RcppRoll

This package provides windowed-versions of commonly-used mathematical and statistical functions.

Install the latest release from CRAN with:

install.packages("RcppRoll")

Or, install the development version with:

install_github("kevinushey/RcppRoll")

Parallelism

When compiled with OpenMP support, RcppRoll parallelizes its rolling computations across threads. Results do not depend on the number of threads used. By default, the thread count is chosen by the OpenMP runtime (typically controllable through the OMP_NUM_THREADS environment variable); pin it with options(RcppRoll.threads = <n>), or set it to 1 to disable parallelism. Check what your installation supports with:

RcppRoll::roll_threads()

which reports the number of threads to be used, or NA if the package was compiled without OpenMP support. The package reports this on attach as well; use options(RcppRoll.quiet = TRUE) to suppress the startup message.

The toolchains normally used on Linux and Windows support OpenMP out of the box, so installations from sources just work. Apple's macOS toolchain ships without OpenMP support; the R project documents how to enable it at https://mac.r-project.org/openmp/, which provides OpenMP runtimes matching Apple's compilers and recommends adding

CPPFLAGS += -Xclang -fopenmp
LDFLAGS += -lomp

to ~/.R/Makevars. Alternatively, with Homebrew's OpenMP runtime (brew install libomp), add the following line to ~/.R/Makevars instead:

SHLIB_OPENMP_CXXFLAGS = -Xclang -fopenmp -I/opt/homebrew/opt/libomp/include -L/opt/homebrew/opt/libomp/lib -lomp

Either way, reinstall the package from sources afterwards, and confirm the result with roll_threads().