Skip to contents

Auxiliary for Controlling GLM Fitting in Parallel

Source: R/parglm.R
parglm.control.Rd

Auxiliary function for parglm fitting.

Usage

parglm.control(
  epsilon = 1e-08,
  maxit = 25,
  trace = FALSE,
  nthreads = parallelly::availableCores(omit = 1L),
  block_size = NULL,
  method = "LINPACK",
  nthreads_auto = missing(nthreads)
)

Arguments

epsilon

positive convergence tolerance.

maxit

integer giving the maximal number of IWLS iterations.

trace

logical indicating if output should be produced during estimation.

nthreads

number of cores to use. Defaults to parallelly::availableCores(omit = 1L), which leaves one core free. You may get the best performance by using all available physical cores if your data set is sufficiently large.

block_size

number of observations to include in each parallel block.

method

string specifying which method to use. Either "LINPACK", "LAPACK", or "FAST".

nthreads_auto

logical; for internal use only. Records whether nthreads was auto-detected (suppresses the thread-reduction warning when the dataset is small). Do not set this argument directly.

Value

A list with components named as the arguments.

Details

The LINPACK method uses the same QR method as glm.fit for the final QR decomposition. This is the dqrdc2 method described in qr. All other QR decompositions except the last are made with DGEQP3 from LAPACK. See Wood, Goude, and Shaw (2015) for details on the QR method.

The FAST method computes the Fisher information and then solves the normal equation. This is faster but less numerically stable.

References

Wood, S.N., Goude, Y. & Shaw, S. (2015) Generalized additive models for large datasets. Journal of the Royal Statistical Society, Series C 64(1): 139-155.

Examples

# use one core
f1 <- parglm(mpg ~ wt + hp, data = mtcars, family = Gamma(link = "log"),
             control = parglm.control(nthreads = 1L))

# use two cores (mtcars has 32 rows, sufficient for 2 threads)
f2 <- parglm(mpg ~ wt + hp, data = mtcars, family = Gamma(link = "log"),
             control = parglm.control(nthreads = 2L))
all.equal(coef(f1), coef(f2))
#> [1] TRUE