NAND2Tetris Note 05 - Hack Assembly Language

The “Hack” computer built during this course has its own instruction set consisting of two instructions: the A-command and the C-command.

Registers

A -> address. Used to access memory D -> General purpose data register M -> Psuedo register that actually accesses RAM[A]

A-instruction

Denoted as @<a> where a is a positive number, this instruction loads the specified literal into the A (or address) register. This is encoded as:

0 <value (15 bits)>

C-instruction

The c-command has the format: dest = comp; jmp, where dest is the destination (if any), comp is the computation to be performed and jmp is a jumpy instruction. This is encoded as the following in a 16-bit instruction:

111 a c1 c2 c3 c4 c5 c6 d1 d2 d3 j1 j2 j3 ^ denotes “C”

a through c6 defines the computation to be performed as inputs to the ALU. There is a table describing every operation allowed here. d1 d2 d3 defines the destination of the computation (or 0 0 0 for null), j1 j2 j3 reperesents a jump instruction (conditional, uncondtional, null etc.)

I/O

I/O is performed by writing/reading from specific memory addresses. Display on the screen is accomplished by writing to 8K 16-bit registers starting at address 16384. Each bit represents a pixel in a 256x512 matrix stored in row-major form.

For pixel at (row, col): 32*row + col/16 -> word index col % 16 -> bit index within word

The keyboard value is represented by a single 16-bit value at the address 24576.

Assembly language

R1-R15 are built-in symbols for addresses from 1-15. This helps remove confusion between literals and addresses to some extent when using the A instruction. SCREEN and KBD and I/O base addresses.

Labels

Labels can be specified as (LABEL). This can be used as jump destinations. Example:

// Infinite loop
(END)
@END
0;JMP

Variables

@

Automatically uses an available memory address as long as there is no corresponding label. Allocated from address 16 onward.