[24hoursupport] Re: 1024 cylinder limit and Partition Magic

  • From: "Ron Allen" <chizotz@xxxxxxxxxxx>
  • To: 24hoursupport@xxxxxxxxxxxxx
  • Date: Sat, 22 Jun 2002 12:20:14 -0500

Hello Fuzzy,

>Here's the long and the short of it.  NTFS uses the same concept as FAT,
>so partition size does matter to some degree.  Note that if you converted
>to NTFS from FAT (or FAT32), you have a 512 byte cluster size.  That is a
>known limitation of the converter utility.

I actually used Partition Magic to convert to NTFS and partition everything
up instead of the Windows converter utility. But PM reports that I do
indeed have 512 bytes per physical sector, and 4K clusters with 8 sectors
per cluster.

>Having a partition on a different partition on the same disk has no effect
>on performance, negative or positive, as compared to having it in the same

The key to (an admittedly minimal in most cases) performance increase has
to do with the swap file being on a partition on a physically different
hard drive from the boot partition. That eliminates or at least greatly
reduces the seek time between reading and writing data to and from the page
file while also reading/writing from the program/data files which are
usually on C:\. This does not give any gain if the data and/or program
files are on a partition on the same physical drive as the swap file, of
course. It actually does speed things up a slight but notiable amount on
slower systems without much RAM. I first adopted this strategy back in the
early Windows 95 days, when I was running on a P120 with 32M RAM. It did
give me a slight performance boost. I quantified it at the time through
some completely unscientific tests and found that I gained about a second
or two on large operations. That was probably about the same as the error
factor, considering I was watching a sweep second hand on a watch and going
strictly by what I saw on the screen <g>. I also force a static swap file
rather than a dynamic one for the additional performance gain that gives,
also pretty slight in the overall scheme of things.

>PM can allow you to create partitions with a different cluster size, I
>believe, but it cannot change the cluster size for NTFS partitions, as far
>as I know (I haven't used the very latest version, though.)

That is true. Unlike using PM on a FAT/FAT32 drive, you can not change the
cluster size directly and have PM calculate the maximum partition size for
you on an NTFS partition. However, I presume that if you change the
partition size over a cluster size boundary that PM will also resize the
clusters. I don't see how it could be otherwise, but then again I don't
know NTFS very well.

Could you also address the original question for me, i.e. will I have a
problem resizing my NTFS boot partition so that it ends beyond the 1024
cylinder limit? Everything I have read so far indicates that this was a
problem on DOS and other older operating systems, but seems to have been
eliminated with newer systems. Yet I can not seem to find an explicit
statement, either on the Power Quest support site nor in the MS KB, that
says that it will work under XP and NTFS. The closest thing I saw was a KB
article that mentioned that the key to having this work is the BIOS and how
it handles INT 13 and the fact that the ATA2 standard is not really an
official standard yet (or wasn't when the article was written, I don't know
about today) and so it is implemented slightly differently by different
BIOS and drive manufacturers. This does not instill me with a lot of
confidence. I have nothing but Maxtor drives, and my BIOS is Award 6.00 SFG
(I think). The primary IDE controllers are whatever was built into  my
Future Power main board, and my add-on IDE controller is also Maxtor (but
is a Promise card licensed to Maxtor and uses Promise drivers). From what I
have read, this should be OK. I just would really, really like to avoid the
hassles of fixing it if it turns out sour :)



- Users can unsubscribe from this list by sending email to 
24hoursupport-request@xxxxxxxxxxxxx with 'unsubscribe' in the 
Subject field OR by logging into the Web interface at

Other related posts: