SubstructSearchConfig#

class nvmolkit.substructure.SubstructSearchConfig(
batchSize: int = 1024,
workerThreads: int = -1,
preprocessingThreads: int = -1,
maxMatches: int = 0,
uniquify: bool = False,
gpuIds: list[int] | None = None,
)#

Configuration for GPU substructure search execution.

Parameters:

  • batchSize – Number of (target, query) pairs per GPU batch. Default 1024.

  • workerThreads – GPU runner threads per GPU. -1 for autoselect.

  • preprocessingThreads – CPU threads for preprocessing. -1 for autoselect.

  • maxMatches – Maximum matches per pair. 0 for unlimited (default).

  • uniquify – If True, remove duplicate matches that differ only in atom enumeration order. Default False.

  • gpuIds – GPU device IDs to use. None or empty list uses current device only.

property batchSize: int#

Number of (target, query) pairs per GPU batch.

property workerThreads: int#

GPU runner threads per GPU. -1 for autoselect.

property preprocessingThreads: int#

CPU threads for preprocessing. -1 for autoselect.

property maxMatches: int#

Maximum matches per pair. 0 for unlimited.

property uniquify: bool#

Remove duplicate matches differing only in atom enumeration order.

property gpuIds: list[int]#

GPU device IDs to use. Empty list uses current device only.