What if I never, never, never ever clean the registry on x64, W7 Sp1?

I think it is not needed, but it may be fine if you can manually delete registry keys you know should not exist anymore.

 

It may be useless in some cases, but definitely not always.

For example: AFAIK, a typical Win7 x64 registry size is 100 MB.

If 10% of this size can be lost with a safe registry "defrag" (garbage collection and compaction), you get 10 MB of free RAM for cache or other things, and less swap.

It may or may not result in a perceived performance boost, depending on what is happening on the system and what resources are available.

But it won't hurt too (unless the tool performing the operation isn't safe and/or you deliberately/accidentally break its procedures and don't restore a backup).

Reference on the matter:

- http://www.quepublishing.com/articles/article.aspx?p=1351988&seqNum=1
- http://www.quepublishing.com/articles/article.aspx?p=1351988&seqNum=2

 

That is interesting. I have known for years that database compaction makes a huge difference, but never was interested to find out about the registry.

If we give a real world example, we look no further than microsofts own access databases. Imagine you had a dbase with lets say 10 tables, each table containing perhaps 10,000 records, and each record housing lets say 40 columns. Thats a good amount of data to start with.

Now, suppose you build a front end for that database with any number of program languages. You read the data and modify/update or create new records. In most instances, the database will not be negatively affected at all.

However, if you were to use access itself to make queries and otherwise "use" the data in those tables, the size of the database grows quite a lot, and the performance will degrade. This must be the "garbage" referred in that article snippet. Access offers a "compact" feature, as does the jet engine if you go that route, but the end result is a vastly smaller size of database and improved performance. This isn't an opinion, but a fact reproducible at will. I know, I do it all the stinking time.

So, it makes me wonder, just what sort of benefits some form of registry "compaction" might have. Whether you could really tell or not is another story, but it is quite interesting to learn more about the topic, knowing what I know anyway.

Sul.

EDIT: and btw, in access at least, I have seen the size grow by 10x when users don't have a proper front end for it but use access itself. Thats quite a difference IMO.

 

Mark Russinovich makes some interesting comments about the registry.

http://www.whatthetech.com/2007/11/25/do-i-need-a-registry-cleaner/

 

Interesting, but I interpret things a bit different from Bill Castner.

Russinovich's statement is about registry cleaners which aim at removing a high quantity of keys for the sake of reducing the size of the registry. This really isn't useful and can be dangerous.

On the other hand, registry "cleaners" which aim at really fixing issues involving the registry (not only at cleaning it of keys for the sake of reducing its size), can be useful in some circumstances.

Later on, Russinovich adds that "developing a safe and effective registry cleaner requires a huge amount of application-specific knowledge".

This is true, and that's why very few registry cleaners are generally trusted.

But one thing that should be noted is that the "false positives" of these generally trusted registry cleaners, nowadays, are frequently of zero impact (keys that get recreated automatically).

Which indicates that (generally trusted) registry cleaners development really improved during all these years since that article.

 

BTW, I would still refrain from recommending any registry cleaner (generally trusted or not) as a tool to troubleshoot Windows and software related problems on random machines - mainly because there are way better methods.

But I just don't see the generally trusted registry cleaners as "something to combat".

 

All registry cleaners essentially do the same thing. Sure, some find more errors than others, and use tactics such as including empty registry keys as errors in order to find more errors. But, in any case it is extremely rare for any registry cleaner to actually fix any problems.

No, it's not that hard to work out what are needed registry keys which should be left alone and therefore make a registry cleaner which is safe. This does not require any knowledge of specific applications. The problem is that almost all registry cleaners are badly written and give false positives.

I don't know where your idea of "zero impact" comes from. From my experience with registry cleaners, I can say that at least 99% of registry cleaners have problems with false positives. If some of the valid registry keys they find are deleted, then problems can arise. In some cases there will be no problems from deleting valid keys. For example if a valid Add/Remove entry is deleted, then the program will no longer be listed in Add/Remove programs, which is not a good thing, but also won't cause any other issues.

I disagree. Almost all new registry cleaners still have problems with false positives, and almost all updates of existing products, do not the fix the issues they have with false positives.

 

Just FYI, I edited my post to clarify some statements before you answered.

 

When I say "improved" I don't mean "perfect". "Slightly better than before" is still "improved" IMO.

 

Rare or not, it can occur. And some problems aren't necessarily visible. They may simply manifest as specific sluggish, lag at times (Windows trying to act on something or in someway that it shouldn't). Small wrong instructions that you don't notice, but, when accumulated, can make an impact (depending on the system, of course).

 

Interesting. I would like to read more about this. Did you ever write or try to write a safe registry cleaner?

 

No, I considered it years ago but never did get round to it. It seems pretty pointless to me now days, as it so extremely rare for a registry cleaner to fix any problems or performance issues. It's just a shame that most programmers who write registry cleaners don't have a clue. If you can write a registry cleaner, surely you should realise if it has problems with false positives, and not release it until they have been fixed!

