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