Re: Remote fork() and Parallel programming

Larry McVoy (lm@bitmover.com)
Tue, 16 Jun 1998 10:03:25 -0700


: If I see someone doing busy waiting, I'd say he is not a very good
: programmer. Busy waiting should be avoided even in normal applications
: running on a single computer. BTW, please refer to Linux's kernel sources to
: see if it uses busy waiting to implement its semaphores (hint: it doesn't,
: processes sleep while waiting).

They do, huh? You're right about processes that are executing Sys5
semaphores. However, the kernel itself is riddled with spin locks.
And most library wrappers around whatever locking mechanism you use will
spin for a while before going to sleep - it's a win on MP machines.

At this point, I think that you need to simpy show up with a pile of data
that supports your position. The other side of the argument has both a
pile of experience and a pile of data and a pile of papers writing up the
data, all of which say that DSM is a bad idea. You need to provide the
/data/, not /arguments/, which supports your position.

And you need to do this in some other forum than this list, we've (you
and me and everyone else) have burned up more than our fair share of
bandwidth on this topic on this list. Feel free to discuss it on the
clusters@bitmover.com list.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu