I've kept going just because i feel the cpu's that are already made aren't really that good at pointers, IO based interrupts, branching, address space, speed; anything that just needs to be there if you want to implement a simple file system/OS. A main goal of mine to let others code on my cpu from a keyboard & screen. Something that hasn't yet been done.
- I try to implement units that make sense on a modern cpu; caching, btb's, experimenting on out of order execution, routers. Even if i dont have a use for them right now.
- I mostly keep several side projects so i don't bore myself to death or get stuck for a while.
- I do not want to waste time on compacting/speeding up components beyond a certain point. There are too many ALU's and too little complete systems.
- rework IS -> rework layout -> rework IS -> rework layout. I don't start building if its too complex and full of uncertanties. No point starting a huge project and seeing it go to waste.
- I try to implement units that make sense on a modern cpu; caching, btb's, experimenting on out of order execution, routers. Even if i dont have a use for them right now.
- I mostly keep several side projects so i don't bore myself to death or get stuck for a while.
- I do not want to waste time on compacting/speeding up components beyond a certain point. There are too many ALU's and too little complete systems.
- rework IS -> rework layout -> rework IS -> rework layout. I don't start building if its too complex and full of uncertanties. No point starting a huge project and seeing it go to waste.
Don't thank me, thank my coffee machine