MUQSAC++

What was done

• The pipeline existed as 14 near-duplicate copies, one per experiment (461 files audited); 28 of the 33 distinct file names were byte-identical across all copies.
• The copies were merged into one source tree. The behaviours that differed between copies are kept as registered variants, selected by a single input-file setting, instead of by copying the tree again.
• Byte-diff equivalence tests run each original variant next to the unified code on the same synthetic input and compare the output files byte for byte.

Direction

Sampling strategies for species-concentration profiles; results are not shown here.

Credit

The related curve-capturing technique is joint work led by Gautam Srikanth (accepted abstract, 64th Symposium (Japanese) on Combustion, Fukuoka, Nov 2026).