diff options
Diffstat (limited to 'astro/oskar')
| -rwxr-xr-x | astro/oskar/run_oskar.py | 17 |
1 files changed, 12 insertions, 5 deletions
diff --git a/astro/oskar/run_oskar.py b/astro/oskar/run_oskar.py index dd45b34..744dda3 100755 --- a/astro/oskar/run_oskar.py +++ b/astro/oskar/run_oskar.py @@ -1,11 +1,14 @@ #!/usr/bin/env python3 # -# Copyright (c) Weitna LI <weitian@aaronly.me> +# Copyright (c) 2017-2018 Weitian LI <weitian@aaronly.me> # MIT License # # Run OSKAR interferometer simulator, with corresponding settings # applied to the common configurations. # +# OSKAR settings doc: +# http://www.oerc.ox.ac.uk/~ska/oskar2/OSKAR-Settings.pdf +# import os import re @@ -17,7 +20,7 @@ from time import time def run_oskar(configfile, model, freq, vis_ms, vis_oskar=None, telescope=None, chunksize=None, use_gpus=None, - num_devices=None, dryrun=False): + gpu_ids=None, num_devices=None, dryrun=False): """ Update simulation settings in the configuration file, and run the simulator ``oskar_sim_interferometer``. @@ -32,6 +35,7 @@ def run_oskar(configfile, model, freq, vis_ms, vis_oskar=None, for item, value in [ ("simulator/max_sources_per_chunk", chunksize), ("simulator/use_gpus", use_gpus), + ("simulator/cuda_device_ids", gpu_ids), ("simulator/num_devices", num_devices), ("sky/oskar_sky_model/file", model), ("telescope/input_directory", telescope), @@ -82,7 +86,10 @@ def main(): "(default: 'visibility/')") parser.add_argument("--num-devices", dest="num_devices", type=int, default=None, - help="number of GPU devices or CPU threads to use") + help="number of CPU threads to use") + parser.add_argument("--gpu-ids", dest="gpu_ids", default=None, + help="comma-separated GPU device IDs to use; " + + "(default: use all GPU devices)") exgrp1 = parser.add_mutually_exclusive_group() exgrp1.add_argument("--use-gpu", dest="use_gpu", action="store_true", help="force to only use GPUs") @@ -140,8 +147,8 @@ def main(): run_oskar(configfile=configfile, freq=freq, model=skyfile, vis_ms=vis_ms, vis_oskar=vis_oskar, telescope=args.telescope, chunksize=args.chunksize, - use_gpus=use_gpus, num_devices=args.num_devices, - dryrun=args.dryrun) + use_gpus=use_gpus, gpu_ids=args.gpu_ids, + num_devices=args.num_devices, dryrun=args.dryrun) t2 = time() print("=============================================================") |