Large-export performance benchmarking • gp3tools

Purpose

This workflow measures elapsed time, approximate R-heap growth, output size, and empirical scaling across increasing Gazepoint export sizes. It is a regression framework rather than a hardware-independent speed claim.

gp3tools_performance_limits()
#>   operation max_seconds_per_million_rows max_heap_delta_mb_per_million_rows
#> 1  generate                           90                               1200
#> 2    import                          240                               1800
#> 3    master                          240                               1800
#> 4  sampling                          180                               1200
#> 5   quality                          180                               1200
#>   max_scaling_exponent
#> 1                  1.6
#> 2                  1.6
#> 3                  1.6
#> 4                  1.6
#> 5                  1.6

The default limits are deliberately conservative. Projects can supply a machine-specific baseline and stricter ratio limits after collecting stable measurements on the same system.

Fast contract run

A small run verifies the complete benchmark contract without slowing package checks.

small <- benchmark_gazepoint_export_performance(
  scales = data.frame(
    total_rows = c(1000L, 5000L),
    n_files = c(1L, 2L)
  ),
  operations = "generate",
  trials = 1L
)

small$summary
#>   scale_id total_rows n_files rows_per_file operation n_trials n_success
#> 1        1       1000       1          1000  generate        1         1
#> 2        2       5000       2          2500  generate        1         1
#>   median_elapsed_s minimum_elapsed_s maximum_elapsed_s median_heap_delta_mb
#> 1            0.003             0.003             0.003             1.949677
#> 2            0.006             0.006             0.006             3.578438
#>   maximum_heap_delta_mb median_output_size_mb
#> 1              1.949677             0.1873322
#> 2              3.578438             0.9200592
small$regression$overall
#>    pass n_checks n_pass n_fail
#> 1 FALSE       12     11      1

Large-export profile

The installed script uses 60,000, 240,000, and 960,000 rows distributed across 1, 4, and 16 files, respectively. Run it outside ordinary R CMD checks:

source(
  system.file(
    "benchmarks",
    "run-large-export-performance.R",
    package = "gp3tools"
  )
)

On Windows CMD, set GP3TOOLS_BENCHMARK_OUTPUT to an external folder before running the script. Benchmark outputs should not be committed when they contain machine-specific paths or private export data.

Baseline-relative regression checks

A saved result can be used as a baseline:

current_audit <- check_gazepoint_performance_regression(
  current_benchmark,
  baseline = previous_benchmark,
  elapsed_ratio_limit = 1.25,
  memory_ratio_limit = 1.25
)

Absolute and baseline-relative failures remain visible in the audit table.