Solid State Drive Fragmantation?

[athiker]

The SSD has its own processor inside, and firmware.

The processor and firmware, run in the background, and rearrange
data to suit wear leveling and future performance.

The reference to "leave it overnight", applies to "4KB-sized torture tests".
The natural block size of SSDs is rather large. And they don't deal with small
files well. If you pummel the SSD with random small files like that (as Anandtech
does in some of its testing), the SSD drive needs significant time in the
background, to unravel the mess. The files are recorded instantly, the mess
is cleaned up later. The blocks would be moved around to consolidate space.
And it really does take all night.

If the drive is not allowed to do that, you could go from a drive that
has a 250MB/sec write rate, to dropping down to 100MB/sec. If you leave
the drive overnight, and powered up, the SSD processor and firmware
move those 4KB file fragments around. The drive then has the 250MB/sec
performance the next morning. Doing such movements, costs something
in terms of SSD lifetime.

More info here if you're curious. Or one of the many articles
on Anandtech probably explains it as well.

http://en.wikipedia.org/wiki/Write_amplification

The 4KB random file test, is a pathological one. If you're doing
activities like email or web surfing, the SSD likely isn't compromised
at all, and ten minutes of background work is enough to maintain it.
But if you use one of those special "torture test" programs, then
you'll need all night to restore full SSD write speed.

Other options for resetting an SSD, might include Secure Erase, which
is a feature of the ATA command set. That might be a quicker way to
straighten up an SSD that has received a torture test. But that
also erases the data.

There's plenty written about the care and feeding of SSDs out
there, but I don't have the time to read it all. It's worth
reading though, if you own one.

Paul

Paul,

Thanks for the links and explanation.

 

[Gene E. Bloch]

I just had a look at Task Mgr / Performance with nothing running except
background tasks. I then exited most of those except virus checker.
After a few seconds it settles down to 0-1% CPU usage, with the idle
task mainly at 99%. That's probable even more idle than the average
British navvy, and should do.

Funny analogy. But being on the wrong side of the ocean, I *did* have to
look up navvy, a word whose meaning I have forgotten (or never knew).
But my spell checker doesn't know it either - par for the course, I'm
afraid

But I was thinking about *disk* activity, not CPU activity. My idea was
that if the firmware and the OS were both writing to disk, surprising
things might happen.

 

[Robin Bignall]

Funny analogy. But being on the wrong side of the ocean, I *did* have to
look up navvy, a word whose meaning I have forgotten (or never knew).
But my spell checker doesn't know it either - par for the course, I'm
afraid

COD:
navvy
· n. (pl. navvies) Brit. dated, a labourer employed in the excavation
and construction of a road, railway, or canal.
– ORIGIN C19: abbrev. of navigator.

I was being rude about British manual labourers.

But I was thinking about *disk* activity, not CPU activity. My idea was
that if the firmware and the OS were both writing to disk, surprising
things might happen.

I guess the firmware sorts that out. I wonder how one knows if and when
the SSD has beavered away doing this good stuff.

 

[Gene E. Bloch]

COD:
navvy
¡P n. (pl. navvies) Brit. dated, a labourer employed in the excavation
and construction of a road, railway, or canal.
¡V ORIGIN C19: abbrev. of navigator.

I was being rude about British manual labourers.

Note that I said that I had to look it up - that implies (to me at least
) that I *did* look it up.

The definition I found was similar and brought it up to date a bit:
"Navvy, pl. navvies: Originally, a laborer on canals for internal
navigation; hence a laborer on other public works, as in building
railroads, embankments, etc." (Webster)

I would've said labourer, but my spell checker is again out of step with
the British way. Also the American way, quite frequently

I guess the firmware sorts that out. I wonder how one knows if and when
the SSD has beavered away doing this good stuff.

You're likely right - I was just being sort of paranoid about the whole
thing, but by proxy: I don't have a SSD (or an SSD).

 

[Dave-UK]

I guess the firmware sorts that out. I wonder how one knows if and when
the SSD has beavered away doing this good stuff.

