Partition/Cluster Size/Wasted Space

I run Windows XP Home, SP2, NTFS.

I heard on a radio show (Online with David Lawrence) that when
partitioning drives, the larger the partition, the larger the clusters,
therefore the more wasted space. I may have misunderstood, but I did
google the subject, and what I found (and am able to understand - most
was very technical) leads me to believe this is correct.

My understanding is that regardless of how many partitions are made, the
same capacity will be reported. But, when data is added/stored on the
drive, a multi-partitioned drive will fill up slower because less space
is wasted. Also, the larger the file size, the less wasted space.

Is this true, and to what extent?

I've got the following IDE drives that I'll be adding to other machines,
which I plan to format as one partition. I've read about single
partition vs multi-partition, and the pros and cons of each. I find it
easier to navigate with fewer logical drives than half the alphabet
showing up on my computer.

250GB external USB Seagate (currently formatted FAT32 and hooked to this
NTFS machine and used as a backup )
160GB internal Seagate (to be used as a primary)
160BG internal WD (to be used as a slave)

For me, space isn't an issue, since my needs are generally low. I don't
do a lot of mult-media stuff (yet?). Also, storage is cheap, but I was
curious about this.

Thanks,
Bruce

Hi!

> I heard on a radio show (Online with David Lawrence) that when
> partitioning drives, the larger the partition, the larger the clusters,
> therefore the more wasted space.

That is generally correct, and no matter the file system type you use, there
will always be some waste. Most, if not all, file systems have one "smallest
unit of storage" that they recognize/define, and the size of this smallest
unit can change as the total disk or partition size goes up. As an example,
let's say that a file is 5 kilobytes in size, and the disk is formatted so
that 4 kilobytes is the smallest unit of storage available. This means that
the 5KB file is going to take up 8KB of disk space. Even though it is only
5KB in size, the clusters are laid out in 4KB...so the 5KB file must use two
of them and leave 3KB of space "unused".

Modern file systems do a lot to alleviate this issue, but there is still
some waste.

As an aside, but on a related note, it is possible with some file systems to
actually define the minimum cluster size. This is done for performance or
efficiency reasons. If you plan to store a lot of large files, using larger
clusters will reduce the overhead encountered each time a cluster must be
found, prepared for use and finally read or written to. With many small
files, however, this method would be very wasteful of disk space...so
smaller clusters are defined in this case. These are slower to access, but
offer more efficient usage of the storage that is available.

> My understanding is that regardless of how many partitions are made, the
> same capacity will be reported.

Yes, but only in a way. If you have several partitions, they will all be
smaller than the total size of the drive. After all, the partions must all
fit on the drive and cannot overlap one another. They can (but do not
necessarily have to) all add up to the total size of the drive. (It is
possible, though usually not done, to leave a portion of a drive totally
unpartitioned.)

> Is this true, and to what extent?

Yes, it generally is true. See the beginning paragraph for an example.

A few big files will waste less space than many small files because there
will not be a lot of partially filled clusters with the large files. There
will be a few, but they will appear at the end of the file.

> I find it
> easier to navigate with fewer logical drives than half the alphabet
> showing up on my computer.

Today the biggest advantage of having separate partitions is just to keep
things separated. Some people like to maintain a separate partition to store
their personal data. Their programs live on another partition. Doing things
this way can provide a little extra protection for data files in the event
that the operating system or a program fails and does some degree of damage.

> Also, storage is cheap, but I was curious about this.

Hope this helps to explain things a little better.

William

"William R. Walsh" < m>
wrote in news:zNEMg.134932$FQ1.53101@attbi_s71:

