SSD partition alignment revisited

Hi All,

I have a trim tool on my netbook. There are two SSD drives, 64GB xSD and an
internal 32GB PATA SuperTalent SSD. Trim of the 64GB drive gives no
complaints and reports, 16 MB partition offset. That is to say, 16MB at the
beginning of drive is free. Trim would complain about bad alignment on C:
drive, so I used Partition Magic LIVE CD to change the front end alignment.
Maybe I used the wrong tool.. Maybe I'm confused about start and end points
of partitions.

BTW, CrystalDiskInfo reports a 1 KB buffer size on 32GB SuperTalent SSD and
CrystalDiskMark reports now improved R/W performance.

1 KB buffer size seems low. Could this reflect formatting options selected?

Thanks.




--- news://freenews.netfront.net/ - complaints: ---

Norm X wrote:
Hi All,

I have a trim tool on my netbook. There are two SSD drives, 64GB xSD and an
internal 32GB PATA SuperTalent SSD. Trim of the 64GB drive gives no
complaints and reports, 16 MB partition offset. That is to say, 16MB at the
beginning of drive is free. Trim would complain about bad alignment on C:
drive, so I used Partition Magic LIVE CD to change the front end alignment.
Maybe I used the wrong tool.. Maybe I'm confused about start and end points
of partitions.

BTW, CrystalDiskInfo reports a 1 KB buffer size on 32GB SuperTalent SSD and
CrystalDiskMark reports now improved R/W performance.

1 KB buffer size seems low. Could this reflect formatting options selected?

Thanks.

PTEDIT32 could report both the beginning of a partition
and the length of a partition, in 512 byte sectors. This
allows users to verify whether all numbers in the display
are divisible by 63 or by some power_of_two number (megabyte
alignment). It gives me some idea of the alignment.

So that's how I would figure that out.

If it took the Parted tool a while to complete the move,
then it probably did rewrite the entire partition.

The buffer on the SuperTalent could be based on
the packet size of the interface, and how large
a packet it can accommodate. The thing would refer
to "Cache", if as a drive, it had a cache RAM. At
first, SSD makers were timid about putting cache
on SSDs and there was none. Now, modern SSDs have
gained a single DRAM chip of respectable size.
A disadvantage of doing that, is if the power
fails, consumer SSDs don't have a supercap inside,
and it's going to be "interesting" to get any
DRAM cache emptied before the power is
completely gone.

So just be happy the only hardware appears
to be a "buffer", meaning there is less volatile
data in danger of being lost.

It doesn't sound like you changed the cluster
size. The alignment affects the alignment of
clusters to flash pages. A power-of-two choice
means it is likely the clusters are on flash
page boundaries.

NTFS clusters range from 4KB to 64KB. The nominal
4KB selection sounds low, but it also happens to
match the size of pagefile allocations, and it
suits both EFS encryption and NTFS compression
(if you choose to turn those on for some reason).

Selecting 64KB clusters is preferred for data-only
partitions, but only as a function of file size. If
you have a 4TB drive, and it will be holding an
infinite number of 7GB DVD ISO files, then 64KB is
perfect for those. If you have an infinite number
of 2KB text files containing your poetry compositions,
64KB clusters is a poor choice for that. Because
the cluster remainder is 62KB and you're wasting the
drive at a high rate.

On this computer, all my data partitions are NTFS 64K.
And any modern OSes with NTFS C:, they would be 4KB.

Paul

On Sun, 23 Oct 2016 14:41:21 -0400, Paul
wrote:

Norm X wrote:
Hi All,

I have a trim tool on my netbook. There are two SSD drives, 64GB xSD and an
internal 32GB PATA SuperTalent SSD. Trim of the 64GB drive gives no
complaints and reports, 16 MB partition offset. That is to say, 16MB at the
beginning of drive is free. Trim would complain about bad alignment on C:
drive, so I used Partition Magic LIVE CD to change the front end alignment.
Maybe I used the wrong tool.. Maybe I'm confused about start and end points
of partitions.

BTW, CrystalDiskInfo reports a 1 KB buffer size on 32GB SuperTalent SSD and
CrystalDiskMark reports now improved R/W performance.

1 KB buffer size seems low. Could this reflect formatting options selected?

Thanks.

PTEDIT32 could report both the beginning of a partition
and the length of a partition, in 512 byte sectors. This
allows users to verify whether all numbers in the display
are divisible by 63 or by some power_of_two number (megabyte
alignment). It gives me some idea of the alignment.

I have tried in vain to find a valid URL for this program.

Any help on a valid currently wording link?

So that's how I would figure that out.

If it took the Parted tool a while to complete the move,
then it probably did rewrite the entire partition.

The buffer on the SuperTalent could be based on
the packet size of the interface, and how large
a packet it can accommodate. The thing would refer
to "Cache", if as a drive, it had a cache RAM. At
first, SSD makers were timid about putting cache
on SSDs and there was none. Now, modern SSDs have
gained a single DRAM chip of respectable size.
A disadvantage of doing that, is if the power
fails, consumer SSDs don't have a supercap inside,
and it's going to be "interesting" to get any
DRAM cache emptied before the power is
completely gone.

So just be happy the only hardware appears
to be a "buffer", meaning there is less volatile
data in danger of being lost.

It doesn't sound like you changed the cluster
size. The alignment affects the alignment of
clusters to flash pages. A power-of-two choice
means it is likely the clusters are on flash
page boundaries.

NTFS clusters range from 4KB to 64KB. The nominal
4KB selection sounds low, but it also happens to
match the size of pagefile allocations, and it
suits both EFS encryption and NTFS compression
(if you choose to turn those on for some reason).

Selecting 64KB clusters is preferred for data-only
partitions, but only as a function of file size. If
you have a 4TB drive, and it will be holding an
infinite number of 7GB DVD ISO files, then 64KB is
perfect for those. If you have an infinite number
of 2KB text files containing your poetry compositions,
64KB clusters is a poor choice for that. Because
the cluster remainder is 62KB and you're wasting the
drive at a high rate.

On this computer, all my data partitions are NTFS 64K.
And any modern OSes with NTFS C:, they would be 4KB.

Paul

wrote:
On Sun, 23 Oct 2016 14:41:21 -0400, Paul

PTEDIT32

I have tried in vain to find a valid URL for this program.

Any help on a valid currently wording link?

Originally, it was a utility in the PowerQuest
Partition Magic product.

Then Symantec bought PowerQuest.

And Symantec made PTEDIT32.exe available on
their FTP site.

The FTP contents were cleared out recently,
and it's gone.

And archive.org only archives web sites, not
FTP sites, so there are no secondary copies.

*******

OK, so let's get creative. This appears to be a
trial version of PM8, with PowerQuest branding
inside. You don't need to install this!

http://www.download3k.com/System-Uti...ion-Magic.html

enpm800retaildemo.zip 23,776,770 bytes

Using 7ZIP, have a look inside the ZIP archive.
Normally, with things like InstallShield cabs, you
can't get inside them with 7ZIP. For some reason
this one is open.

L:\enpm800retaildemo.zip\Setup\PMagic.cab\
PTEDIT32.EXE 503,808 bytes September 16, 2002, 2:24:48 AM

And these are the copies I have already. I don't
have the 2002 version, as I bought PM7 to get
the 2001 version. So this file is new to me.

379,392 bytes February 18, 1999, 3:01:00 AM
501,760 bytes August 10, 2001, 7:00:00 AM

The scan on the new one is clean. The file also has
a special status. "National Software Reference Library (NIST)"
Normally, that status was reserved for some big CDs that
were too big to upload to virustotal, so the entries
are manually inserted by the staff. I don't know what
that means for a crusty old file like this.

https://www.virustotal.com/en/file/b...d5f0/analysis/

I tried it, and it has the traditional appearance.

Note that, on newer OSes, if you just double click,
it will give "Error 5". You need to right-click
the EXE and "Run as Administrator" to use it on
something like Win10.

Good luck,

Paul

On Thu, 01 Dec 2016 01:04:23 -0500, Paul
wrote:

wrote:
On Sun, 23 Oct 2016 14:41:21 -0400, Paul

PTEDIT32

I have tried in vain to find a valid URL for this program.

Any help on a valid currently wording link?

Originally, it was a utility in the PowerQuest
Partition Magic product.

Then Symantec bought PowerQuest.

And Symantec made PTEDIT32.exe available on
their FTP site.

The FTP contents were cleared out recently,
and it's gone.

And archive.org only archives web sites, not
FTP sites, so there are no secondary copies.

*******

OK, so let's get creative. This appears to be a
trial version of PM8, with PowerQuest branding
inside. You don't need to install this!

http://www.download3k.com/System-Uti...ion-Magic.html

enpm800retaildemo.zip 23,776,770 bytes

Using 7ZIP, have a look inside the ZIP archive.
Normally, with things like InstallShield cabs, you
can't get inside them with 7ZIP. For some reason
this one is open.

L:\enpm800retaildemo.zip\Setup\PMagic.cab\
PTEDIT32.EXE 503,808 bytes September 16, 2002, 2:24:48 AM

And these are the copies I have already. I don't
have the 2002 version, as I bought PM7 to get
the 2001 version. So this file is new to me.

379,392 bytes February 18, 1999, 3:01:00 AM
501,760 bytes August 10, 2001, 7:00:00 AM

The scan on the new one is clean. The file also has
a special status. "National Software Reference Library (NIST)"
Normally, that status was reserved for some big CDs that
were too big to upload to virustotal, so the entries
are manually inserted by the staff. I don't know what
that means for a crusty old file like this.

https://www.virustotal.com/en/file/b...d5f0/analysis/

I tried it, and it has the traditional appearance.

Note that, on newer OSes, if you just double click,
it will give "Error 5". You need to right-click
the EXE and "Run as Administrator" to use it on
something like Win10.

Good luck,

Paul


Well I got it so thank you. Have no idea how long that URL will last
before Symantec yanks it. I'd grab it now and stow it in case one
needs it later.