Hugh,

Consider that most current spec hard drives PATA and SATA tend to have 8, or
16Mb cache, and the ability to 'read-ahead'
I.E. read an entire track into cache.
The more data packed onto a track, the more benefit is gained from that
ability.


This means that, if a file occupies a contiguous portion of the disk, then
the 'track read' facility can read, and store - say 32Kb of data in 1
revolution of the drive, and will not need to reposition the heads and wait
for the allocation unit ( 4KB perhaps) to come round to the heads for the
next 7 read requests ( regardless of the rpm that's still only going to use
slightly over 1/8th of the disk access time that would be required to read a
fragmented version of the same file.
And - you also save on the time needed to search the FAT, or MFT for those
additional 7 allocation units, because with an unfragmented file, they all
(usually) are in the same block of allocation entries.

That situation will change slightly with SATA-II 'intelligent command
queuing, but only for a system doing enough intermingled access requests to
build a queue.

Re the paging file
Try comparing application performance with a contiguous, and then with a
fragmented pagefile, when the system is actively paging in & out - 2
applications with ongoing pagefaults as each want more 'real memory' than is
currently available.

However - you can also speed up a system by simply removing redundant file
entries - such as the thousands of zero length files that can be generated
by IE in the InternetTemporaryStorage.

As per my earlier waffle on planning and allocating appropriate groups of
files to particular partitions -
a well structured and setup system may not gain much benefit from defrag

JimB

----- Original Message ----- 
From: "Hugh Gundersen" <[EMAIL PROTECTED]>
To: <[email protected]>
Sent: Tuesday, February 28, 2006 5:45 PM
Subject: Re: SOFT: Diskeeper vs. XP defrag


On Tue, 28 Feb 2006 11:17:35 -0600, you wrote:
Apart from our off list chats I have read that in lab tests that defragging
a disk does very little
to improve the speed of an Hdd.  Defragging the swap file also does very
little in speed
improvements.

According to PC Pro and a couple of other tech mags - "it's nice to have
things in a neat pile but
if there is no logical reason then there is no real point"  or words to that
effect.

After all we cannot open the Hdd and see the neat rows of data all lined up
and if it doesn't
improve anything then why do it.

It might have done something in the long and distant past when computers ran
at tortoise top speed
and Hdds were worth a fortune for 4Mb and a Kings ransome for 25Mb.


Sir Hugh of Bognor
-- 

Remember. You may honestly believe that you understood everything
            you thought I said but what you thought you heard wasn't
           exactly what I said.

[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]

--
                ----------------------------------------
To Change your email Address for this list, send the following message:
 CHANGE  WIN-HOME  your_old_address  your_new_address
 to:  [EMAIL PROTECTED]
Note carefully that both old and new addresses are required.

--
                ----------------------------------------
To Change your email Address for this list, send the following message:
 CHANGE  WIN-HOME  your_old_address  your_new_address
 to:  [EMAIL PROTECTED]
Note carefully that both old and new addresses are required.

Reply via email to