Will formatting my hard drive in fat32 improve the performance?

I have an old Toshiba Pentium 4 laptop with 384M ram and 60G hard drive.

I can clearly see that the bottleneck for its performance is the 4200rpm of
its hard drive.

The hard drive is 60G in capacity and is still working fine, therefore it is
a pity to replace it with a new faster one.

I wonder now whether I should reformat it in fat32 foromat instead of the
NTFS format because somebody says that fat32 is faster than NTFS especially
for larger files.

It also looks obvious to me because each cluster of a fat32 is 8k (or 16k or
32k???) while for NTFS is 4k. Larger cluster means less disk seeks.

Is my idea reasonable enough?

But smaller clusters mean less wasted space.........and the drive will not
spin(seek)any faster than 4200rpm.....or tranfer any faster than your IDE
Channel allows.
http://en.wikipedia.org/wiki/AT_Attachment
So switching from NTFS to FAT32 will not give you any perceptual speed
advantage.
peter
"John" <rds1226@sh163.net> wrote in message
news:f97cqr$9ald$1@netnews.upenn.edu...
>I have an old Toshiba Pentium 4 laptop with 384M ram and 60G hard drive.
>
> I can clearly see that the bottleneck for its performance is the 4200rpm
> of its hard drive.
>
> The hard drive is 60G in capacity and is still working fine, therefore it
> is a pity to replace it with a new faster one.
>
> I wonder now whether I should reformat it in fat32 foromat instead of the
> NTFS format because somebody says that fat32 is faster than NTFS
> especially for larger files.
>
> It also looks obvious to me because each cluster of a fat32 is 8k (or 16k
> or 32k???) while for NTFS is 4k. Larger cluster means less disk seeks.
>
> Is my idea reasonable enough?
>

On Mon, 6 Aug 2007 10:56:26 -0400, "John"
<rds1226@sh163.net> wrote:

>I have an old Toshiba Pentium 4 laptop with 384M ram and 60G hard drive.
>
>I can clearly see that the bottleneck for its performance is the 4200rpm of
>its hard drive.
>

Agreed it is one area that effects many uses, but depending
on the tasks it might benefit from a memory upgrade as well.


>The hard drive is 60G in capacity and is still working fine, therefore it is
>a pity to replace it with a new faster one.

No way around that, if you want to get above 4200 RPM.


>
>I wonder now whether I should reformat it in fat32 foromat instead of the
>NTFS format because somebody says that fat32 is faster than NTFS especially
>for larger files.

The difference isn't enough to notice.


>
>It also looks obvious to me because each cluster of a fat32 is 8k (or 16k or
>32k???) while for NTFS is 4k. Larger cluster means less disk seeks.
>
>Is my idea reasonable enough?

The disk seeks are a matter of whether there is contiguous
space available to storage a file. If your hard drive is
not very fragmented it shouldn't be an issue. The most
significant performance loss from small clusters if if they
are much smaller like 512 bytes which might result from a
conversion from FAT32, at 4K you're not going to reap enough
of a benefit to matter.

You can use larger clusters with NTFS though, but AFAIK you
have to be running WinNT (XP? in your case) to format
specifying the cluster size. There are some 3rd party
utilities that allow specifying or changing NTFS cluster
size like Partition Magic or Paragon Partition Manager, but
I don't know of any free utilities to do it.

On 2007-08-06, John <rds1226@sh163.net> wrote:

> I have an old Toshiba Pentium 4 laptop with 384M ram and 60G hard drive.

> I wonder now whether I should reformat it in fat32 foromat instead of the
> NTFS format because somebody says that fat32 is faster than NTFS especially
> for larger files.
>
> It also looks obvious to me because each cluster of a fat32 is 8k (or 16k or
> 32k???) while for NTFS is 4k. Larger cluster means less disk seeks.

The short answer: unless you need some version of FAT for
interoperability reasons (accessing the drive with older versions
of Windows, or a non-Windows OS), go with NTFS.

A slightly more detailed answer, beginning with your concerns about
performance: Whilst you are correct that a larger block size means
that the worst case fragmentation will be better than a smaller
block size, in practice this worst case never occurs. Even in a
heavily fragmented filesystem, _most_ blocks will be contiguous
with their neighbours on disk.

Filesystem design has a big effect here, and while NTFS lacks some
of the techniques seen on some of the more advanced Unix filesystems,
the way metadata and file contents are separated means that it is
less prone to fragmentation that FAT. Thus I would expect to see
a real-world reduction in fragmentation with NTFS.

In addition, FAT has a number of limitations due principally to its
age and origins on small, MSDOS, floppy based machines. There are
no file permission or ownership details recorded: vital for security
on a multi user system, or even these days protecting the system
against malware.

More importantly, I believe you are approaching the issue from the
wrong angle if performance is your primary motivation. For a file
system, one thing matters and that is reliability - your disk must
not lose your data no matter what happens. That is of course an
impossible dream, but NTFS uses several tricks that collectively
make it substantially more robust than FAT.

