98 Guy wrote:

Paul wrote:

Is FAT32 "efficient" on a 2TB partition ? Not really.

Actually, it probably is.

Micro$haft designed their FAT-32 format programs to scale up the
cluster-size along with hard-drive size in order to keep the total
number of allocation units to 2 million or less.

Format.com has a /Z command-line switch that allows you to specify an
alternate cluster size - but it doesn't work.

Having large clusters (either 32 or 64 kb) is not necessarily wasteful
on a volume where you primarily store large files (media files, for
example).

I've used hard-drive formatting tools supplied by drive makers (like
Seagate, WD, etc) to format FAT32 volumes using custom cluster-sizes in
order to get around the intentional custer-size strategy that Micro$oft
designed into format.com.

For example, I've installed win-98se on a 500 gb sata hard drive that
was formatted as a single volume with 4kb cluster size (same as any
NT-based OS would do). This resulted in about 125 million allocation
units (far beyond what Macro$haft claimed was possible for either DOS or
Win-98 to handle).


I would expect the size of the FAT is a bit of an issue.

http://en.wikipedia.org/wiki/Design_...cation_Tabl e

"Because each FAT32 entry occupies 32 bits (4 bytes) the
maximal number of clusters (268435444) requires..."

1073741776 bytes or a gigabyte of RAM to hold the whole FAT.

How many of those could you have sloshing around, without
needing to re-read the FAT ? That's got to have some impact.

Win98 has some funny address space limitations I don't
understand, so a FAT that big might even cause problems
with the dimensions of some of the addressing.

I've not tested this, mainly because I wouldn't
recommend it as a configuration to anyone. It would
be fun to test, but I don't have a spare disk that
size which is completely empty.

Paul