I bought a Kingston 120 G/B SSD the other day to speed up a laptop.
It came in a plastic shrink-wrap with an 'installation' leaflet of
many languages. The installation instructions were mostly about
physically fixing the drive into a computer.
It just said the BIOS should recognize the drive automatically on switch on.
If not, then enter the BIOS and 'instruct the system to autodetect the new drive'.

There is no mention of any low level sector alignment, defrag settings or any
of the dangers often mentioned in newsgroup posts. If it were that important
I would have thought the manufacturer of the drive would mention it somewhere.

There is a data migration video on their website and it is basically a simple
cloning operation, no messing about with BIOS settings.
http://www.kingston.com/en/ssd/v

 

[Scott]

I bought a Kingston 120 G/B SSD the other day to speed up a laptop.
It came in a plastic shrink-wrap with an 'installation' leaflet of
many languages. The installation instructions were mostly about
physically fixing the drive into a computer.
It just said the BIOS should recognize the drive automatically on switch on.
If not, then enter the BIOS and 'instruct the system to autodetect the new drive'.

There is no mention of any low level sector alignment, defrag settings or any
of the dangers often mentioned in newsgroup posts. If it were that important
I would have thought the manufacturer of the drive would mention it somewhere.

There is a data migration video on their website and it is basically a simple
cloning operation, no messing about with BIOS settings.
http://www.kingston.com/en/ssd/v

Mine (Samsung) says you should switch off defragmentation and indexing
to reduce wear and because they are unnecessary given the way a SSD
works. It provides a program to do this (Samsung Magician).

It also recommends changing the BIOS to AHCI but my computer shop
advises this is marginal and may cause compatibility issues.

 

[Robin Bignall]

Mine (Samsung) says you should switch off defragmentation and indexing
to reduce wear and because they are unnecessary given the way a SSD
works. It provides a program to do this (Samsung Magician).

My Crucial just had physical fixing instructions. No mention of any
settings, defrag, indexing at all.

It also recommends changing the BIOS to AHCI but my computer shop
advises this is marginal and may cause compatibility issues.

Did I not read that AHCI is a prerequisite for enabling the Windows TRIM
command?

 

[Paul]

I bought a Kingston 120 G/B SSD the other day to speed up a laptop. It
came in a plastic shrink-wrap with an 'installation' leaflet of many
languages. The installation instructions were mostly about physically
fixing the drive into a computer. It just said the BIOS should recognize
the drive automatically on switch on. If not, then enter the BIOS and
'instruct the system to autodetect the new drive'.

There is no mention of any low level sector alignment, defrag settings
or any
of the dangers often mentioned in newsgroup posts. If it were that
important
I would have thought the manufacturer of the drive would mention it
somewhere.

There is a data migration video on their website and it is basically a
simple
cloning operation, no messing about with BIOS settings.
http://www.kingston.com/en/ssd/v

SSDs work, with no intervention.

And to some manufacturers, that's all that matters.

The fact they don't work *well*, is your problem.

Many users will never benchmark their purchase, and
discover it could be working better.

To give another example of this, I have a Seagate hard drive.
Performance was suboptimal under NTFS. I couldn't figure out why.
Just the other day, I reformatted the partition, and used 64K
as the allocation size, and performance popped right up.
So even hard drives, can need some tuning, if the hard drive
design (controller board) is deficient. I have no idea what's
going on there. Something related to their cache design perhaps
(a design which was changed to support 512e, and is
quite different than previous designs).

http://en.wikipedia.org/wiki/Advanced_Format#512e

So even so-called mature technology, it's still your
responsibility, to tune it. If that means visiting
"overclocker sites" for instructions on how to do it,
so be it.

My hard drive was working, out of the box, but not
working all that well. Transfer speed was lethargic.
And initially, I thought it was bad blocks.

Paul

 

[Dave-UK]

SSDs work, with no intervention.

And to some manufacturers, that's all that matters.

The fact they don't work *well*, is your problem.

Many users will never benchmark their purchase, and
discover it could be working better.

