andyc@hplsla.HP.COM (Andy Cassino) (04/22/89)
Fellow Atarians:
I recently got set up with a hard disk on my Mega ST2 and have
experimented with a couple of shareware programs, L-CACHE and
TURBODOS. I'm puzzled by some of the results!
First, neither of these programs affects results obtained by the ICD
RATE.HD program - I guess that this program operates at a low enough
level that caches and VDI-speedups are bypassed, so that's not too
puzzling. But it does mean I need another way to compare disk
performance.
So, my method of assessing the performance improvements due to these
programs is to move a group of 5 folders containing about 30 files of
medium size from one segment of the hard disc to another. This is
something I do fairly often and find to be irritatingly slow, even
with the hard drive (about 30 seconds).
Given this benchmark, TURBODOS yields a solid 30% improvement over
TOS. I would not expect L-CACHE to make any difference in this
experiment, but it did - L-CACHE slows things down by 50 to 150%,
depending on the size of the cache. (The larger the cache, slower
things go). This doesn't make sense to me!
Further, in experiments where I expect L-CACHE to make a difference,
it usually doesn't. No matter the size of cache that I define, only
small files seem to get cache'd, the larger files never do. I find
that most of my programs and data files are fairly large, and L-CACHE
just seems to be a drag on the system.
Both TURBODOS and L-CACHE conflict with another shareware program I
use occasionally - ST-TALK. If I try to run ST-TALK with either
program loaded, I get 3 to 12 bombs and have to reboot. This sort of
situation makes me pretty nervous, what with 65 MB of corruptable disk
lying around. I've tried TURBODOS with alot of other programs and
there doesn't seem be any other conflicts, my guess is that the
problems are with ST-TALK.
Comments appreciated!
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Andy Cassino %
% uucp: hplabs!hplsla!andyc domain: andyc%hplsla@hplabs.hp.com %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%