[yt-dev] rockstar issues

[Christopher Moody]

Hi Stephen, Matt,
I haven't run rockstar with multiple readers at all. It also doesn't
segfault for me, and my scripts successfully continue churning on.

Also, I've found that in a few places the force resolution is particularly
touchy. What I naively thought -- that setting the force resolution to 3x
the native resolution would decrease the number of low-mass halos and nail
down the high mass ones -- actually created spurious, *high*-mass halos.
I've seen one instance where two large halos are found only ~5kpc from each
other, but the second one clearly does not exist, and disappears when the
resolution is dropped to from 200pc -> 60pc. I only mention this because we
were talking about this off-list, and because in casual testing I noticed
this counterintuitive trend. Of course, it could just be that I'm only
looking at a few snapshots... but we should definitely keep an option open
for manually setting the force resolution.

chris


On Fri, Nov 9, 2012 at 1:02 PM, Stephen Skory <s at skory.us> wrote:

> Hi Matt,
>
> Thanks for the reply!
>
> >> - When you run rockstar, does it always die with a segfault? On both
> >> mac and linux, after everything is done the run dies with a segfault
> >> that I can't make heads or tails of right now. It doesn't seem to
> >> affect the content of the output, but it's obviously wrong.
> >>
> >
> > This sounds familiar, but I thought it had been addressed.  I think it
> > might be related to communicators not being deallocated correctly.
> > Figuring out where it dies would be step 0 to fixing it ... :)
>
> I think I've found it. What was essentially happening is the
> workgroups are supposed to be deleted inside a loop:
>
> for wg in workgroups:
>     del wg
>
> but this doesn't actually delete items from the list, of course.
> However, because there was some deletion of things inside wg in that
> loop, I think there were references that were left dangling. I'll
> issue a PR soon for just that fix.
>
> Regarding the duplicate halos, I haven't tracked that one down yet.
>
> --
> Stephen Skory
> s at skory.us
> http://stephenskory.com/
> 510.621.3687 (google voice)
> _______________________________________________
> yt-dev mailing list
> yt-dev at lists.spacepope.org
> http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.spacepope.org/pipermail/yt-dev-spacepope.org/attachments/20121109/420b788b/attachment.htm>