. Adding a selection of software to the baseline VPCs produced representative dirty systems — with longer boot times, more disk clutter, and Registry bloat.

Note that the differences in startup/shutdown times between “clean” and “dirty” systems include the percentage change. This is intended to provide a rough comparison for what you might see with your specific system.

For example, if your PC normally takes, say, one minute for shutdown/startup, and one of the following tests shows a 100 percent increase in time, you can guesstimate that your PC might take around two minutes for its shutdown/startup under similar circumstances. (But again, you’ll get the most accurate results by performing your own tests.)


A Control Panel uninstall leaves behind debris

After installing and running the 15 apps, I then uninstalled everything via Control Panel’s standard uninstall applet. Because this is not an aggressive tool, it typically leaves behind at least some orphaned files, excess Registry settings, and so forth.

This is borne out by the test results shown in Figure 3. Win7 was able to return the startup/shutdown time to the previous condition (within the limits of the one-second timing sensitivity), but the disk and Registry gained some modest bloat. Win8 kept the Registry clean but allowed some disk bloat and a small increase in startup/shutdown time.

After standard uninstall

Figure 3. After uninstalling the third-party apps, neither Win7’s nor Win8’s uninstall applet was able to return all three measured attributes to their baseline states.

Taken alone, these measured differences are small — one round of software installs/uninstalls doesn’t result in huge changes. But each cycle of software setup and knockdown can leave behind digital debris. It adds up over time and can become a significant source of slowdowns and clutter.

That’s where cleanup tools can come in handy — at least in theory.

To see what common third-party cleanup tools could do, I cloned the final Win7/8 VPC setups — in the somewhat dirty, freshly uninstalled apps condition. I then set loose the different tools, each on its own identically cloned setup, to see how they would affect startup/shutdown times, disk clutter, and Registry bloat.


Test 1: jv16 PowerTools X

I installed and ran jv16 PowerTools X (free trial/$30; site) on the cloned Win7 and Win8 VPCs.

It reported alarmingly poor conditions on both setups. For example, on the Win7 VPC, it stated that “Overall System Health” was just 23 percent, “Registry Health” was only 1 percent, “Startup Health” was 54 percent, and “File System Health” was 87 percent.

Drilling into the details, it reported 263 Registry errors plus various other issues, as shown in Figure 4.

jv16 PowerTools report

Figure 4. Macecraft’s jv16 PowerTools X reported finding a substantial number of system problems.

I selected the one-button “Clean and fix my computer” option and let the software perform its self-assigned tasks.

When it was done, I again measured my virtual Win7 and Win8 systems. Figure 5 shows the results.

jv16 PowerTools X results

Figure 5. Instead of ‘at least 47 percent faster,’ jv16 PowerTools X made the test PCs’ startup/shutdown times at least 11 percent slower — with a bit of file and Registry bloat, too.

In fairness, some of the file and Registry bloat is due to the installation of jv16 PowerTools X itself, and that can’t be avoided. But I can think of no excuse for the disparity between the promised and delivered startup/shutdown times.

Your tests will, of course, yield somewhat different results. They might show significant system-performance improvements, but on my machines, jv16 PowerTools X didn’t come close to delivering on its extravagant promises.


Test 2: Advanced SystemCare

Next, I installed and ran Advanced SystemCare (free trial/variable cost thereafter; site) on freshly cloned copies of the Win7 and Win8 VPCs. In other words, it had exactly the same starting point as jv16 PowerTools X.

Advanced SystemCare reported many problems on both the Win7 and Win8 setups. For example, on the Win8 VPC, it reported 263 Registry errors and 58.6MB of junk files, plus 16 unspecified “performance issues,” four “security holes,” 100 “privacy issues,” 33 “Internet problems,” and a whopping 9,798 “browser security issues. ” See Figure 6.

Advanced SystemCare report

Figure 6. Advanced SystemCare claimed to find literally thousands of problems on the test PCs.

I let Advanced SystemCare run to completion (which included downloading a number of Windows Update items to the already fully updated test machines) and then took new measurements. Figure 7 shows the results.

