RBenchmarking

Directories and files descriptions:

• RHPCBenchmark - This is the R HPC benchmark package directory containing source code and vignette. The package directory does not contain autogenerated directories and files needed for a complete R package. To generate manual page documentation and perform checks on the package use the update_benchmark_package.R script. To generate the package use, generate_benchmark_package.R

• RHPCBenchmarkData - This is the companion data package to the R HPC benchmark package. It contains the sparse matrices used by the sparse matrix benchmark function in the R HPC benchmark package. To generate manual page documentation and perform checks on the package use the update_data_package.R script. To generate the package, use generate_benchmark_package.R.

• dense_matrix_driver.R - This script runs the dense matrix microbenchmarks.

• sparse_matrix_driver.R - This script runs the sparse matrix microbenchmarks.

• machine_learning_driver.R - This script runs the machine learning microbenchmarks.

• dense_matrix_driver_win.R - Tests the dense matrix benchmark on the IUAnyware Microsoft Windows R environment.

• sparse_matrix_driver_win.R - Tests the sparse matrix benchmark on the IUAnyware Microsoft Windows R environment.

• test_machine_learning.R - Tests the machine learning benchmark on the IUAnyware Microsoft Windows R environment.

• update_benchmark_package.R - Builds the manual page documentation for all R objects in the RHPCBenchmark package and performs validation checks.

• update_data_package.R - Build the manual page documentation for all R objects in the RHPCBenchmarkData package and performs validation checks.

• update_benchmark_vignettes.R - This script builds and installs the benchmark package vignettes as part of a full package install. Installaing the package is the only way to view the vignettes for proofreading.

• update_data_vignettes.R - This script builds and installs the data vignettes as part of a full package install. Installaing the package is the only way to view the vignettes for proofreading.

• build_benchmark_package.R - Builds the benchmark package.

• build_data_package.R - Builds the companion data package.

Package validation and build process

To build the R HPC benchmark package:

1. Run the update_benchmark_package.R script
2. Fix any validation WARNINGs and ERRORs reported in the standard output and log file
3. Repeat steps 1 and 2 until problems are resolved
4. Run the build_benchmark_package.R script
5. Run the following on the built package for further validation: R CMD check --as-cran [gzipped tar file package]
6. Fix any validation WARNINGs and ERRORs reported in the standard output and log file
7. Repeat process from step 5 until no WARNINGs or ERRORs are detected

A similar process applies to checking and building the companion RHPCBenchmarkData package.