The situation seems as nightmare. Less than a kilobyte for code+data or just data? I remember one good chess program from the old MS-DOS days but it also had hundreds of kilobytes to function. I've a doubt if you can write a code for a movegen assuming just 12bit for single move you are unable to ...

It seems normal behavior. Why you expect to see some thing different.
In my opinion its better to clear the table every time engine starting to search, because it stores only a hash value (just like a checksum), there is no guaranty that there will be no conflict between two position's hash, so you ...

It look like incomplete work. Gives me error for every any UCI command!!

This was also one of my problems on first day. This code from my engine (at this moment simply alpha beta TT) maybe useful. I think its much better for understanding than explanation
function Search(Board: TBitBoard; Depth: Integer; Alpha, Beta: Integer; CurrentBoardKey: UInt64; MoveInfo: TMoveInfo ...

Yes, exactly. because i just set that to search very limited fixed depth since there was no improvement to raw alpha beta such as move ordering or PVS or any pruning techniques.

I thought about many kind of ways to implement good search but some are not possible due to memory limitation and some ...

I experienced this problem when i include check moves on quiescence search.as a simple debug technique you may want to try putting depth limit for including checks in quiescence. or exclude them to see if the problem persist
// move list
moves := pos.generate_moves()
just as a doubt: does this ...

Bad move ordering will make the search take much longer, but eventually it should get the right answer.
Do you have a quiescence function? Sometimes bad moves are made when a quiescent search would have shown that the move fails due to a simple chain of captures.

A good thing to do to debug your ...

Move order will not make the engine play ridiculous moves. Are most of the moves nonsense or just the odd one?

Other moves are not perfect and according to the engine strength are acceptable. but some of moves are buggy.
Can you post an example of a position and the move in question?
position ...

Hi everyone,
After finishing some of the base parts of engine such as (Simple UCI Adapter)/MovGen/DoMove/(Simple fixed depth Alpha Beta Search)/QuiescenceSearch(excluding check moves) and very simple evaluation (as simple as material and score arrays for piece on position) I noticed that engine ...