/bytecode_interpreter

Primary LanguageC++

My attempt at writing a bytecode interpreter

How the instructions are structured

  • Each instruction is 64-bits
  • The rightmost 8 bits hold values for different flags
  • The next 8 bits from the right hold information about which register to use
  • The next 16 bits from the right hold information about which operation to execute
  • The last 32 bits hold information about the data.

Registers

  • It has 8 general purpose (32-bit) named R1..R8.
  • It has 8 floating-point registers (32-bit) named FPR1..FPR8.

Structure of the assembly code

OPERATION DEST, SOURCE. The accepted data are integer literals and floating-point literals and other registers. For accessing the data stored in a register: OPERATION REGISTER, $REGISTER.

CPU Flags

As mentioned above, the 8 rightmost bits are used for CPU Flags.

FROM RIGHT TO LEFT:

  1. REGISTER_FLAG - Tells the CPU if the data is a literal or if it should retrieve data from a register.
  2. FLOAT_FLAG - Tells the CPU if the current instruction consists of floating-point data and operations.

Commands supported currently

  • Commands for general purpose registers.

    • MOV - Moves a value into a register.
    • ADD - Adds a value to a register.
    • SUB - Subtracts a value from a register.
    • MUL - Multiplies a value to a register.
    • DIV - Divides a value by a register.
    • RPRINT - Prints the register's name and its value.

  • Commands for floating-point registers.

    • FMOV - Moves a floating-point value into a register.
    • FADD - Adds a floating-point value to a register.
    • FSUB - Subtracts a floating-point value from a register.
    • FMUL - Multiplies a floating-point value to a register.
    • FDIV - Divides a floating-point value by a register.
    • FPRINT - Prints the floating-point register's name and its value. Optionally takes an integer value for precision.