benchmarks.md

Benchmarks

The repository provides also a simple implementation of a bulk load function in order to benchmark the general speed of the network in terms of tps (transactions-per-second).

fabkit benchmark load [jobs] [entries]

# e.g.
fabkit benchmark load 5 1000

The example above will do a bulk load of 1000 entries times 5 parallel jobs, for a total of 5000 entries. At the completion of all the jobs it will be prompted on screen the elapsed time of the total task.

Note: Kepe the number of jobs not higher than your CPU cores in order to obtain the best results. This implementation does not provides a complete parallelization.

To achieve the optimal result it is recommended to install Gnu Parallel and use as it follows:

parallel ./fabkit benchmark load {} ::: [entries])

# e.g.
parallel ./fabkit benchmark load {} ::: 20
# prefix with FABKIT_DEBUG=true to see more logs