COMS10015: Hex 8 ISA assembler

This is a simple assembler to convert Hex 8 assembly programs into machine code, suitable to be loaded into the ModuleSim NRAM.

Building

The code should build just by running:

make 

The code uses C++14, and so g++ version 5 or higher is required.

Running

The binary takes a single argument: a *.hex8 file with Hex 8 assembly instructions. The machine code is output to a file called a.hex in the current working directory.

Input file format

The *.hex8 input file should be a valid Hex 8 program.

The assembler performs some basic error checking, but this is not exhaustive.

Operands

There are 16 Hex 8 instruction mnemonics which can be used. In the assembly program, the opcode mnumonic is followed by a single space and then an opcode represented in decimal (base 10).

Negative numbers are allowed. The range of valid operands is therefore [-127, 255] inclusive.

The assembler will insert appropriate PFIX instructions for negative or large operands.

Labels

Labels are supported. Labels are of the form L123: the capital letter L followed by a number. They must be on their own line.

Labels can be used as operands.

For example:

L1
ADD
BR L1

The assembler will always insert a PFIX instruction before any instruction using a label. A better assembler would remove ones that were not needed with additional passes.

Data statements

Support is also included for the DATA mnemonic.

For example:

DATA 42
DATA L1

The assembler simply outputs the operand into the instruction stream. For large absolute numbers and all labels, the assembler outputs PFIX instructions.

Labels can be applied to DATA statements so they can be used as memory addresses. For example:

L1 
DATA 0
STAM L1