Re: execution terminating right after loading in particle file

From: Volker Springel <volker_at_MPA-Garching.MPG.DE>
Date: Wed, 29 Dec 2010 20:52:16 +0100

Hi Chris,

This crash appears to happen when the code initializes a look-up table for integrals of cosmological factors, which are then used in the drift- and kick operations. This is done in the routine driftfac.c, and whatever goes wrong here, it should not depend on your initial conditions, but only at the cosmological parameters and the starting and end redshifts that you specify. Can you pass on an example of a parameter set that causes the problem?

It might simply be that the convergence criterion specified for the gsl_integration_qag() calls is set to something too aggressive for peculiar parameter combinations. In this case, a poor man's solution might be to lower the 1.0e-8 to 1.0e-5 or so.


On Dec 29, 2010, at 5:08 PM, Chris Orban wrote:

> Hi all,
> Although I've had much success running Gadget2 simulations, I
> recently ran across an error I haven't seen before:
> gsl: qag.c:261: ERROR: could not integrate function
> Default GSL error handler invoked.
> [node49:21808] *** Process received signal ***
> [node49:21808] Signal: Aborted (6)
> [node49:21808] Signal code: (-6)
> I'm running ensembles of simulations and some realizations in the
> ensemble are failing with this message immediately after loading in the
> initial conditions file. (n.b. I can't just run more realizations,
> discarding the failed ones.)
> As far as I can tell the particle positions in the initial conditions
> file look fine and I've measured the power spectrum of the initial
> conditions file and it looks good too. Also, the problem persists
> regardless of whether I set up the ICs+code to begin at a very early or
> a very late starting redshift. I'm running out of ideas for why this
> might occur.
> Any thoughts?
> Chris
> -----------------------------------------------------------
> If you wish to unsubscribe from this mailing, send mail to
> with a subject of: unsubscribe gadget-list
> A web-archive of this mailing list is available here:
Received on 2010-12-29 20:52:16

This archive was generated by hypermail 2.3.0 : 2023-01-10 10:01:31 CET