--
Andrew Smallshaw
andrews@sdf.lonestar.org

In article <f97cqr$9ald$1@netnews.upenn.edu>, rds1226
@sh163.net says...
> I have an old Toshiba Pentium 4 laptop with 384M ram and 60G hard drive.
>
> I can clearly see that the bottleneck for its performance is the 4200rpm of
> its hard drive.
>
> The hard drive is 60G in capacity and is still working fine, therefore it is
> a pity to replace it with a new faster one.
>
> I wonder now whether I should reformat it in fat32 foromat instead of the
> NTFS format because somebody says that fat32 is faster than NTFS especially
> for larger files.
>
> It also looks obvious to me because each cluster of a fat32 is 8k (or 16k or
> 32k???) while for NTFS is 4k. Larger cluster means less disk seeks.
>
> Is my idea reasonable enough?
>
>

No, I doubt the increase will be enough to even notice.

The best thing you can do to increase performance is to
up your ram to 512MB or 1024MB.

Bill
--
Gmail and Google Groups. This century's answer to AOL and
WebTV.

"John" <rds1226@sh163.net> wrote in message
news:f97cqr$9ald$1@netnews.upenn.edu...
> I have an old Toshiba Pentium 4 laptop with 384M ram and 60G hard drive.
>
> I can clearly see that the bottleneck for its performance is the 4200rpm
of
> its hard drive.
>
> The hard drive is 60G in capacity and is still working fine, therefore it
is
> a pity to replace it with a new faster one.
>
> I wonder now whether I should reformat it in fat32 foromat instead of the
> NTFS format because somebody says that fat32 is faster than NTFS
especially
> for larger files.
>
> It also looks obvious to me because each cluster of a fat32 is 8k (or 16k
or
> 32k???) while for NTFS is 4k. Larger cluster means less disk seeks.
>
> Is my idea reasonable enough?
>
>

I've done a lot of experimenting and in general cannot notice any difference
in performance between FAT32 and NTFS.

I once had an ancient 4 gig drive that did work a lot better as fat32...but
it was really an old drive.

That was really the one exception.



--
Posted via a free Usenet account from http://www.teranews.com

"John" <rds1226@sh163.net> wrote...
> I can clearly see that the bottleneck for its performance is the 4200rpm
> of its hard drive.
>
> The hard drive is 60G in capacity and is still working fine, therefore it
> is a pity to replace it with a new faster one.
>
> I wonder now whether I should reformat it in fat32 foromat instead of the
> NTFS format because somebody says that fat32 is faster than NTFS
> especially for larger files.

Either live with the slow performance or buy a faster HD.

When the 4200 RPM HD died in my HP laptop (which my wife wouldn't use
because it was too slow), I bought a 5600 RPM HD to replace it. The
difference was AMAZING! My wife was even happy!

In message <46b7e8d1@kcnews03> "John Weiss"
<jrweiss@nospamattglobal.net> wrote:

>"John" <rds1226@sh163.net> wrote...
>> I can clearly see that the bottleneck for its performance is the 4200rpm
>> of its hard drive.
>>
>> The hard drive is 60G in capacity and is still working fine, therefore it
>> is a pity to replace it with a new faster one.
>>
>> I wonder now whether I should reformat it in fat32 foromat instead of the
>> NTFS format because somebody says that fat32 is faster than NTFS
>> especially for larger files.
>
>Either live with the slow performance or buy a faster HD.
>
>When the 4200 RPM HD died in my HP laptop (which my wife wouldn't use
>because it was too slow), I bought a 5600 RPM HD to replace it. The
>difference was AMAZING! My wife was even happy!

<pedantic>5400rpm</pedantic>

Try a 7200rpm some time -- Chews more power, runs hotter, but much
faster

--
Americans couldn't be any more self-absorbed if they were made from equal
parts water and papertowel.
-- Dennis Miller

On Mon, 06 Aug 2007 22:22:49 -0600, DevilsPGD
<spam_narf_spam@crazyhat.net> wrote:

>In message <46b7e8d1@kcnews03> "John Weiss"
><jrweiss@nospamattglobal.net> wrote:
>
>>"John" <rds1226@sh163.net> wrote...
>>> I can clearly see that the bottleneck for its performance is the 4200rpm
>>> of its hard drive.
>>>
>>> The hard drive is 60G in capacity and is still working fine, therefore it
>>> is a pity to replace it with a new faster one.
>>>
>>> I wonder now whether I should reformat it in fat32 foromat instead of the
>>> NTFS format because somebody says that fat32 is faster than NTFS
>>> especially for larger files.
>>
>>Either live with the slow performance or buy a faster HD.
>>
>>When the 4200 RPM HD died in my HP laptop (which my wife wouldn't use
>>because it was too slow), I bought a 5600 RPM HD to replace it. The
>>difference was AMAZING! My wife was even happy!
>
><pedantic>5400rpm</pedantic>
>
>Try a 7200rpm some time -- Chews more power, runs hotter, but much
>faster

.... or SSD of course, if one only procrastinates long enough
they'll be down to a reasonable price-point.

"DevilsPGD" <spam_narf_spam@crazyhat.net> wrote...
>>
>>When the 4200 RPM HD died in my HP laptop (which my wife wouldn't use
>>because it was too slow), I bought a 5600 RPM HD to replace it. The
>>difference was AMAZING! My wife was even happy!
>
> <pedantic>5400rpm</pedantic>
>
> Try a 7200rpm some time -- Chews more power, runs hotter, but much
> faster

It is a viable option for a newer laptop, and I have one in my IBM.
However, the price premium may make it undesirable for an older computer.