Hello.

Do you use a 3rd party disk defragger (e.g., O&O Defrag Pro.) in W10 4
SSDs & HDDs?

Thank you for reading and hopefully answering soon. :)
--
Dang Moderna booster shot took me down again with the same symptoms, but longer! Horror for Halloween time! Worst COVID-19 shot ever so far. :( It's November and cold winter again! :O
Note: A fixed width font (Courier, Monospace, etc.) is required to see this signature correctly.
/\___/\ Ant(Dude) @ http://aqfl.net & http://antfarm.home.dhs.org.
/ /\ /\ \ Please nuke ANT if replying by e-mail.
| |o o| |
\ _ /
( )

On Wed, 03 Nov 2021 21:12:44 -0500, ant@zimage.comANT (Ant) wrote:

>Hello.
>
>Do you use a 3rd party disk defragger (e.g., O&O Defrag Pro.) in W10 4
>SSDs & HDDs?
>
You might also try asking Fire fox for a list of free defrag
programs - try asking for - free disk defragger

On 2021-11-03 9:12 p.m., Ant wrote:
> Hello.
>
> Do you use a 3rd party disk defragger (e.g., O&O Defrag Pro.) in W10 4
> SSDs & HDDs?
>
> Thank you for reading and hopefully answering soon. :)
>

No 3rd party program needed, Windows Optimize looks after SSD trim and
HdDD degrament all on its own, no need for anything else.

Rene

On Wed, 3 Nov 2021 22:23:08 -0500, Rene Lamontagne <rlamont@shaw.ca>
wrote:

>On 2021-11-03 9:12 p.m., Ant wrote:
>> Hello.
>>
>> Do you use a 3rd party disk defragger (e.g., O&O Defrag Pro.) in W10 4
>> SSDs & HDDs?
>>
>> Thank you for reading and hopefully answering soon. :)
>>
>
>No 3rd party program needed, Windows Optimize looks after SSD trim and
>HdDD degrament all on its own, no need for anything else.

+1

On 11/3/2021 10:12 PM, Ant wrote:
> Hello.
>
> Do you use a 3rd party disk defragger (e.g., O&O Defrag Pro.) in W10 4
> SSDs & HDDs?
>
> Thank you for reading and hopefully answering soon. :)
>

The built-in defragmenter does a relatively good job.

Originally, it had a policy of not defragmenting files
larger than 50MB, but I think today there may be
exceptions to the original policy. For example, the
20GB .vhd files on my C: partition, they never get
defragmented by the built-in tool. But there have been
a few other, smaller things, that it may have messed with.

What you do, is you run the built-in one first,
then if you don't like the cosmetic details, you
run your third-party tool to defragment the large files.

I use JKDefrag 336 for visualization. This colors the blocks
and prepares jkdefrag.log if you want to read about it.

jkdefrag -a 1 -d 2 c: # make the diagram, for visualization
jkdefrag -a 5 -d 2 c: # squash files to origin, makes a mess
jkdefrag -a 2 -d 2 c: # defragment, and mostly do large files
# if the built-in already ran before this

By not defragmenting the really large files, that reduces
the run time somewhat. You won't go to bed, get up in the
morning, and find the defragmenter is still running eight
hours after you started it. The built-in is better than that,
sometimes, much better. Like ten or fifteen minute runs.

The pass count is variable. It could do three passes. It could
do a dozen passes. The number of passes is obviously dependent on
some quality metric we can't see. It gives a level of fragmentation
later as a percentage, but the built-in one does not show a
block display with the colored blocks.

*******

SSDs only need defragmentation under some corner conditions.
Mostly, they don't need it. You would not run JKDefrag on
an SSD, as it doesn't know SSDs should be treated differently.
The built-in defragmenter knows the difference between
the application of TRIM and defragment. What the built-in has
had problems with several times though, is it doesn't always
successfully determine what is an SSD and what is an HDD.
If your third-party tool is particularly adept at that,
who knows, maybe that third party tool should be used exclusively.

The built-in defragmenter, will defragment an SSD if there is
some kind of Copy On Write performance problem. That means,
there must be some kind of metric it can use, to determine
that corner case has arisen.

I have tested fragmentation on an SSD, and if the level of
fragmentation is extreme enough, yes, it actually slows
the SSD down by a tiny bit :-) So it isn't precisely true
that an SSD is "impervious" to fragmentation. But to get that
high a level of fragmentation, that never happens in normal usage.
I used the Passmark Fragmenter to test that. That's a tool
that fragments a disk on purpose. And no, you *don't* run
the Passmark Fragmenter directly on the SSD. You set up a
RAMDisk and fragment that, then clone over to the SSD, then
measure the performance. That's to reduce wear on the SSD.

