Is the only one i haven't completed (let alone improved on). I've seen the ending (and had no clue what it meant) and programmed most of the puzzles with multiple solutions (some gaining a mere 3 cycles), but i'm not sure how to approach it.

Part of it is the layout is annoying, and part of it is that each processor just has so little memory, another 2-4 instructions and i'd probably nail it no problem.

Perhaps another problem is if you try to get an element from an empty stack, it just hangs so you can't let it grab unless you know there's something in there (even if it's a dummy data).

So, if anyone else has solved this one and has some ideas and pointers, great. If not, i'll re-try it a few more times over the next few days before i give up.

On the other hand i can probably answer and hint at all the other puzzles...

The game was only released two days ago? :)

Maybe... Writing a sorting algorithm is easy enough, and the comparator too from scratch, but it's the sheer limitations of the system.

Just finished... Ugg!!!

I took a better design based on my index one, where the two in the middle only handle moving the data from one stack to the other (pretty much), and the stack closest to the output holds the sorted data. If the number is smaller, it goes in the lower stack, if it's larger it moves that stack's top value to the other stack. Once the proper spot is found it resets all the data back.

Curiously it's one of the faster sorts on the list apparently, basically being insertion. If i had just a little more memory to work with, i could make it much faster by having a count instead of 0 terminate for shifting the data back and forth.

You've hit the nail on the head here with these three issues.

God know I love low level programming, I've enjoyed brainfuck a lot. I love doing assembly programming. But these three issues add an element of annoyance. Not difficulty; just annoyance.

I've almost finished both campaigns. Just did the exponential level where I had an extremely neat solution using 4 processors only without stacks. Except that I needed 2 more lines of code to implement it and I couldn't fit them. Because of that, after an hour of frustration, I ended up using 2 additional processors, 10 or so more lines of code and all that had effect on cycles too.

I love this game; I love the architecture and the parallel work it allows you to do, but because of these annoyances I gave up on the sorting level and 2 or 3 more levels from the second campaign (the one where you find the mode value was another noteworthy pain in the ass; not least because of the annoying additional condition that if there is a tie on top you have to output 0).

Maybe I will go back to these levels in the future.

It would be cool if the game was changed so that you could:
_ Change the layout a bit. i.e. move the output and input nodes, and also be able to move/change (within reason) some of processors around.
_ Give the processors more memory. 20 lines instead of 15 would make a world of a difference.
_ Change the stack implementation so that for example "read all elements till empty" could be implemented more easily.

Yeah... same with Shenzen IO, where it's like 'i need 1 more line... damn... gotta use a bigger processor or split it up'...

Actually made an engine for a game, haven't finished it though, the idea being you're working in similar code with like 2k-4k of room, but you're going to be building more or less the microcode to emulate a 6502 (all 60 documented instructions). Haven't gotten around to finishing it, especially not while on a laptop.

Hmmm i only got 1/3rd of the way through the secondary campaign and some of the problems were just getting too tedious to try and solve. Sorting was by far the hardest i wanted to do, and took me like 4-6 hours to properly solve, whereas if you just had say 50 lines and 2 stacks, you could implement a good sorting routine. The more you have to break up the routine into parts, the less effective it becomes as you have more overhead, as instructions move from say 6 on one processor, to 2-3 and the instructions of work go on the other processor... bloat mostly.

Just do what I did. If it's too tedious, copy a solution from the internet. Some others are interesting nonetheless. So far I've copied 4 solutions from the net. Sorting from the main campaign and 3 others from the second campaign.

As you said, they're not difficult per se, just tedious. So to save me the frustration and keep the game enjoyable that's what I did.

And I finished the game.

The last levels of second campaign were pretty good. Overall I liked the game. But it marked the first time in Zachtronics game where I had to get help.

Needed help on the following levels.

From the original campaign:
_Sequence sorter
From the second campaign:
_Sequence mode calculator
_Back reference reifier
_Prolonged sequence sorter
_Prime factor calculator

I suggest you give the remaining levels a try, rtcvb. If only for the story.

The second story seemed more of a hack job of what i went through, and more like a fan fiction rather than something official.

Yeah. Spacechem story this is not.

I'm pretty sure the second campaign puzzles are all user submitted.

Probably. There's a lot of problems you can solve with very little.

The puzzles, yes. But not the story. I found the puzzles good though.