-----Original Message-----
>> he uses as example. It is a defective assumption.
>>
>I'm not convinced that the structure of the BIOS matters at all if one
uses
>the BDOS to fetch or point to whatever he's after. I do believe it's
fair
His utilities break on three of my systems as I do some things like
bank part of the bios and in one case part of the BIOS is resident
and executed by a seperate cpu.
>> >In those cases, among which one finds the CCS example, wherein a
simple,
>> >"dumb" BIOS is loaded into a 20-K CP/M requiring a 32K memory in
which
>> >to run, and subseqeuently used to run a "smarter" and more fully
>developed
>>
>> Not really.
>>
>What do you mean, Allison? I can assure you that it works fine as a
larger
I didn't mean you cant load a smarter bios. Only that in the case of the
CCS
if you want a nice system loading a smarter bios is a good thing to do.
>> Yes it is. I have floppies that use 1k (most SD though some are DD!),
>> many that use 2k and even some nut case using 4k(not me honest!).
>>
>Yes but that's easily recognizable if you check the directory entries
and
>disk parameters. Remember, I propose extracting this information from
the
>physical medium, not from some speculation.
Suit yourself. There are a few programs outthere that "image" the
disk and they didn't have it easy either.
>The stepper-driven Rodime 204E (1982) was about as fast as that Quantum
and
>had 640 cylinders instead of the Q540's 512. I used one for years and
am
I presume the rodime was larger than 31mb as well so more cylinders are
not
amazing. Was it an 8headed disk?
>be posted sometime soon, I hope) was 3 ms. It would do the job for
sure, in
>3 ms. I checked many of them. The Tandon equivalent and the Shugart
>equivalent both did the same. Others came later.
There is no dispute about that and I still have the spec sheat and all
for mine.
But to go from track 71 to track 0 still takes 213Ms, you can go faster
if
you ramp the step rate to 2.2mS and look for track 000.
>So, what did you do when you wanted to use 17 512-byte sectors (commonly
>used) and a 5-head drive, like the ones form Miniscribe that plagued us
now
>and again? Or, for that matter, the 6-head ST225, which was somewhat
later.
>How about 1K sectors? You could get 9 of them per track. People were
after
>capacity, even though they hadn't yet figured out how to waste it.
did that too sector(s) is a 16bit value though most bios will truncate it
on
assumption that it will be 256 or less. The only effect doing to the
miniscribe
( have two) was I had to divide out using non binary integer math rather
than
shifts, no biggie. it was much faster drive too.
>I'm not sure it helped much, but since the early ('506-class) HDD's
stepped
>at 3 ms regardless, and since the controllers didn't take advantage of
>momentum, I put the logical zero track of every partition in the
physical
>middle of the corresponding region of the drive. That made the
worst-case
>directory-seek half as long. Once drives capable of buffering step
commands
>became available that stunt wasn't necessary. Avoiding the use of
physical
>track zero was an important trick, however, since almost every drive
homed
>to that track on power-on, and if anything went wrong, it took a "rest"
>there. I had a lot less trouble with drives once I learned not to use
>physical track zero for anything that mattered at all.
Yep you can do that. You can take multi platter drive and designate
each platter or head as a drive. In some cases thats faster.
For a 5mb st506 partitions were sorta pointless as I was going for
something
bigger than the 1mb floppies I already had. A 2.5mb partition was not
considered or desired.
For the larger 31mb drives I went to 8mb partitions and let them fall on
cylinder
boundaries and left it at that. By then caching was in use and directory
caching
as well so where things were was not much of an issue.
>> Memorex 102 I gave away working, CTRL-C was a real wait and
>> a kick to watch the head creap back.
>>
>If it truly "crept" back, it was probably because it was being stepped
too
>fast and got lost, finally having to do a recal, which it did slowly.
No 512 tracks or so and a really slow step rate it crept. the controller
was doing step out, are we at 000, repeat. Morrow disk jocky and
controller with their code... very slow.
Allison
Received on Sat Oct 07 2000 - 10:30:06 BST
This archive was generated by hypermail 2.3.0
: Fri Oct 10 2014 - 23:33:15 BST