Microbenchmarks for wasm function calls, collected from various sources. See run.sh for some rudimentary information.
The fib benchmarks are newer, they take a principled approach to exploring all the cases.
The sum-add benchmarks are older, they are a little ad-hoc still, and there are some TODOs. Use with care.
- Note, sum-add/private-table/same-module has been observed to have some very nasty microarchitectural artifacts on Intel systems (both a Xeon and an i7). Beware.
To see summary information about the programs, grep INFO *.js