Skip to content

Latest commit

 

History

History
71 lines (55 loc) · 2.72 KB

README.md

File metadata and controls

71 lines (55 loc) · 2.72 KB
Raw

Simulador SISA

The code is incredibly self documenting (on purpose). Usage is below.

Options and flags can be shown by running the -h (summary) or --help (full) flags.

Usage

To solve the alien that I had for 13c (which gave instuctions and initial memory, with an initial PC of 0 and no initial register state), use (Note the use of --simple):

cargo run -- --simple examples/alien/alien.sisa -m examples/alien/alien.smem -i examples/alien/alien.sio

To solve the alien that I had for Tema 14 (Note the lack of --simple):

cargo run -- examples/complete/first_example.sisa

You may also specify some initial values for the registers and IO, regardless of if the input is --simple or not:

cargo run -- [code file] -r [registers file] -i [io file]

IO

Use [cpu].update_io(new_io) to change the IO status in between executes. This cannot be done from the cli, at the moment, because I have no idea how (and it doesn't seem that useful, to be honest).

Registers

File must contain eight lines (or less, for a computer with less registers), each with a decimal number. E.g.

0
0
0
0
0
0
5
17

Register 6 will hold 0x0005 and Register 7 will hold 0x0011.

You cannot use more than eight registers, it will abort (this is intended behavior).

Ideas

Notes

  • .text and .data are not followed by :, only labels are
  • Don't .set words like verbs, register names or labels. That's UB.
  • Don't mismatch parens. That's UB, though I try to catch them
  • Jumps are always relative, so JALR may behave... unexpectedly (wrong). Careful!
  • Don't try to make jumps too large. There's a check for this but it's untested
  • Don't feed it malformed input like MOVI R5, 0x555.If you feed it invalid input, you're gonna get UB (good luck lmao), though I do try to catch them when I can.
  • The memory and data memories are separate because I didn't stop to think before I started writing down code. Just, like, assume they're the same. I'm protecting you from yourself (this actually has already helped me from programmer error lmao).
  • The LD/ST instructions have offsets in their command: these are always in number of bytes, not words (changing it would be easy, the source code is actually very readable and intuitive).
  • If the program were to not halt, the output would grow VERY large, so there's a built in check: if the number of instructions ran exceeds 10000 (MAX_INSTRUCTION_RUN_SIZE), it halts execution.

License

Licensed under the GPLv3 (see COPYING file for full text).