PCBUILD Archives

Personal Computer Hardware discussion List

PCBUILD@LISTSERV.ICORS.ORG

Options: Use Forum View

Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Don Cooley <[log in to unmask]>
Reply To:
PCBUILD - Personal Computer Hardware discussion List <[log in to unmask]>
Date:
Sun, 28 Jun 1998 14:56:44 -0700
Content-Type:
text/plain
Parts/Attachments:
text/plain (78 lines)
Jim

I agree with you in principal.  However taking the same example if you have
a file that is 1K it will be deposited in a 4K cluster or 8L cluster, 16K
cluster or 32K cluster depending upon the formatting of the Fat.  If you
look at your hard drive you will find a lot of files that are small and
files that are just over the cluster size - which means it takes two
clusters.

I ran a check on a drive that had been formatted at 32K clusters and
calculated the waste space.  It calculated to about a 40% waste.  So if you
had a 4GB drive on a Fat 16 with 32K clusters, you in fact will only store
about 60% of the drive size or 2.4GB.

Another way that I noticed it is when I moved the 300MB of files off of an
old drive to the new drive with 32K clusters - the space used was something
like 500+MB.  You in fact lose space because you have empty space in the
clusters that is not usable.  The larger then cluster size - the more you
lose.

It does make a difference.

I would agree with you on speed differences with having to access one
cluster vs. four.  However I am not sure that it would be too noticeable to
the operator - until you started running scandisk and defrag!

Don


=============================================
   Don Cooley / San Jose, California / [log in to unmask]
  New Prostate-Help Forums / My Prostate Cancer story
 My family story & history / All at  http://prostate-help.com
     To discuss Prostate Cancer call me at 408 268-6400
=============================================

-----Original Message-----
From: PCBUILD - Personal Computer Hardware discussion List
[mailto:[log in to unmask]]On Behalf Of Jim Meagher
Sent: Saturday, June 27, 1998 7:21 PM
To: [log in to unmask]
Subject: Re: [PCBUILD] FAT32/WIN98-What the manual says.


> -----Original Message-----
> From: PCBUILD - Personal Computer Hardware discussion List
> [mailto:[log in to unmask]]On Behalf Of Don Cooley
>
The size of the cluster has nothing to do with the total amount of
wasted space.  It is actually the total number of files stored (and
the size of them) that will affect how much space is wasted.

If I have a 15K file (doesn't matter what KIND of file) and I store it
on a disk using 4K clusters, the file will use one complete cluster for
the first 4K of data, a second complete cluster for the next 4K of
data, a third cluster for the next 4K and a fourth cluster for the
remaining 3K of data.  The fourth cluster (as well as the first three)
is now reserved for that file only.  So I have wasted one kilobyte
of storage space in that last cluster. (4K total minus 3K used)

If I store the same file on a disk with 8K clusters, then it will
need two clusters to store the file and will have 1K of unused space
in the second cluster.  So again, I have the same amount of wasted space.

If I store the file on a disk with 16K clusters, then it will only
need one cluster which will also have 1K of unused space.

Naturally a different file size will have a different outcome, but one
of the interesting side points is that (in my example) the FAT for the
16K disk will need to keep track of the location of only one cluster for
the file.  But the FAT of the 4K disk would need to keep track of 4
clusters -- thereby using up four times as much of the FAT storage area
to keep track of that one file.  So the 4K disk __could__ end up
being less efficient than the 16K disk.

PS:  When you format an 8Gig drive, you will get 8Gigs of storage
space regardless of whether it has 8K, 16K, or 32K clusters.

ATOM RSS1 RSS2