To give another example of this, I have a Seagate hard drive.
Performance was suboptimal under NTFS. I couldn't figure out why.
Just the other day, I reformatted the partition, and used 64K
as the allocation size, and performance popped right up.
So even hard drives, can need some tuning, if the hard drive
design (controller board) is deficient. I have no idea what's
going on there. Something related to their cache design perhaps
(a design which was changed to support 512e, and is
quite different than previous designs).

http://en.wikipedia.org/wiki/Advanced_Format#512e

So even so-called mature technology, it's still your
responsibility, to tune it. If that means visiting
"overclocker sites" for instructions on how to do it,
so be it.

My hard drive was working, out of the box, but not
working all that well. Transfer speed was lethargic.
And initially, I thought it was bad blocks.

Paul

It isn't a 'fact' that they don't work well.
They do work well, so I don't have a problem.

I would rather believe what an SSD manufacturer says about their
product than mess about benchmarking and trying to achieve an
'improvement' that for practical use in the real world is irrelevant.

Your Seagate problems are immaterial as this thread is about solid
state drives.

 

[charlie]

I bought a Kingston 120 G/B SSD the other day to speed up a laptop. It
came in a plastic shrink-wrap with an 'installation' leaflet of many
languages. The installation instructions were mostly about physically
fixing the drive into a computer. It just said the BIOS should recognize
the drive automatically on switch on. If not, then enter the BIOS and
'instruct the system to autodetect the new drive'.

There is no mention of any low level sector alignment, defrag settings
or any
of the dangers often mentioned in newsgroup posts. If it were that
important
I would have thought the manufacturer of the drive would mention it
somewhere.

There is a data migration video on their website and it is basically a
simple
cloning operation, no messing about with BIOS settings.
http://www.kingston.com/en/ssd/v

The Kingston 120GB SSD I bought was from Newegg, several months ago.
First intended use was for a laptop, running Vista. The 120GB was just a
shade small, so I eventually ended up using a 240GB SSD. This desktop,
an older MBD based Phenom IIx4, was the recipient.

Things worked well, except that an SSD firmware update was needed.
That update was sort of a pain, since the SATA chipset drivers had to be
changed from the normal AMD drivers for the chipset back to the MS
drivers, do the firmware update, and back again to the AMD drivers.
All five of the SSD's I bought in the last few months had a one back
firmware version out of the box. (2 120GB. 2 240GB, and one 512GB.)

I don't know if it's a gimmick to sell larger drives, but some of the
mfrs claim that the features of the SSDs work much better if the drive
has plenty of free space, much as a conventional HD and defrag.

 

[Dave-UK]

The Kingston 120GB SSD I bought was from Newegg, several months ago.
First intended use was for a laptop, running Vista. The 120GB was just a
shade small, so I eventually ended up using a 240GB SSD. This desktop,
an older MBD based Phenom IIx4, was the recipient.

Things worked well, except that an SSD firmware update was needed.

Who or what told you a firmware update was needed ?
Did the drive fail somehow ?

 

[charlie]

Who or what told you a firmware update was needed ?
Did the drive fail somehow ?

The update had to do with the Samsung SSD controller chip and, as I
remember, Trim. MS driver and Intel SSD chipset drivers supposedly work
for the updates, and the AMD drivers do not. The firmware update
utilities give a cannot update error with the AMD drivers installed.
(At least with Vista and Win 7, didn't try with Win 8)
The older (2007) Vista laptop has an Intel processor, with an AMD
support chipset, and the other desktops use AMD processors and chip sets.
After initial does it work testing, I usually check components such as
HDs and SSDs for firmware updates, and install them if it makes sense to
do so.

 

[Dave-UK]

The update had to do with the Samsung SSD controller chip and, as I
remember, Trim. MS driver and Intel SSD chipset drivers supposedly work
for the updates, and the AMD drivers do not. The firmware update
utilities give a cannot update error with the AMD drivers installed.
(At least with Vista and Win 7, didn't try with Win 8)
The older (2007) Vista laptop has an Intel processor, with an AMD
support chipset, and the other desktops use AMD processors and chip sets.
After initial does it work testing, I usually check components such as
HDs and SSDs for firmware updates, and install them if it makes sense to
do so.

