Speed degradation over time

[Mark]

Can somebody explain why a computer would go slower over time? I assume it
to be one or more of the following:

too many [windows and internet ] temp files
spyware
viruses
installed applications (using memory)
insufficient disk space (reducing swap file size)

So my typical actions aim at resolving these issues, but the cleanup never
seems to be complete.

 

[Will Denny]

Hi

Please try the following article:

http://www3.telus.net/dandemar/slowcom.htm

--


Will Denny

MS MVP Shell/User
Please Reply to the News Groups

 

[Mark]

Thanks that article seems to just about confirm most of my ideas and adds a
few more.

I guess I'm basically on the right track.

I use Tune Up Utilities to clean up a lot of stuff mentioned there, if you
know it + defrag and check disk.

 

[Uncle Grumpy]

I use Tune Up Utilities to clean up a lot of stuff mentioned there, if you
know it + defrag and check disk.

Defrag and check disk will not make a noticeable speed difference in
the responsiveness of your XP system.

 

[Mark]

Well they do seem to. Why do you say not?

 

[Uncle Grumpy]

Well they do seem to. Why do you say not?

Because it's true.

 

[Mark]

Well it seems logical that if all files are contiguous that seek times will
be lower and that if files are uncorrupt that performance will be better

 

[Uncle Grumpy]

Well it seems logical that if all files are contiguous that seek times will
be lower and that if files are uncorrupt that performance will be better

but you won't notice it given the amount of RAM today's systems have
and the speed of today's processors.

It worked with Win98... it can only be detected in XP systems by
software that techies use to compare speeds.

 

[jm]

but you won't notice it given the amount of RAM today's systems have
and the speed of today's processors.

It worked with Win98... it can only be detected in XP systems by
software that techies use to compare speeds.

I agree. I have a suspicion that registry fragmentation may have
something to do with the problem. Even if you move all the data blocks
containing the registry into one contiguous block by defragmentation,
the various keys that are commonly accessed could be scattered about
within it. Some registry cleaners claim to be able to defragment the
registry, but I have never found a way to see if they really do or
not.

Sometimes I've even thought that Microsoft has an idle loop in the
Windows kernel that gets executed more times as the date gets
larger.

Julian Moss
www.tech-pro.net - tested quality Windows software

 

[Noncompliant]

Well they do seem to. Why do you say not?

I know when I restore an instance of XP from an image file, its usually
terribly fragmented. There usually seems to be an appreciable 30 second or
so delay in boot time from logon to completion.

I use a retail version of Diskeeper to defragment the XP partition. After
defragmenting, and doing a reboot, the apparent delay previously mentioned
is no longer present. Have done this procedure a few times. I cannot deny
its result.

Whether or not this affects XP and 3rd party application opening and working
time after XP is fully finished booting including user files, I've not
established.

 

[HeyBub]

I agree. I have a suspicion that registry fragmentation may have
something to do with the problem. Even if you move all the data blocks
containing the registry into one contiguous block by defragmentation,
the various keys that are commonly accessed could be scattered about
within it. Some registry cleaners claim to be able to defragment the
registry, but I have never found a way to see if they really do or
not.

You're assuming that Windows reads the registry sequentially. If that were
true, doubling the size of the registry would automatically double (on
average) the time taken to find something. But I don't think Windows does
read the registry sequentially. If I'm right, the time required to fetch a
registry entry is not related to the degree of "fragmentation."

 

[Mark]

I agree. I have a suspicion that registry fragmentation may have

You're assuming that Windows reads the registry sequentially. If that were
true, doubling the size of the registry would automatically double (on
average) the time taken to find something. But I don't think Windows does
read the registry sequentially. If I'm right, the time required to fetch a
registry entry is not related to the degree of "fragmentation."

This could be a long debate, since the way that windows reads the registry,
sequentially or not is not the issue, but a separate one. If a registry
entry or anything else is stored on the disk, it has to be found. The more
difficult to find due to the need to follow pointers, the longer it takes,
right?