Hello Simon, Can you please provide the test cases that you used to plot the performance graph you've attached. Also do you know if your optimization will be useful for a very large amount of subtransactions per transaction (around 1000)?
-- Thanks, Yaroslav
