lkml.org 
[lkml]   [1999]   [Mar]   [23]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [OFFTOPIC] Re: disk head scheduling
Date
From
: > > "Richard B. Johnson" <root@chaos.analogic.com> writes:
: > > > [snip]
: > > > Expensive disk drives now do full track buffering. This costs
: > > > money because RAM costs money. To buffer one full track on a
: > > > Disc drive requires CAPACITY / (HEADS * CYLINDERS) which can
: > > > be upwards of 100 megabytes of high-speed SRAM.

You are confusing a track with a cylinder. A track sits on one platter
and is typically quite small. A cylinder is the same track on multiple
platters, and I don't know of any drives which buffer all tracks in a
cylinder at the same time. In fact, I'll bet long odds that none do.
Here's why. If the drive manufactors could read all tracks at the same
time, they would do so. That would bump the drive speed from the current
platter speed to N_HEADs * (platter speed), which would be extremely nice.

However, they can't do that. The tracks within the same cylinder are not
all at exactly the same place. Depending on which platter you are going for,
the heads settle at a slightly different place. Very annoying that. If it
weren't for that little problem, we could have 5 platter drives which were
65MByte/sec drives.

Anyway, the point is that nobody puts 100MB of buffer on their drives.
I know Quantum is thinking about fairly large buffers but I think those
are more like 32MB, and regardless, the reason for big buffers is not to
buffer a cylinder (as nice as that would be).

--lm

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:50    [from the cache]
©2003-2011 Jasper Spaans