NTFS partition increment size?

[Ham Pastrami]

When using disk utilities to change partition sizes I notice that the
possible sizes are stepped in increments which are larger than the size of a
cluster. For example, 8GiB is exactly 2,097,152 * 4KiB clusters. However,
disk utilities seem unable to set a partition size to exactly 8GiB, the
nearest approximations are 7.999 GiB and 8.005 GiB. How do you know or
calculate what the possible disk sizes/increments are?

 

[Andrew Rossmann]

When using disk utilities to change partition sizes I notice that the
possible sizes are stepped in increments which are larger than the size of a
cluster. For example, 8GiB is exactly 2,097,152 * 4KiB clusters. However,
disk utilities seem unable to set a partition size to exactly 8GiB, the
nearest approximations are 7.999 GiB and 8.005 GiB. How do you know or
calculate what the possible disk sizes/increments are?

I think partition information still uses a cylinders/heads format, even
though most drives just use sector counts these days. The sector counts
are converted into a pseudo cylinder/head/sector scheme.

Partitions are cylinder based. Most schemes use 63 sectors, but the
heads can be 16 to 255. 255 heads * 63 sectors * 512 bytes = 8,225,280 =
7.84M.

It's been a while since I dealt with this, so I don't know how out of
date it is. I think the basics haven't changed, though.

A few references:
http://en.wikipedia.org/wiki/Master_boot_record
http://en.wikipedia.org/wiki/GUID_Partition_Table
http://www.partitionsupport.com/partitionnotes.htm

 

[Ham Pastrami]

I think partition information still uses a cylinders/heads format, even
though most drives just use sector counts these days. The sector counts
are converted into a pseudo cylinder/head/sector scheme.

You are correct. Using heads*sectors*512 as a boundary, the math is
consistent with the figures I stated previously. Thanks for the help.