Simple and efficient GPU parallelization of existing H-Matrix accelerated BEM code

Connections are not prepared for this record yet, or its source metadata has changed. The original record remains available while background snapshots are rebuilt.