Forums - Open Redstone Engineers

Hello,

There has been discussion recently about another CPU build-off, so let's do it! Post your ideas below. Please include a brief description of how the winner would be determined, whether teaming would be allowed and possible rules concerning prebuilt parts or other things you deem important. You can also mention what you think would be a good time limit for the challenge. Feel free to discuss ideas.

We will look into doing more build-offs (not just CPUs!) and other events in the future, so if CPUs aren't your thing, don't worry - more is coming. Don't be afraid to give it a try though .
Vote on the poll and discuss if you think 30.-31.3.2019 (Saturday and Sunday) is a fine time for the build-off. Perhaps we can make it so that not everyone has to start at the same time - that way more people could have a chance to participate.

Id love to do a challenge where u can only use wires and torches and blocks (ofc) but thats probs gonna not be the most popular idea but I do like the sound of no premade parts also another time no W/E would also be fun

Repeater only CPU

The coolest part about repeater-only CPU's is their resemblance to real-life electronics. You could do the same thing with comparators, but they don't need a power-supply like repeaters need. To resemble real life as closely as possible, you can use repeaters as a sort of transistor that needs a power and a switch signal. Regular CPU grading conventions would apply. Teams should definitely be allowed as there isn't much prebuild repeater-only stuff floating around.

My suggestion is allowing only blocks, wires, torches and repeaters. No comparators, hoppers, pistons or anything else like that. Just like good old times.
I think W/E should definitely be allowed for something like this as manual bussing is tedious.
Btw, this is my first post since I applied for the server in 2014 lol. I love you guys for still being so active after all these years.

Maybe a limited amount of space to build in or a 4 bit ISA. Also I think to decide the winner we need some sort of benchmark system where ur score in each category (speed, CPI, mem access time etc) will give u points and then all ur points r added up therefore cpu with most benchmark points wins. Or if we do a cpu with space restrictions we could make the incentive to where the winning cpu would go on pol.

We could also maybe do single cycle CPUs that might make it easier to score CPUs cuz with pipelining we might see repetitive designs that r mostly the same

Also with scoring we need to have a good way to score peoples ISAs. If we’re going to test CPUs by running programs in them I suggest running more complex programs so people’s ISAs can really show how they affect performance. If we do something simple like fibs any basic cpu can do that so we won’t really take into affect the instruction architecture aspect of CPUs when scoring

I'd like to see a requirement for some basic multiprocessing/scaling capabilities. Something that I've always waned to mess with more

I'm still up for a cpu that can do the ackermann function: https://en.wikipedia.org/wiki/Ackermann_function. It is one area ORE never really got into.

1 operand ISA