 

True. IMO, software like these are much more useful (and safe) for general troubleshooting purposes nowadays:

Tweaking.com Windows Repair - http://www.tweaking.com/content/page/windows_repair_all_in_one.html

Microsoft Fix It - http://support.microsoft.com/fixit/

Although some people confuse the first one with registry cleaners, which is funny.

I'm seeing some logical errors in your conclusion, but I'm a bit tired to elaborate.

 

I've been using Auslogics Registry Cleaner on an old XP computer for 10 years and never had any problem related to Auslogics Registry Cleaner.

I had problems with
1.SpyBot Search & Destroy/Antivirus Programs finding and deleting False Positives/legit files,
2.with aggressive disk cleaners (my fault), but never with Auslogics Registry Cleaner.
When I got Windows 8 computers, I decided to listen to knowledgeable computer specialists and never use any registry cleaners.
After a numerous refreshing/resetting, and even clean-installations of Windows 8, I figured out that Auslogics Registry Cleaner was not the worst thing in my life.
Every morning, I'm running Auslogics Registry cleaner on Windows 8, Windows 7, Vista, and XP computers.
At this forum, advanced computer specialists know more than the average computer user, and at this forum, advanced users have some custom-made programs/utilities that Auslogics Registry Cleaner could erase; therefore, I would not recommend Registry Cleaner HERE.

I remember, when CCleaner used to always find one problem but could not fix it. I found out that it was an empty key for Avira Antivir - False Positive.

I completely disagree with some statements that "no reason to clean hundreds of registry errors, when you have hundreds of thousand of registry keys."
One key could easily slow down your computer or even create real headaches.

And the most important thing, on old XP computers, I can clearly see improvement.

To sum up.

I can only recommend to the average user
Auslogics Registry Cleaner,
CCleaner,
Wise Registry Cleaner (not the deep mode),
and possibly, Ashampoo WinOptimizer.

There are hundreds of other registry cleaners - I don't use them. They could be dangerous for sure.

 

Compacting the registry or removing keys WONT save any ram as parts of the registry are loaded on demand, the whole thing is NOT loaded into memory:

To quote myself from https://www.wilderssecurity.com/showpost.php?p=2140283&postcount=44:

 

Yes, much better than using registry cleaners, and I use Windows Repair and various Fix-It solutions all the time when fixing computers.

Even Windows 7 offers a lot of problem solving abilities as part of the OS, if you right click on the Action Center flag and go to Troubleshoot a problem, you can search for fixes to various problems.

 

Doesn't it save RAM if those "parts of the registry that need to be loaded on demand" are smaller than before?

Also, Russinovich himself wrote a tool that "defrags" the registry.

It's called PageDefrag: http://technet.microsoft.com/en-us/sysinternals/bb897426.aspx

Look at what Russinovich says while describing his tool (among other things):

"Paging and Registry file fragmentation can be one of the leading causes of performance degradation related to file fragmentation in a system."

 

No as the chunks that are loaded into memory are fixed size and are not directly mapped to the file, but queried like a database.

That is correct it is purely about file access performance. Relevance to today’s machines - I don't honest know (when pagedefrag was written, there was for XP machines).

Cheers, Nick

 

A registry cleaner is useful for me in 2 situations:

After a malware removal: when running Ccleaner registry tool, i can see lots of leftover registry keys related to adware/spyware/malware being removed.

When having problems uninstalling some application: by using RegSeeker for targeting and removing keys related to the software many times is possible to solve the problem.

 

It's really strange that knowledgeable IT professionals/experts sooner or later resort to Rusinovich as "the one and only all-Windows-OS guru" in the world.

I'm an average computer user with my own experience.

I have a question.
I cleaned my Windows 8 laptop with
a. Auslogics Registry Cleaner, (daily)
b. WinOptimizer, (daily)
c. System Cleaner 7.(started two days ago)

This morning, I decided to run CCleaner Registry Cleaner after a one month gap. CCleaner found 6 registry issues. All of them were related to Ashampoo Office 2010, that I uninstalled 3 weeks ago.
Why Auslogics, WinOptimizer, System Cleaner could not remove those empty entries?

On a different Windows 8 computer, I ran Auslogics, WinOptimizer, and CCleaner Registry twice. No more registry issues shown.
After that, I ran Wise Registry cleaner. Wise Registry Clean found 126 registry issues.



Do the registry cleaners use different algorithm?

 

Yes - every registry cleaner is different.

Of the registry cleaners you are using only Auslogics and CCleaner (if you disable the check for obsolete software) are the only one which will not delete needed registry keys. So it's best to stay well away from the others.

 

""Small wrong instructions that you don't notice, but, when accumulated, can make an impact""

This is very true, sometimes a big impact, I find these so called repair
programs cause far more problems than "most" Reg cleaners

The one I use is Vit Registry Fix and it has never giving me a problem
will it be the same for everyone? "don't Know"

But as I said just make sure you have a good image
and then try it for yourself, only you can be the judge

 

Yes, and that's why I said it is extremely rare and not never. I used registry cleaners hundreds of times, and often use them on computers with problems.
There are probably 5 times I can think of when a registry cleaner has fixed problems, or made the computer appear to run faster.

 

Somewhat O/T question

May I use a RegCleaner on SSDs?

 

yeah, of course.