PASM Interpreter
hippy
Posts: 1,981
Before re-inventing the wheel; has anyone already written a PASM interpreter in PASM ideally or Spin ?
This would be an interpreter which could execute PASM as is, supporting self modifying code, arbitrary access to all 512 registers / cog locations, standard PASM call / return handling. I'd expect the PASM to be interpreted to be held in 496 longs of hub memory.
This is similar to but quite an extension of LMM interpretation. It would consequently be a lot slower but its advantage is that PASM doesn't need to be changed to LMM-style code to work.
Ultimately the PASM interpreter would be coded as LMM and interpreted, but as real PASM running an LMM Interpreter running a PASM interpreter, potentially running a Spin interpreter become such a 'mind-blow' it's probably best not to think about any of that
This would be an interpreter which could execute PASM as is, supporting self modifying code, arbitrary access to all 512 registers / cog locations, standard PASM call / return handling. I'd expect the PASM to be interpreted to be held in 496 longs of hub memory.
This is similar to but quite an extension of LMM interpretation. It would consequently be a lot slower but its advantage is that PASM doesn't need to be changed to LMM-style code to work.
Ultimately the PASM interpreter would be coded as LMM and interpreted, but as real PASM running an LMM Interpreter running a PASM interpreter, potentially running a Spin interpreter become such a 'mind-blow' it's probably best not to think about any of that
Comments
I think we are now missing someone to write a spin complier so we can put all the pieces together. If someone targets Basic and C, etc to the spin interpreter, I think we could have a great system, given the other work being done on SD drivers, etc.
Mindblowing
Postedited: I already have a short, fast, variable length·overlay loader which hits the sweet spot (16 clocks).
Post Edited (Cluso99) : 7/29/2008 4:58:07 AM GMT