Re: multiply files in one (was GNU/Linux stance by Richard Stallman)

Larry McVoy (lm@bitmover.com)
Mon, 29 Mar 1999 23:28:56 -0800

: > loop and only you can break out of it - go think about it. Model it.
: > Figure out exactly how long it takes to read a bunch of little files
: > all in their own blocks and contrast that with how long it would take
: > to read all those blocks in one large I/O.
:
: *You're* the one with the proposal, *you* model it. Do it with Linux, since
: this is linux-kernel, not sunos-kernel. Do it with several different
: distributions of file sizes. Everyone can do the maths, but sufficient people
: have come up with sufficient potential problems to indicate that the way you
: want to do the maths might not necessarily prove to be
: effective/efficient/useful/winning in real life.

It's statements like this that make the people who know how to solve
the problems just shrug and say "let 'em figure it out on their own,
they don't want any help". Most of the people who know how to solve
problems don't read this list anymore because of the incredibly low
signal/noise ratio.

As for me modeling it, I have, buddy, I have. At the disksort level,
at the file system level (both allocation/write and read paths), and at
the application level. Repeatedly in multiple different file systems.
Oh, yeah, and at the disk drive level itself.

What I've told you is all based on over 10 years of modeling this stuff.
If you choose to ignore that work, that's fine, but maybe you should
check into what has already been done before dismissing the results.

-
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/