Currently only works with csv files.

To run dev version simply use: cargo run my-input.csv

Supports concurrent mode, which distributes clients across different threads to go though transactions faster.

In order to run in concurrent mode, pass in the option to the cli:

cargo run -- --concurrent my-input.csv

It uses client's id to decide to which thread to send it to. Simply applies modulo(%) operator to the client_id and based on result sends to corresponding thread.

So thread to which client goes to = client_id % thread_count. thread_count depends on cpu core count, can be customized from code though.

This means that in very rare and worst cast scenario, all clients will go to the same thread. In such case it will be a little bit slower than single threaded version, since we have to pay cost for sending transactions through the channel.

For example if we have 8 cores, hence 8 threads and we receive transactions for clients: 1, 9, 17, 25, 33, ... will all run on the same thread.