Pete wrote:
>
> (RD) ~10 secs
>
> Clarence--
> *How* did you get your 33MHz's RD to run this test in 6 secs??
How many buffers do you have ? <G> Ramdisk is VERY sensitive to this.
Buffers are only slighty faster than ramdisk due to slightly lower DOS
execution overhead. If you waste time looking thru too many buffers for
stuff that's not there - you lose. My optimum for RD is less but I use 15
as a compromise to help the HD.
Another factor could be the ramdisk. Some are better than others.
I used Vdisk and am going to try several others when I get a few minutes
- some year. <g>
And the DOS version. Newer and fatter is slower. That was PCDOS 3.30.
p.s. Note that my HD time was 38 sec. I don't think the default PC-Cache
is very good for us. My cache was 800k in extended memory.
I KNOW that it's very difficult to get EMM#*^ hehe to provide EXT for
PC-Cache. Piece of c#*^ ! QEMM does it just fine and gives me 3k more
low memory. But if I use QEMM on my test box, then I'm not running the
flavor of the year.:(
BTW. With the NEW (modified) version of speedchk I get 7 seconds consistently
on my Ramdisk and 36 sec first pass on HD plus 33 sec subsequent passes.
BTW2: I think there must be a problem with the WD caviar 325Mb HD on my P90.
Maybe someone set the wrong interleave before I got it. :(((
The '486 uses a Maxtor 171Mb with 1:1 interleave.
- Clarence Verge
--
- Help stamp out FATWARE. As a start visit: http://home.arachne.cz/
--
