Reduce memcpys in substruct search and improve benchmark autotuning#240

scal444 merged 4 commits into

NVIDIA-BioNeMo:mainNVIDIA-BioNeMo/nvMolKit:mainfrom

scal444:substruct-autotune-and-pattern-uploadscal444/nvMolKit:substruct-autotune-and-pattern-uploadCopy head branch name to clipboard