Parallel algorithms for mining frequent structural motifs in scientific data