rcpp_hdf5dataset_multiply: General matrix product for HDF5 datasets (R6 wrapper)

In BigDataStatMeth: Title: Scalable Statistical Computing with HDF5-Backed Matrices

General matrix product for HDF5 datasets (R6 wrapper)

Description

Computes A %*% B (or transposed variants) for two HDF5 datasets referenced by external pointers, using the BigDataStatMeth block-wise multiplication algorithm.

Usage

rcpp_hdf5dataset_multiply(
  ptr_a,
  ptr_b,
  transpose_a = FALSE,
  transpose_b = FALSE,
  paral = NULL,
  block_size = NULL,
  threads = NULL,
  compression = NULL,
  outgroup = NULL,
  outdataset = NULL
)

Arguments

ptr_a	External pointer (SEXP) for matrix A
ptr_b	External pointer (SEXP) for matrix B
transpose_a	Logical; transpose A before multiplying
transpose_b	Logical; transpose B before multiplying
paral	Logical or NULL; enable OpenMP parallelisation
block_size	Integer or NULL; block size (NULL = auto)
threads	Integer or NULL; thread count when paral = TRUE
outgroup	Character or NULL. Output group in the HDF5 file. Default "OUTPUT".
outdataset	Character or NULL. Output dataset name. Default "A_x_B" where A and B are the input dataset names.

Value

Named list with filename (character) and path (character) locating the result dataset within the HDF5 file.

BigDataStatMeth documentation built on May 15, 2026, 1:07 a.m.