One thing that Windows 10 has done, is it switched to a
64K buffer per write handle, up from a 4K buffer. Thus, when
the Passmark Fragmenter runs today, on an up to date Win10,
the fragments cannot be smaller than 64K. If I installed
Win10 10240, and ran the Fragmenter, then I could get
fragments as small as 4KB each (one cluster). Presumably this
is specifically an optimization to reduce wear on SSDs.
SSDs would prefer to work with larger quanta.

That also has implications for things like CrystalDiskMark and
others which claim to be doing a "4K write test". Do they
still do a 4K random test, or does the Win10 optimization defeat
that ? Dunno.

There is always something they are fiddling with. For example,
a while back, HDTune benchmarks curves were in error by ten percent
or so on Win10. As a result, if I needed to characterize new hardware
I bought, I'd use some other OS (and not Linux, either).
Win2K, WinXP, Win7 are OK. And the results on those
three, might agree with one another. Linux has so much caching
and cheating, it's a waste of time to make a curve there. Even
if you use an "un-cached" feature, there is still caching
going on in Linux.

HTH,
Paul

On 11/3/21 10:12 PM, this is what Ant wrote:
> Hello.
>
> Do you use a 3rd party disk defragger (e.g., O&O Defrag Pro.) in W10 4
> SSDs & HDDs?
>
> Thank you for reading and hopefully answering soon. :)
>
I use My Defrag 4.3.1
https://web.archive.org/web/20150811001837/http://www.mydefrag.com/index.html
It's an oldie but I like the features and presentation. It frags in stages, boot files, folders, recent used, remainder. Defrags first
then compacts. There is a view option, defrag only option, or defrag & compact.
Note that the block display in not blocks but one pixel per unit on the screen. So it's a bit had to see detail, but you will see overall
detail as it works for a minute or two.

--
Linux Mint Cinnamon 20.2 64bit, Dell Inspiron 5570 laptop
Quad Core i7-8550U, 16G Memory, 512G SSD, 750G & 1TB HDDs

Monty <monty@home.invalid> wrote:

> ant@zimage.comANT (Ant) wrote:
>
>> Do you use a 3rd party disk defragger (e.g., O&O Defrag Pro.) in W10
>> 4 SSDs & HDDs?
>
> You might also try asking Fire fox for a list of free defrag
> programs - try asking for - free disk defragger

Firefox won't speak to you, but an online search engine will.

Ant <ant@zimage.comANT> wrote:

> Do you use a 3rd party disk defragger (e.g., O&O Defrag Pro.) in W10 4
> SSDs & HDDs?

If you use a 3rd party defragger, you should not use the one in Windows,
or any other defragger. Each has different algorithms on how they think
is the best layout for clusters in a partition. Using one will use a
layout different than another, so they compete, and keep making changes
to get the layout to match their algorithm. Competing defraggers will
defrag differently, so you end up with lots of disk I/O that has no real
benefit.

So, if you use any 3rd-party defragger, you had better disable the
scheduled events in Task Scheduler that runs Microsoft's defragger on
Windows startup, and at regular intervals.

Task Scheduler -> Task Scheduler Library -> Microsoft -> Windows ->
Defrag -> ScheduledDefrag. Disable it. I don't remember where the
scheduled event gets created if you run the GUI defrag tool, and select
to run it at intervals. For the boot-time defrag, in the registry go
to:

HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Dfrg\BootOptimizeFunction

and change the Enable data item from Y to N. Missing registry entries
mean using the default value for them. If missing, and you want to
change away from the default, you'll have to create the registry entry
to set it to a non-default value. Sorry, I don't remember the data type
of the "Enable" named data item.

After testing several, I found no significant difference between using
the Microsoft or 3rd party defragmenters. There may be some difference
depending on how often you access huge files. Some defraggers like to
put huge files on the inside slower tracks of the disk, and leave the
faster outside tracks for highly used files, like system files.
Rearranging the food on your plate in a different order doesn't alter
what food is available on the plate. Microsoft's defragger is less
aggressive than many 3rd party defraggers.

SSDs should normally not be defragged. There is a special case where it
is beneficial. If a file is highly fragmented, all those fragments must
be recorded in the file system. Having fragments means the heads have
to bounce around a lot to find them, and on an HDD that incurs lag. But
not on SSDs. Defrag on an SSD affords no speedup in file segment
access, but the defrag (that is not SSD-aware) will pressure the SSD
with superfluous writes, and SSDs eventually die due to write volume.
However, it may be required to reduce the number of segments for a file
to under a count lower than the max number of segments the file system
can track. I don't remember the exact count, but the max segments for a
file in NTFS is something like a million. It would be very rare any
files gets that severely defragmented, but it can happen, and why defrag
on an SSD might be needed. If you're running a file server where
millions of users write to the same file, it could happen. On a
workstation or end-user box, highly unlikely a file will get that
severely dragged.

