lane@cs.dal.ca (John Wright/Dr. Pat Lane) (04/10/90)
Some questions about 1:1 controllers. I understand about hard disk
interleaving and how the optimal interleave factor depends on the
disk drive electronics, controller, and the machine they are operating
in. The preponderance of controllers that advertise being "1:1 capable"
seems to suggest that it's only the controller that determines the optimal
interleave and that a "1:1 controller" will achieve 1:1 optimality in
any machine with any drive. Is that so?
On the subject of which components affect optimum interleave factor,
I would suppose that drives are standardized enough that there tends
not to be a significant difference with respect to interleaving. On
the machine side, CPU speed seems not to make a difference: I would
suppose that this is because hard disk I/O is handled through DMA and
that bus speed and the DMA electronics were the big factors here. Am
I correct on these points?
I've always assumed that 1:1 interleaving was possible (both to format
a drive and to read it) on any drive/controller/machine combination; just
that if any of them were too slow, 1:1 would be a (very) sub-optimal
interleave factor.
So I wonder if these controllers advertised as 1:1 can deliver 1:1 as the
optimal interleave in *any* machine with *any* drive?
What started this off was a DTK XT controller that was supposedly 1:1.
(Curses! I don't have the model no. but it was also an RLL controller).
It was in a 12-MHz XT with a 30 Meg RLL drive (Miniscribe, I think) that
was, in fact, interleaved at 1:1 according to two programs I have that
test such things. However, both pgms suggested 1:4 was the optimal
interleave and so it certainly seemed to be.
A Western Digital tech guy said they did not make a 1:1 controller for
XT's and doubted that it was possible.
So, I'm ultimately wondering if DTK (or anyone else) really does make a
1:1 XT controller, if so, does it really work in all machines, and, if so,
why did 1:1 appear to be sub-optimal in the machine I saw.
Thanks very much for any enlightenment on this subject and, please, if you
post a response, please mail me a copy as I'll likely miss it on the news.
(This site only keeps things for a day or two and I just can't log in that
often! :-( ). Thank's again.
--
John Wright ////////////////// Phone: 902-424-3805 or 902-424-6527
Post: c/o Dr Pat Lane, Biology Dept, Dalhousie U, Halifax N.S., CANADA B3H-4H8
Cdn/Eannet:lane@cs.dal.cdn Uucp:lane@dalcs.uucp or {uunet watmath}!dalcs!lane
Arpa:lane%dalcs.uucp@uunet.uu.net Internet:lane@cs.dal.cascjones@sdrc.UUCP (Larry Jones) (04/12/90)
In article <1990Apr10.095716.1689@cs.dal.ca>, lane@cs.dal.ca (John Wright/Dr. Pat Lane) writes: > Some questions about 1:1 controllers. I understand about hard disk > interleaving and how the optimal interleave factor depends on the > disk drive electronics, controller, and the machine they are operating > in. The preponderance of controllers that advertise being "1:1 capable" > seems to suggest that it's only the controller that determines the optimal > interleave and that a "1:1 controller" will achieve 1:1 optimality in > any machine with any drive. Is that so? Every 1:1 controller I know of has enough memory on the controller to hold one track of data from the drive, guaranteeing that it can read or write an entire track in a single revolution no matter how long it takes to transfer data to and from the main cpu. > On the subject of which components affect optimum interleave factor, > I would suppose that drives are standardized enough that there tends > not to be a significant difference with respect to interleaving. On > the machine side, CPU speed seems not to make a difference: I would > suppose that this is because hard disk I/O is handled through DMA and > that bus speed and the DMA electronics were the big factors here. Am > I correct on these points? All drives rotate at the same speed, so they have no effect at all. Cpu speed can make a difference, but bus speed is more important. DMA in only used on XTs, ATs use the cpu instead because it is marginally faster and the original designers, blinded by DOS, never considered that there might actually be other productive work the cpu could do while waiting for the disk. > I've always assumed that 1:1 interleaving was possible (both to format > a drive and to read it) on any drive/controller/machine combination; just > that if any of them were too slow, 1:1 would be a (very) sub-optimal > interleave factor. Exactly. > So I wonder if these controllers advertised as 1:1 can deliver 1:1 as the > optimal interleave in *any* machine with *any* drive? See above re track cache. > What started this off was a DTK XT controller that was supposedly 1:1. > (Curses! I don't have the model no. but it was also an RLL controller). > It was in a 12-MHz XT with a 30 Meg RLL drive (Miniscribe, I think) that > was, in fact, interleaved at 1:1 according to two programs I have that > test such things. However, both pgms suggested 1:4 was the optimal > interleave and so it certainly seemed to be. > > A Western Digital tech guy said they did not make a 1:1 controller for > XT's and doubted that it was possible. All of the 1:1 controllers I know of are 16 bit controllers, so they only work in ATs, but there's no technical reason to prevent making one for an XT. ---- Larry Jones UUCP: uunet!sdrc!scjones SDRC scjones@SDRC.UU.NET 2000 Eastman Dr. BIX: ltl Milford, OH 45150-2789 AT&T: (513) 576-2070 "You know how Einstein got bad grades as a kid? Well MINE are even WORSE!" -Calvin