Hi Jun,
speed-wide, upgrading to the newest STP would not make much
difference. Almost all time in Hampi is spent in the encoding step
before STP and the actual STP solving is maybe 5% of time (I'm
guessing). So even if latest STP was 1000% faster, Hampi would still
see only 4.5% improvement. That's why I've been reluctant to putting
effort into upgrading to latest STP.
However, if STP were enhanced to handle string equations or regexp
constraints, then it may be beneficial to move to the latest STP. But
then, a more involved migration would be required to push more work
down to the STP level.
So, for now, feel free to compare the 'something' to the current
version of Hampi. Let me know how it goes.
./adam