af2a046Von Karman modal solver on a Cartesian grid masked to the disk, site, and workflowJeremy Magland 1
f124192Chain procedure: 8 mm on CPU, 4 mm on GPU up to 600 modesJeremy Magland 2# Produce the vonkarman-modal chain for cymbal-flat-disk-v1. Run locally, never on the vault runner.
3# The 8 mm members run on a CPU (minutes); the 4 mm members need a GPU (--device cuda, PyTorch) to be
4# practical, since the per-step cost is the memory bandwidth of the coupling tensor (1.7 GB at 600 modes).
e396cedChain script for the 1000 N strike; drop the solutions of the superseded strikeJeremy Magland 5# Usage: bash run_chain.sh <cymbal-repo>/challenges/cymbal-flat-disk-v1/challenge.json [gpu|cpu] (cpu: hours for the 4 mm members)
af2a046Von Karman modal solver on a Cartesian grid masked to the disk, site, and workflowJeremy Magland 6set -e
7here="$(cd "$(dirname "$0")" && pwd)"
8ch="$1"
af2a046Von Karman modal solver on a Cartesian grid masked to the disk, site, and workflowJeremy Magland 10out="$here/../../solutions/cymbal-flat-disk-v1/vonkarman-modal"
11mkdir -p "$out"
12cd "$here"
e396cedChain script for the 1000 N strike; drop the solutions of the superseded strikeJeremy Magland 13dev=""
14[ "$what" = cpu ] || dev="--device cuda"
15python3 vk_modal_disk.py "$ch" "$out" $dev --fields g4-w300-a300-m2 g8-w100-a100-m2 g8-w200-a200-m2 g4-w200-a200-m2 g4-w300-a300-m2
16python3 make_submission.py "$out" g8-w100-a100-m2 g8-w200-a200-m2 g4-w200-a200-m2 g4-w300-a300-m2