Skip to content

SWIFT stuck on first step #74

Description

@titanbin

Hello, I am trying to run SWIFT on an isolated galaxy setup on the TGCC cluster, but after running the steps before starting the configuration, it seems to be stuck forever after this message in the log :

[00009.8] main: Running on 2500000 gas particles, 0 sink particles, 13000000 stars particles 0 black hole particles, 0 neutrino particles, and 4998152 DM particles (20498152 gravity particles)
[00009.8] main: from t=0.000e+00 until t=1.000e+01 with 1 ranks, 16 threads / rank and 128 task queues / rank (dt_min=1.000e-09, dt_max=1.000e-02)...
My config setup
--with-fftw=/ccc/products/fftw3-3.3.9/intel--20.0.0__openmpi--4.0.1/default \
--enable-parallel-hdf5 \
--enable-optimization=yes \
--enable-mpi-mesh-gravity \
--with-hydro=sphenix \
--with-cooling=EAGLE \
--with-chemistry=EAGLE \
--with-feedback=EAGLE \
--with-entropy-floor=EAGLE \
--with-stars=EAGLE \
--with-star-formation=EAGLE \
--with-tracers=EAGLE \
--with-tbbmalloc \
--with-parmetis=/ccc/products/parmetis-4.0.3/intel--20.0.0__openmpi--4.0.1/default \
--with-metis=/ccc/products/metis-5.1.0/intel--20.0.0/default \
--with-numa=/ccc/products/numaprof-1.0.0/system/default \
--with-gsl=/ccc/products/gsl-2.6/intel--20.0.0/default
My run setup
# Use -N node
#MSUB -N 1
# Use -n MPI ranks in total
#MSUB -n 1
# -c cores per MPI rank
#MSUB -c 16
# Exclusive use of the node
#MSUB -x

set -x

ccc_mprun ../swift_eagle_2 --threads=16 --feedback --self-gravity --stars --star-formation --cooling --hydro --limiter --sync Happy2.yml 2>&1 | tee output.log

My simulation has 2.5M hydro parts, 13M star parts and 5M DM parts. I waited several hours and not a single step has completed. I'm guessing it has to do with the initial domain decomposition ?

Any idea how I can fix this ?

Thanks !

Metadata

Metadata

Labels

No labels
No labels

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions