On Wed, 20 Apr 2022, Tom Gardner via cctalk wrote:
Likewise, I don't know it for certain, but I am pretty sure that it is
true that virtually all controllers switch heads sequentially when
transferring blocks beyond the end of the track,
Are you implying that data/file that is more than one track long has its
next data on a track that is a different head of the same cylinder?
If that is, indeed what you are saying, . . .
It would make sense, and is common. Since it is obvious that
switching heads should take less time than stepping to the next
cylinder. BUT, it is a choice by the file system, not by the controller.
As a simple example, when floppy disks went from single sided to double
sided, SOME OS programmers chose to switch heads before stepping to the
next cylinder. (Cylinder 0 side A, cylinder 0 side B, Cylinder 1 side A,
cylinder 1 side B, etc.)
BUT, some chose to "keep what they had", and use the second side as an
"extension" of the first side, and chose to not switch heads
until all tracks on the first side were exhausted. (Cylinder 0 side A,
Cylinder 1 side A, Cylinder 2 side A . . . ) Of those, most
"recalibrated" (seek to zero) for the second side (cylinder 0 side A,
Cylinder 1 side A . . . Cylinder 75 side A, Cylinder 76 side A, Cylinde 0
side B, Cylinder 1 side B) (that's for 77 track 8") while others started
using the second side starting at the high end (to avoid the seek to zero
delay). (cylinder 0 side A, Cylinder 1 side A . . . Cylinder 75 side A,
Cylinder 76 side A, Cylinder 76 side B, cylinder 75 side B, . . .)
There were a few more variations, because it was the programmer making the
decision, not the controller, and we can come up with some amazing
cockamamy ideas.
--
Grumpy Ol' Fred ci...@xenosoft.com
