Skip to Content.
Sympa Menu

sm-grimoire - Re: [SM-Grimoire] Starting from scratch (again)

sm-grimoire AT lists.ibiblio.org

Subject: Discussion of Spells and Grimoire items

List archive

Chronological Thread  
  • From: Seth Woolley <seth AT tautology.org>
  • To: Andrew <afrayedknot AT thefrayedknot.armory.com>
  • Cc: sm-grimoire AT lists.ibiblio.org
  • Subject: Re: [SM-Grimoire] Starting from scratch (again)
  • Date: Tue, 10 Jun 2003 17:18:26 -0700 (PDT)

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Slight benefit for those programs that can't handle parallel make, I
suppose, but you're right, and we do support parallel compiling ... just
run the casts in two separate terminals and I've never had anything really
break because of that. It might want to detect to see if a cast is coming
up in the queue of another existing cast and then ask if you really want
to compile that twice or wait for the spell to finish casting in the other
cast session and then continue (skipping the new cast for that spell)
after it's compiled in the other cast.

There's extensive locking already in sorcery, but I suppose there could be
more.

I think what they were talking about that sorcery does was to handle
parallel dependency casting (not sure though). Does it compile as many at
once unless a dependency is missing and then it lock-waits for the
dependency to finish (like a summon almost) and then continues? Each
spell would have to keep track of its recursive dependencies, but once
they had that info, one could do that pretty easily with a for...sleep.

Seth

On Tue, 10 Jun 2003, Andrew wrote:

> >
> > It won't unless you are on SMP.
> >
>
> you could just edit compile_config and have more than one make job
> per cast, which effectively turns the argument around again.
>
> Although given appropriate locking (nudge nudge), theres no reason why we
> cant run multiple casts simultaneously.
>
> As it stands now, cast generates a makefile from the dependancy tree,
> then calls make on that file, which calls cast (with an envivonment
> variable set) on each of the targets, theres no reason why that first
> call cant just have a -j $parallel_casts_number set on it also.

- --
Seth Alan Woolley <seth at tautology.org>, SPAM/UCE is unauthorized
Key id 7BEACC7D = 2978 0BD1 BA48 B671 C1EB 93F7 EDF4 3CDF 7BEA CC7D
Full Key at seth.tautology.org and pgp.mit.edu. info: www.gnupg.org
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.0 (FreeBSD)

iD8DBQE+5nVU7fQ833vqzH0RArboAKDAkWnGngKINSReiBtdaj5wf3Wz1gCdHKq0
wFxczHeNd5RgWFIaQ1IDb68=
=OWl6
-----END PGP SIGNATURE-----





Archive powered by MHonArc 2.6.24.

Top of Page