9
98 Guy
Gordon Freeman wrote:
> I don't think there's any way you could format a 250GB drive
> with less than 32KB clusters, FAT32 has a limit of how many
> clusters the FAT can contain,
Please have a look at this:
http://groups.google.com/group/micr...3eea1ce2fbf?lnk=st&q=&rnum=2#0e0e53eea1ce2fbf
(or try this link: http://tinyurl.com/2zy2jx)
It's a link to a google archived thread in this newsgroup with the
subject:
Cluster size and exploring the limits of FAT-32 Options
As I point out, I've already seen that win-98 can run on a large hard
drive (160 gb, partitioned as a single volume) and with 4kb cluster
size to boot.
FAT-32 does not have to increase cluster-size with increasing volume
size. That is a choice made by the win-9x and ME format tool, to keep
the total number of clusters under 4 million (and in most cases under
2 million).
I believe that strategy was designed so that scandisk could run on a
system with the specified minimum requirements that Microsoft spelled
out for windows 98, which was 16 mb of system memory. Given a system
with that amount of memory, scandisk could not load in a FAT table
containing more than 4 million clusters. Given a system with more
than 4 million clusters (in my case, I've tried up to 40 million) I've
seen scandisk handle a drive scan with no problems (ie - scandisk will
use all available system memory, not just the first 16 mb).
Others say that 4-million cluster limit was rooted in the idea that
Windows must load the entire FAT table as part of it's normal startup
and use, so having a large FAT table would consume inordinate amounts
of available system memory. I countered that argument by saying that
there is no evidence from looking at system memory usage that windows
loads the entire fat table during normal operation, and it really
doesn't have to. It only needs to load the fat entries for the files
that it opens.
> I don't think there's any way you could format a 250GB drive
> with less than 32KB clusters, FAT32 has a limit of how many
> clusters the FAT can contain,
Please have a look at this:
http://groups.google.com/group/micr...3eea1ce2fbf?lnk=st&q=&rnum=2#0e0e53eea1ce2fbf
(or try this link: http://tinyurl.com/2zy2jx)
It's a link to a google archived thread in this newsgroup with the
subject:
Cluster size and exploring the limits of FAT-32 Options
As I point out, I've already seen that win-98 can run on a large hard
drive (160 gb, partitioned as a single volume) and with 4kb cluster
size to boot.
FAT-32 does not have to increase cluster-size with increasing volume
size. That is a choice made by the win-9x and ME format tool, to keep
the total number of clusters under 4 million (and in most cases under
2 million).
I believe that strategy was designed so that scandisk could run on a
system with the specified minimum requirements that Microsoft spelled
out for windows 98, which was 16 mb of system memory. Given a system
with that amount of memory, scandisk could not load in a FAT table
containing more than 4 million clusters. Given a system with more
than 4 million clusters (in my case, I've tried up to 40 million) I've
seen scandisk handle a drive scan with no problems (ie - scandisk will
use all available system memory, not just the first 16 mb).
Others say that 4-million cluster limit was rooted in the idea that
Windows must load the entire FAT table as part of it's normal startup
and use, so having a large FAT table would consume inordinate amounts
of available system memory. I countered that argument by saying that
there is no evidence from looking at system memory usage that windows
loads the entire fat table during normal operation, and it really
doesn't have to. It only needs to load the fat entries for the files
that it opens.