I see, so if I've read your post correctly the drive was performing well
but you decided to seek out a firmware update from the manufacturer's site
just because you like to have the latest version number.
That's something that a normal user would have no need to do or would even
know about, because the drive was working OK.

 

[Paul]

I see, so if I've read your post correctly the drive was performing well
but you decided to seek out a firmware update from the manufacturer's site
just because you like to have the latest version number.
That's something that a normal user would have no need to do or would even
know about, because the drive was working OK.

Dave, some of these SSD firmware updates, are to avoid drives that brick themselves.
It's not all knob polishing. There have been some generations of SSD drives,
that were an unmitigated disaster with respect to firmware. Check back
through the articles on Anandtech, for some of the history.

Paul

 

[Yousuf Khan]

Did I not read that AHCI is a prerequisite for enabling the Windows TRIM
command?

TRIM is supported even under IDE mode, I've done the test. It's not
directly part of the ATA specs, it's just an extended command that's
part of SSD's.

Yousuf Khan

 

[Robin Bignall]

TRIM is supported even under IDE mode, I've done the test. It's not
directly part of the ATA specs, it's just an extended command that's
part of SSD's.

Thanks.

 

[charlie]

Dave, some of these SSD firmware updates, are to avoid drives that brick
themselves.
It's not all knob polishing. There have been some generations of SSD
drives,
that were an unmitigated disaster with respect to firmware. Check back
through the articles on Anandtech, for some of the history.

Paul

When I get involved with (to me) new hardware, I generally try to make
sure that I am aware of the various details. Then there is the concern
that's related to buying sale items. (Why were/are they on sale?)
SSD's were a new area to deal with, older windows ops systems were not
"optimized" for use with them, and so forth. As a result, I generally
look up what a mfr has to say about things like video cards, SSD's, HD's
and so on before I get into putting them into actual use.

Initially, the concern with SSD's was to establish a reliable process to
deal with transferring the contents of a larger HD boot drive to a
smaller SSD. Then, how to "optimize" for best speed of the SSD.

I will say that at least one of the SSD firmware updates supposedly
improved reliability and service life at the expense of a small decrease
in speed.

It was also an exercise in finding the differences between spec'd
performance numbers and real world numbers.

I'd rather tell a customer/user what can be reasonably expected, based
upon experience, rather that a mfr's ideal spec.

 

[Scott]

I see, so if I've read your post correctly the drive was performing well
but you decided to seek out a firmware update from the manufacturer's site
just because you like to have the latest version number.
That's something that a normal user would have no need to do or would even
know about, because the drive was working OK.

To be fair, I routinely carry out software updates without spending
time investigating what the update has to offer and what the benefits
might be. I just think it makes sense to have the latest version. I
expect a lot of people think the same.

 

[Dave-UK]

To be fair, I routinely carry out software updates without spending
time investigating what the update has to offer and what the benefits
might be. I just think it makes sense to have the latest version. I
expect a lot of people think the same.

I think software updates are different from a firmware update.
The SSD can't phone home or tell you it has an update available.
Yes, you or I might update firmware because we tinker around with
the hardware but if the SSD is working OK then the average user
will have no idea about any update that may be available.
Do you update your motherboard's BIOS just because there's one
available, because I don't unless I have a problem that the update will fix.

 

[Scott]

I think software updates are different from a firmware update.
The SSD can't phone home or tell you it has an update available.
Yes, you or I might update firmware because we tinker around with
the hardware but if the SSD is working OK then the average user
will have no idea about any update that may be available.
Do you update your motherboard's BIOS just because there's one
available, because I don't unless I have a problem that the update will fix.

With my SSD the Samsung Magician will readily indicate whether a
firmware upgrade is available and I would install it anyway on the
assumption that Samsung know what they are doing. But I would not
update the BIOS so I am not consistent!