bench/Micro.purs contains a microbenchmark exercising decoding various structures.
This is done using synthetic datatypes, to test how the operations behave with regards to the size of the datatype. For example R3 is a record with 3 fields, R10 is a record with 10 fields, Enum10 is an enum with 10 constructors, and so on.
Decoding is performed using multiple libraries (currently unscramble, argonaut, foreign-generic and simple-json) for comparison.
Use npm run bench-micro to run it using Node.js.
You can pass additional command line arguments to filter the cases executed. For example, npm run bench-micro Enum10 Enum30 will run only benchmarks related to Enum decoding.