Advanced SystemCare results

Figure 7. Advanced SystemCare (ASC) was ambitious in its repairs but yielded disappointing results on my test systems.

As you can see, Advanced SystemCare wasn’t useful for clearing the test systems’ Registry and file bloat nor for trimming startup/shutdown times. (However, as with jv16 PowerTools X, some of Advanced SystemCare’s file and Registry bloat is unavoidable — from its own added files.)

Again, your tests, on your system, might yield better results — or not.


Test 3: Piriform’s CCleaner

As with the previous two products, I installed Piriform’s CCleaner (free/$30 and up; site) on Win7 and Win8 VPCs, cloned from the post–Windows uninstall systems. Figure 8 shows the basic CCleaner interface.

CCleaner report

Figure 8. Piriform CCleaner is often recommended because it tends to be conservative on its system-fault findings.

Unlike jv16 PowerTools X and Advanced SystemCare, CCleaner reported finding a relatively modest number of deletable temp files and Registry entries. For example, it noted only 88 invalid Registry Keys on the Win8 setup.

As with the other tools, CCleaner’s own installation added some disk and Registry bloat, but not so much as the other products. Notably, only CCleaner returned the Win7 and Win8 VPC startup/shutdown times to their baseline levels. Figure 9 shows the results of CCleaner’s efforts.

CCleaner results

Figure 9. CCleaner was the only third-party cleanup tool to return startup/shutdown times to their baseline levels.


Summarizing the post-cleaning test results

For your convenience, Figure 10 shows the aggregate results, with the best and worst cleaning-tool results in each category highlighted in green and red, respectively.

Summary results

Figure 10. Aggregate results, with the best cleanup results shown in green and the worst in red.


Bottom-line conclusions and caveats

One of the key takeaways from these tests isn’t about the tested products themselves. Rather, it’s that Windows systems regularly maintained with the operating system’s built-in tools are already running relatively lean and clean. Based on the above tests, third-party tools probably won’t improve system performance by much — if at all.

It proves the value and validity of the common, free, do-it-yourself maintenance methods I used to set up and maintain those baseline systems.

Those maintenance techniques and tools are thoroughly documented in Windows Secrets articles such as the Jan. 16, 2014, Top Story, “Keep a healthy PC: A routine-maintenance guide,” and the Jan. 10, 2013, Top Story, “Let your PC start the new year right!”

If you follow the recommendations in those articles, then your PC will most likely also be lean, clean, and running well — without the assistance of any third-party cleaning tool.

A second related takeaway from these tests: If your PC is at all healthy, it’s unlikely you’ll see miraculous results from any commercial cleaning tool — and certainly no “47 percent faster boot times.”

My third and final takeaway: I’ll continue using — and recommending — lightweight tools such as CCleaner for routine cleanups. As mentioned earlier, this test (one cycle of software setup and knockdown on already-clean systems) didn’t leave a lot to clean up. But each cycle of software installs, upgrades, and uninstalls leaves behind some digital debris. Over time, it adds up and can become a significant source of slowdowns and clutter.

These tests tend to confirm that the manual cleaning methods mentioned above, plus lightweight cleaning from third-party tools such as CCleaner, might be all you need to keep a PC operating at or near best-available levels of performance.

I also worry about over-cleaning. I have to wonder about the high fault numbers reported by jv16 PowerTools X and Advanced SystemCare and what the two products are actually cleaning. (In some cases, the cure might be worse than the disease.) It’s difficult to know whether some of those “faults” are misdiagnoses, minor errors that have no real effect on system performance, or true issues — or possibly exaggerated numbers designed to make the product seem more valuable than it really is. I just don’t know. (Again, a recent full backup is your safety net for recovering from any damage done by overly aggressive cleaning.)

But those are my personal conclusions. As before, I recommend drawing your own conclusions, using the test concepts and procedures described in “The Windows Maintenance Challenge” Part 1 and Part 2.

That’s the true test for definitively answering which tools are right for your own unique mix of hardware, software, skill level, and personal preferences! Testing these products in virtual machines will provide a safe platform and let you make your own customized analysis.

Are system/Registry cleaners worthwhile?


If you need any help contact Robert at