## Deja Vu impossible?

### Deja Vu impossible?

Hi,

i found that new challenge 'Remeber me?' pretty solvable. But this one keeps bugging me like:

HVM run ERROR: too many cycles: 10002 (PC=42, STACK_SIZE=4)

HVM run ERROR: too many cycles: 10002 (PC=42, STACK_SIZE=4)

HVM run ERROR: too many cycles: 10002 (PC=42, STACK_SIZE=4)

10002 cycles?? if you calcualte the worst case, then you have to check 128!/(2!*126!) = 64*127 = 8128 tuples, which leaves only a sparse 10002/8128 =~ 1.230.... operations (!) per commparsion, which is somehow impossible to archieve.

also submitting the program '1<p!' gives me the feeling that this schallenge is not meant to be 'fuzzed' ...

Cheating, like only returning a single element out of the first N memcells seems also impossible, as even my second program cannot compare more than 50 or so of the 128 memcells, and the dups never seem to be all in the beginning by chance...

Is there more to it or is this just a bug?

i found that new challenge 'Remeber me?' pretty solvable. But this one keeps bugging me like:

HVM run ERROR: too many cycles: 10002 (PC=42, STACK_SIZE=4)

HVM run ERROR: too many cycles: 10002 (PC=42, STACK_SIZE=4)

HVM run ERROR: too many cycles: 10002 (PC=42, STACK_SIZE=4)

10002 cycles?? if you calcualte the worst case, then you have to check 128!/(2!*126!) = 64*127 = 8128 tuples, which leaves only a sparse 10002/8128 =~ 1.230.... operations (!) per commparsion, which is somehow impossible to archieve.

also submitting the program '1<p!' gives me the feeling that this schallenge is not meant to be 'fuzzed' ...

Cheating, like only returning a single element out of the first N memcells seems also impossible, as even my second program cannot compare more than 50 or so of the 128 memcells, and the dups never seem to be all in the beginning by chance...

Is there more to it or is this just a bug?

You need a complete different method here.

My program really checks every number (in the worst case) and still takes less than 10000 cycles.

Actually, there are some inputs where my program does not work properly, but the probability for this is very low.

With further improvements I think it is feasible to print out the duplicate number for ALL valid input arrays. That would mean a 100 percent probability of success.

I'm not saying that your way of looking at it is wrong, but it is a logical step you have to take in finding the solution. And maybe in a way, with discussing this, we are giving a big spoiler for finding the correct solution.