best defrag tool

[Frank]

You have your OS misconfigured

Just joking. Try PerfectDisk's 30 day free trial.
You'll be surprised.

I have no doubt in my mind that PerfectDisk is a good
product, but you know that if the OEM defragger takes
hours to complete with constant restarts, there is way
too many FUDwarz running in the background.

 

[Clamshell]

Either their disk in question was not badly fragmented to begin with, or
their test did not include video capture.

Anyone who has ever done video capture work understands the importance
of a defragged disk in minimizing dropped frames.

 

[Alexander Grigoriev]

A properly written video capture application can sustain even 12 MB/s
uncompressed stream, no matter how the disk is fragmented. By "properly
written" I mean, which employs separate threads for capture and disk write,
a memory buffer big enough, and FILE_FLAG_NO_BUFFERING mode when the file is
opened. I won't say I know any app that meets these conditions, but I haven
t tried many.

 

[Arno Wagner]

A properly written video capture application can sustain even 12 MB/s
uncompressed stream, no matter how the disk is fragmented. By "properly
written" I mean, which employs separate threads for capture and disk write,
a memory buffer big enough, and FILE_FLAG_NO_BUFFERING mode when the file is
opened. I won't say I know any app that meets these conditions, but I haven
t tried many.

This is definitely untrue. If the disk is heavily fragmented, you get
the full seek time + latency time for every disk cluster.

As an example assume 32kB clusters, radmon seek at 9ms, latency at 4ms
(7200rpm). These are realistic values. That gives you 13ms for every
32kB and equals a sustained read rate of about 2.5MB/sec.

A little bit might be gained by access reordering, but that is an OS
taks, the application cannot do this. Of course this is a pathological
example. But it clearly shows that your claim is wrong in general.

However in the average case with a decent filesystem layer (don't know
how well MS is doing, but on a long used Linux system I have about 2%
file fragmentation), you get very little fragmentation unless you fill
the drive to capacity or have unusual access patterns, such as
many sparse files that are updated within the sparse areas in small
blocks.

So yes, there is still room for defragging-tools. And some people
may even have desparate need for them. (Actually a backup-restore cycle
does the same...).

Arno

 

[wlloo]

This is definitely untrue.

Nope, yours is.

If the disk is heavily fragmented, you get the full
seek time + latency time for every disk cluster.

Bullshit. Its never that fragmented and you never
get a full seek for every cluster even if it was.

As an example assume 32kB clusters, radmon seek at 9ms,
latency at 4ms (7200rpm). These are realistic values.

But a full seek on every cluster aint.

That gives you 13ms for every 32kB

Mindlessly silly.

and equals a sustained read rate of about 2.5MB/sec.

Mindlessly silly.

A little bit might be gained by access reordering,
but that is an OS taks, the application cannot do
this. Of course this is a pathological example.

Yep, taint ever seen in real life.

But it clearly shows that your claim is wrong in general.

Crap, just shows yours is.

However in the average case with a decent filesystem layer
(don't know how well MS is doing, but on a long used Linux
system I have about 2% file fragmentation), you get very little
fragmentation unless you fill the drive to capacity or have
unusual access patterns, such as many sparse files that
are updated within the sparse areas in small blocks.

So your numbers are completely silly, stupid.

So yes, there is still room for defragging-tools.
Nope.

And some people may even have desparate need for them.
Crap.

(Actually a backup-restore cycle does the same...).

No need.

 

[lol]

Nope, yours is.


Bullshit. Its never that fragmented and you never
get a full seek for every cluster even if it was.


But a full seek on every cluster aint.


Mindlessly silly.


Mindlessly silly.


Yep, taint ever seen in real life.


Crap, just shows yours is.


So your numbers are completely silly, stupid.


No need.

This guy sounds an awful lot like Rod Speed doesn't he ?

Singularly obnoxious and that wonderful habit of never giving any
basis for his positions. One would say almost mindlessly silly.

Gad, is Rod's disease spreading or is he just avoiding our kill
filters ?

 

[Arno Wagner]

[...]


This guy sounds an awful lot like Rod Speed doesn't he ?

Singularly obnoxious and that wonderful habit of never giving any
basis for his positions. One would say almost mindlessly silly.

Gad, is Rod's disease spreading or is he just avoiding our kill
filters ?

Yes, I noticed that too. Another entry in my killfile...

Arno

 

[John Turco]

[...]

This guy sounds an awful lot like Rod Speed doesn't he ?

Singularly obnoxious and that wonderful habit of never giving any
basis for his positions. One would say almost mindlessly silly.

Gad, is Rod's disease spreading or is he just avoiding our kill
filters ?

Yes, I noticed that too. Another entry in my killfile...

Arno

Hello, Arno:

Don't you realize, that "wlloo <[email protected]>" IS Rod Speed?
He often uses other names/e-mail addresses, when replying to certain
posters (such as yourself), merely to avoid their killfiles. Otherwise,
they wouldn't even see his responses to them, in the first place.

Ain't that right, Rod? ;-)


Cordially,
John Turco <[email protected]>

 

[Rod Speed]

[...]

This guy sounds an awful lot like Rod Speed doesn't he ?

Singularly obnoxious and that wonderful habit of never giving any
basis for his positions. One would say almost mindlessly silly.

Gad, is Rod's disease spreading or is he just avoiding our kill
filters ?

Yes, I noticed that too. Another entry in my killfile...

Don't you realize, that "wlloo <[email protected]>" IS Rod Speed?
He often uses other names/e-mail addresses, when replying to certain
posters (such as yourself), merely to avoid their killfiles. Otherwise,
they wouldn't even see his responses to them, in the first place.

Ain't that right, Rod? ;-)

Yep. And its hilarious that clowns like that havent even noticed.

Just another terminal ****with that trys to close
its eyes to the exposure of its terminal stupiditys.

Rather pathetic, really.

 

[Arno Wagner]

[...]

This guy sounds an awful lot like Rod Speed doesn't he ?

Singularly obnoxious and that wonderful habit of never giving any
basis for his positions. One would say almost mindlessly silly.

Gad, is Rod's disease spreading or is he just avoiding our kill
filters ?

Yes, I noticed that too. Another entry in my killfile...

Arno

Hello, Arno:

Don't you realize, that "wlloo <[email protected]>" IS Rod Speed?
He often uses other names/e-mail addresses, when replying to certain
posters (such as yourself), merely to avoid their killfiles. Otherwise,
they wouldn't even see his responses to them, in the first place.

Hello John,

actually I don't think much about this entity. I am still not
convinced that he/it is a person and not a chatterbot. You have to
admit the setting is ideal for a long-term AI experiment and his
stereotypical answers sound like something straight out of a database.

If I would spend some effort tracking him, I probably would have known
right away. But he/it is not a menance, just a small annoyance, so I
mostly ignore him/it and from time to time add some new alias to the
killfile.

Arno