Big Al <Bears@invalid.com> wrote:

> Ant wrote:
>
>> Do you use a 3rd party disk defragger (e.g., O&O Defrag Pro.) in W10
>> 4 SSDs & HDDs?
>
> I use My Defrag 4.3.1
> https://web.archive.org/web/20150811001837/http://www.mydefrag.com/index.html
> It's an oldie but I like the features and presentation. It frags in stages, boot files, folders, recent used, remainder. Defrags first
> then compacts. There is a view option, defrag only option, or defrag & compact.
> Note that the block display in not blocks but one pixel per unit on the screen. So it's a bit had to see detail, but you will see overall
> detail as it works for a minute or two.

MyDefrag is a front-end to the Microsoft defragmenter.

"MyDefrag is extremely solid because it is based on the standard
defragmentation API by Microsoft"

On 11/4/2021 5:04 AM, VanguardLH wrote:
> Big Al <Bears@invalid.com> wrote:
>
>> Ant wrote:
>>
>>> Do you use a 3rd party disk defragger (e.g., O&O Defrag Pro.) in W10
>>> 4 SSDs & HDDs?
>>
>> I use My Defrag 4.3.1
>> https://web.archive.org/web/20150811001837/http://www.mydefrag.com/index.html
>> It's an oldie but I like the features and presentation. It frags in stages, boot files, folders, recent used, remainder. Defrags first
>> then compacts. There is a view option, defrag only option, or defrag & compact.
>> Note that the block display in not blocks but one pixel per unit on the screen. So it's a bit had to see detail, but you will see overall
>> detail as it works for a minute or two.
>
> MyDefrag is a front-end to the Microsoft defragmenter.
>
> "MyDefrag is extremely solid because it is based on the standard
> defragmentation API by Microsoft"
>

But that's not the essence of it.

What they're referring to, is the defragmenter API.

Originally, Microsoft did not know how to move files safely
at the cluster level.

One of the third party defragmenter products, demonstrated how
to do it. Microsoft "adopted" the method and added a library or
interface in the OS, doing the same thing.

The intention was to make a power-safe data movement method, with
atomic update. That means, if a "chunk" of data, of limited size,
is moved and the power goes off half-way, then because the change was
not "committed", the file system reverts to the previous state
and no data loss (of the "chunk") occurs.

The chunk size is limited, and this is a bit of a nuisance, but
it is what it is.

*******

OK, so like *every* commercial defragmenter as well as the Windows one,
they all use the built-in data movement API. The data movement API is
also "shadow compatible", which means it does not fuck up the
performance of shadows, by using the wrong size of quanta.

When MyDefrag claims it is safe to use, it is safe to use because
like everyone, it uses the API provided by Microsoft.

The value-added MyDefrag contributes, like all defragmenters, is "custom policy".

Every designer has clever ideas as to how best to optimize data position
on a disk. For Jerome, this involves moving the large files up high
on the partition, maybe with a bit of space between them, so if they're
appended in some later operation, they don't fragment.

Other companies may do "boot optimization". They watch the boot process,
the order of load of files. They move the files into ascending LBA order.
This means the head, during boot, follows a smooth if somewhat jerky
path. This might chop three seconds off boot time.

Doesn't matter.

What matters is, each commercial tool has its own set of policies,
and the users either "love the policies" or "hate the policies", and
that is how one product is purchased and another product rejected.

All the products are data-safe, due to the usage of the API built
into the OS. And that API would not be there, would not exist,
except for Microsoft adopting the work of a third party. As
it is the history of it. Today, I'm sure at least half the
staff could figure that one out, with their eyes closed, but
back in the day, the claim was they hadn't figured it out (yet).

It's the same with the "shrink partition" interface on Disk Management.
That too, involves the data movement API. But, guess what ? They
don't know how to move some of the metadata, metadata strategically
positioned at exactly 50% of the partition. One suspects, they *do*
know how to do it, however, they would be hurting the feelings
of the four person NTFS team, so... they don't touch it :-)
Whereas commercial tools, like the Paragon I was using minutes
ago, are not bothered in the least, to shrink a partition below
50%, which does involve moving that crappy set of blocks out of
the way.

