Engineering evidence
A growing collection of reproducible large-file tests, output validation, discovered limitations and documented optimization work.
Published entries
New tests will be added here as separate articles so that each dataset, application build and method remains clearly identifiable.
Benchmark E57-01 · 15 August 2026
A pre-optimization test of CloudCutter 1.1 beta using a 26.089 GiB registered industrial E57 with 960 million points and 39 structured scans.
Benchmark E57-02 · 16 August 2026
A two-run large-scale E57 test with 249 registered scans and six outputs, documenting near-linear full-read scaling, delayed preview and the current export read amplification.
Publication principles
The series is intended as an engineering log, not a collection of isolated promotional numbers.
Dataset fingerprints, software versions, hardware, storage and test modes are disclosed where possible.
Reading, operator preparation and export are treated as different parts of the workflow.
Point counts, parser checks and independent application checks are recorded together with limitations.
