On Sat, 02 Dec 2000 23:24:11 +0100 (CET), [EMAIL PROTECTED] (Richard Menedetter) wrote:
> ??? sector size _IS_ 512 bytes !!
> M$ does not use large sectors (who has told you so ??)
You did ... or so it appeared in the message I was responding to:
>>> TM> cluster size/allocation unit for FAT32?
>>> Ricsi> it could be 512 bytes ... but than MUCH RAM would be wasted
>>> Ricsi> ...
>>> so M& has chosen to use 4KB for sizes up to 8 GB and 8 KB for
>>> larger partitions ... and maybe even larger clusters for huge
>>> partitions
And, according to the software I have [called DOS] this
> LB> Would someone explain to me please how cluster/sector size on a HDD
> LB> has any effect on RAM usage?
> sector size not ... it's usually 512 bytes.
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ isn't true. I have my
RAM drive set with 512 sectors, but my HDD has sector size 2048 on the
primary drive, and temporarily 30K on the 2nd HDD.
> CLuster size yes ...
> you have to store the FAT in RAM ...
> with FAT16 there are at most 65535 clusters ....
How is this computed, and what determines cluster size?
> with FAT32 there are a LOT more ... these take up additional space.
Same questions.
> LB> Am I to assume that RAM "stores" instructions and data in active
> LB> memory by "sector" rather than by byte?
> usually many sectors are grouped together ... these are called clusters.
> LB> Thus 1Mb SDRAM can only store x number of sectors even if those
> LB> sectors don't add up to anywhere near 1 MegaBYTES of datapoints??
> ??? I don't see the point
You say that FAT32 takes up more RAM if "allocation unit" is smaller.
The only way that could be true would be if said "allocation unit"
were smaller than the smallest memory segment of a RAM chip... i.e.
the only way the same amount of data in FAT32 v FAT16 would take more
space in RAM for FAT32 would be if FAT32 somehow produced "slack" in
RAM.
> RAM stores bytes, and HDD as well ...
> on HDD there can be slack space ... because data is allocated in x bytes
> chunks.
According to what's been written, that's not the case. You clearly
stated that FAT32 would eat up more RAM ...
> LB> I don't see how that could be the case. If it were, then every
> LB> memory mapping piece of software ever released to the public would be
> LB> a tissue of lies ...
> ?????
No memory program I've seen has shown "slack" in RAM. i.e. if
you subtract the space used [for files & directories] from available
space, you get the free space shown. No slack means that it makes no
sense when you say smaller "allocation units" in FAT32 would result in
higher RAM usage.
> The discussion is about FAT32 ...
> there can be more clusters, so you can have larger HDDs, and usually
> SMALLER clusters ...
> But performance degrades ...
WHY?
Everything you said this time around supported my stance that RAM
stores by byte, yet you're saying cluster size determines RAM usage.
WHY? If I have 1Mb of data in my FAT16 table, or 1Mb of data in a
FAT32 table, how can cluster size affect amount of RAM used???
l.d.
[Gonna ask dumb questions until I get answers I can understand!]
-- Arachne V1.67, NON-COMMERCIAL copy, http://arachne.cz/