Paul

Paul <nospam@needed.invalid> wrote:

> VanguardLH wrote:
>
>> Big Al <Bears@invalid.com> wrote:
>>
>>> I use My Defrag 4.3.1
>>>
>>> https://web.archive.org/web/20150811001837/http://www.mydefrag.com/index.html
>>>
>>> It's an oldie but I like the features and presentation. It frags in
>>> stages, boot files, folders, recent used, remainder. Defrags first
>>> then compacts. There is a view option, defrag only option, or
>>> defrag & compact. Note that the block display in not blocks but one
>>> pixel per unit on the screen. So it's a bit had to see detail, but
>>> you will see overall detail as it works for a minute or two.
>>
>> MyDefrag is a front-end to the Microsoft defragmenter.
>>
>> "MyDefrag is extremely solid because it is based on the standard
>> defragmentation API by Microsoft"
>
> But that's not the essence of it.
>
> What they're referring to, is the defragmenter API.
>
> Originally, Microsoft did not know how to move files safely
> at the cluster level.
>
> One of the third party defragmenter products, demonstrated how
> to do it. Microsoft "adopted" the method and added a library or
> interface in the OS, doing the same thing.

Windows is rife with products that Microsoft acquired or duplicated from
elsewhere. Not everyone wants to reinvent the wheel. Remember from the
very start that Bill Gates didn't write a whole new OS. He bought
Seattle's DOS (https://en.wikipedia.org/wiki/Seattle_Computer_Products).

> OK, so like *every* commercial defragmenter as well as the Windows one,
> they all use the built-in data movement API. The data movement API is
> also "shadow compatible", which means it does not fuck up the
> performance of shadows, by using the wrong size of quanta.
>
> When MyDefrag claims it is safe to use, it is safe to use because
> like everyone, it uses the API provided by Microsoft.

I'm not sure that is true. As with VSC (Volume Shadow Copy) provided by
Microsoft starting in Windows XP, not all backup programs use
Microsoft's shadows. I remember one backup product where you could
select using their scheme or Microsoft's. Possibly the backup software
was introduced before Microsoft established their defrag API, so the
backup software had to use something proprietary. At one time, DirectX
didn't yet exist, so there were proprietary graphics and audio libs.
After DirectX showed up, and after awhile to allow adoption, it was
easier to use DirectX. However, just because Microsoft provides an API
or service doesn't mean that's the only way to accomplish a task.

Was it Paragon that offered 2 choices of using Microsoft's VSC or their
own properietary method? I don't remember they ever well-explained why
I would use their method instead of Microsoft's.

> What matters is, each commercial tool has its own set of policies,
> and the users either "love the policies" or "hate the policies", and
> that is how one product is purchased and another product rejected.

And that's why I warned that only ONE defragmenter should be used to
avoid conflicting policies regarding layout. Defrag with one will
conflict with defrag by another, and you end up moving clusters that no
longer need to move. You'll always end up with one moving clusters
after you just finished using another that said it was done. Use only
one dragger to employ just one policy (their default or however you
configure it to operate).

Rene Lamontagne <rlamont@shaw.ca> wrote:
> On 2021-11-03 9:12 p.m., Ant wrote:
> > Hello.
> >
> > Do you use a 3rd party disk defragger (e.g., O&O Defrag Pro.) in W10 4
> > SSDs & HDDs?
> >
> > Thank you for reading and hopefully answering soon. :)
> >
>
> No 3rd party program needed, Windows Optimize looks after SSD trim and
> HdDD degrament all on its own, no need for anything else.

Assuming the Windows 10 version is about the same as the Windows 8.1
version, I agree that the built-in 'Optimise Drives' is sufficient.

I just checked the results of an 'Optimise' with the (Windows')
'Optimise Drives' tool, by doing a Analyze Drive' in (an old version
of) Piriform Defraggler.

That showed that there were only 52 [1] fragmented files and that
there were only two areas of the disk (read: C partition) used by files/
folders and the biggest of those two areas was neatly at the start of
the disk.

So not only did (Windows') 'Optimise Drives' do a good job of
defragging the disk, but also of 'compacting' it.

This 'compacting' result confirms what I have experienced from my
Differential Macrium Reflect image backups, which neatly grow over time
with actual in-use usage, while they would grow disproportionally if the
disk was badly/not compacted.

N.B. I do *not* use Piriform Defraggler to defragment the drive,
because as others have said, you should not mix disk defragmenters. I
only use Piriform Defraggler to analyze/view/<whatever>.

[1] It's unclear why (Windows') 'Optimise Drives' did not defragment
some of these files. While some of these files could be in-use/open,
others clearly are_not/cannot_be. Several are 1 to 7 years old. Anyway,
52 files is peanuts. So no worries mate.