On Wed, May 15, 2013 at 11:28:23PM +0100, Dave Love wrote:
Jesse Becker <[email protected]> writes:
I seem to recall hearing that "5 halflives" is how long radioactive
stuff has to decay before it's "safe." Don't quote me on that though.
:)
That rather depends on how unsafe it was to start with...
Nah, 5 halflives is 5 halflives. It's just a question of *time* more
than "initial danger". ;-)
It also looks like a real exponential decay is used, instead of a linear
decrease (as in some of the load calculations).
That's implied by half-time/half-life, though I can spell it out in the
doc.
Well, at least in one place (load adjustment?), there's a similar sort
of "time-based weighted backoff", but it's linear in nature (and
documented as such). Sometimes people are sloppy about terminology, and
I wanted to check. :)
This is especially interesting since it implies that negative halftimes
are acceptible. Sure enough, setting a negative value zeros out
historical usage:
https://blogs.oracle.com/sgrell/entry/a_couple_lines_on_halftime
[It's in sched_conf(5).]
A value of "0" is documented, but negative values are not.
So yes, you'd need to keep your accounting files around for some number
of halftimes.
The accounting data just aren't generally useful in this context. See
reporting(5) about recording sharelog, which you can stuff in postgres
with dbwriter, for instance.
Agreed.
--
Jesse Becker
NHGRI Linux support (Digicon Contractor)
_______________________________________________
users mailing list
[email protected]
https://gridengine.org/mailman/listinfo/users