> Hi!
>
>> I heard on a radio show (Online with David Lawrence) that when
>> partitioning drives, the larger the partition, the larger the
>> clusters, therefore the more wasted space.
>
> That is generally correct, and no matter the file system type you use,
> there will always be some waste. Most, if not all, file systems have
> one "smallest unit of storage" that they recognize/define, and the
> size of this smallest unit can change as the total disk or partition
> size goes up. As an example, let's say that a file is 5 kilobytes in
> size, and the disk is formatted so that 4 kilobytes is the smallest
> unit of storage available. This means that the 5KB file is going to
> take up 8KB of disk space. Even though it is only 5KB in size, the
> clusters are laid out in 4KB...so the 5KB file must use two of them
> and leave 3KB of space "unused".
>
> Modern file systems do a lot to alleviate this issue, but there is
> still some waste.
>
> As an aside, but on a related note, it is possible with some file
> systems to actually define the minimum cluster size. This is done for
> performance or efficiency reasons. If you plan to store a lot of large
> files, using larger clusters will reduce the overhead encountered each
> time a cluster must be found, prepared for use and finally read or
> written to. With many small files, however, this method would be very
> wasteful of disk space...so smaller clusters are defined in this case.
> These are slower to access, but offer more efficient usage of the
> storage that is available.
>
>> My understanding is that regardless of how many partitions are made,
>> the same capacity will be reported.
>
> Yes, but only in a way. If you have several partitions, they will all
> be smaller than the total size of the drive. After all, the partions
> must all fit on the drive and cannot overlap one another. They can
> (but do not necessarily have to) all add up to the total size of the
> drive. (It is possible, though usually not done, to leave a portion of
> a drive totally unpartitioned.)
>
>> Is this true, and to what extent?
>
> Yes, it generally is true. See the beginning paragraph for an example.
>
> A few big files will waste less space than many small files because
> there will not be a lot of partially filled clusters with the large
> files. There will be a few, but they will appear at the end of the
> file.
>
>> I find it
>> easier to navigate with fewer logical drives than half the alphabet
>> showing up on my computer.
>
> Today the biggest advantage of having separate partitions is just to
> keep things separated. Some people like to maintain a separate
> partition to store their personal data. Their programs live on another
> partition. Doing things this way can provide a little extra protection
> for data files in the event that the operating system or a program
> fails and does some degree of damage.
>
>> Also, storage is cheap, but I was curious about this.
>
> Hope this helps to explain things a little better.
>
> William
>
>
>

Thanks very much, you've reinforced my intuition and filled in some gaps
for me. Very nice reply which I'll archive.

Bruce

Bruce wrote:
> I run Windows XP Home, SP2, NTFS.
>
> I heard on a radio show (Online with David Lawrence) that when
> partitioning drives, the larger the partition, the larger the clusters,
> therefore the more wasted space.

Yes, that's true. The term is called slack space. Files must be
allocated at least 1 cluster, and files cannot share clusters (It's
called cross-linking, and it's bad). So a small text file containing
"Hello World.", even though it's only 12 bytes long will still take up
512 bytes, Or however many bytes the cluster size happens to be.
Leaving 500 bytes of slack.

Don't worry overmuch about it, you have LOTS of room. Since most files
will take up multiple clusters anyway. Very few people bother with
cluster size. Radio show hosts should avoid scaring people into
reformatting their drives just to "recover" the couple megabytes total
a smaller cluster size would save.

For example. My "Program Files" Folder is 2,471,894,576 Bytes, but is
taking up 2,489,942,016 Bytes. Leaving 18,047,440 Bytes in "slack" This
works to be 17.21MB out of 2.31GB, or 0.73% slack.


I may have misunderstood, but I did
> google the subject, and what I found (and am able to understand - most
> was very technical) leads me to believe this is correct.
>
> My understanding is that regardless of how many partitions are made, the
> same capacity will be reported. But, when data is added/stored on the
> drive, a multi-partitioned drive will fill up slower because less space
> is wasted. Also, the larger the file size, the less wasted space.
>
> Is this true, and to what extent?

Larger average file size will take up more clusters. And therefor more
clusters will be 100% full. Reducing "slack".

The part about multi-partitioned hard drives filling slower is also
true, but.... it's a bit weak, the effect would not be enough to
notice. And, each partition would lose some space to the File
Allocation Table, or the NTFS equivlent. So the effect would be
partially cancelled out.

>
> I've got the following IDE drives that I'll be adding to other machines,
> which I plan to format as one partition. I've read about single
> partition vs multi-partition, and the pros and cons of each. I find it
> easier to navigate with fewer logical drives than half the alphabet
> showing up on my computer.
>
> 250GB external USB Seagate (currently formatted FAT32 and hooked to this
> NTFS machine and used as a backup )
> 160GB internal Seagate (to be used as a primary)
> 160BG internal WD (to be used as a slave)
>
> For me, space isn't an issue, since my needs are generally low. I don't
> do a lot of mult-media stuff (yet?). Also, storage is cheap, but I was
> curious about this.
>
> Thanks,
> Bruce

Bruce wrote:
> I run Windows XP Home, SP2, NTFS.
>
> I heard on a radio show (Online with David Lawrence) that when
> partitioning drives, the larger the partition, the larger the clusters,
> therefore the more wasted space. I may have misunderstood, but I did
> google the subject, and what I found (and am able to understand - most
> was very technical) leads me to believe this is correct.

What the radio guy described is true for the FAT2 file system, the
bigger the HDD/partition the bigger the clusters. It's forced by the
FAT2 design. NTFS is another animal entirely in that you can define your
300GB HDD using 4K clusters, resulting in much less slack space.

Thanks to whoever posted re: using larger clusters for HDDs with larger
files - I have 2 unpartitioned 300GB HDDs devoted to video and I think
I'll convert from 4K to 8K clusters, thinking that along with reduced
access time in general it might result in less overhead when capturing,
thus less chance of dropping a frame.

This is consistently a great NG.