6502 Instructions - Introduction
It is still under construction and content may change. Do not rely on this information.
The 6502 processor has a compact instruction set, consisting of just 56 instructions:
ADC AND ASL BCC BCS BEQ BIT BMI BNE BPL BRK BVC BVS CLC
CLD CLI CLV CMP CPX CPY DEC DEX DEY EOR INC INX INY JMP
JSR LDA LDX LDY LSR NOP ORA PHA PHP PLA PLP ROL ROR RTI
RTS SBC SEC SED SEI STA STX STY TAX TAY TSX TXA TXS TYA
This page groups these instructions and explains their basic function.
Contents
Addressing Modes
There are 13 6502 Addressing Modes. All of these instructions work with at least one addressing mode, and many work with several addressing modes. See the Resources section for Opcode tables that define which instructions work with which addressing modes.
Registers
Most of these instructions work with registers. Refer to the 6502 page, Register section for details on the 6502's internal registers.
Performance
Each 6502 instruction takes a defined number of machine cycles to execute. In some cases,the number of cycles will vary depending on the circumstances of the instruction's execution - for example, the conditional branch instruction BRE (Branch if EQual) takes:
- 2 cycles if the branch is not taken
- 3 cycles if a branch is taken to an address in the same page
- 4 cycles if a branch is taken to an address in another page
Remember that the Program Counter (PC register) contains a pointer to the next instruction to be executed. When the BEQ instruction has been loaded into the CPU, the PC points to the instruction following the BEQ. The branch works by adding a signed integer value (in the range of -128 to +127) to the Program Counter; the extra cycle required when the branch is taken is used to process the addition. If the high byte of the Program Counter changes (because the branch crosses in to another page), one additional cycle is required to adjust the high byte.
You can find the execution time in the instruction charts found in the Resources section below.
To convert the number of cycles to time, multiply the cycles by the time between system clock pulses. Many 6502 systems operated at 1 MHz (1 million operations per second), and therefore 1 cycle corresponded to 1 millionth of a second, or 1 microsecond (uS). Therefore, an instruction that took 4 clock cycles would take 4 uS to execute.
Loading and Storing Data (to/from Memory)
Register-Memory Loads and Stores
There are three instructions to load data from memory to a register:
LDA ; load the accumulator LDX ; load the X register LDY ; load the Y register
And there are three matching instructions to store data from a register to a memory location:
STA ; store the accumulator STX ; store the X register STY ; store the Y register
Push/Pull on the Stack
When a value is pushed to the stack, the stack pointer register (SP) is decremented and the selected register is written to memory location $0100+SP.
When a value is pulled from the stack, the stack pointer register (SP) is incremented and the selected register is loaded from memory location $0100+SP.
There are two instructions to push data onto the stack:
PHA ; push the accumulator PHP ; push the processor status register (SR)
And two matching instructions to pull data from the stack:
PLA ; pull the accumulator PLP ; pull the processor status register (SR)
Note that some other operations, such as JSR, interrupts, RTI, and RTS, cause data to be pushed to or pulled from the stack.
Transferring Data between Registers
The X and Y registers can be transferred to/from the accumulator:
TAX ; transfer A to X TAY ; transfer A to Y TXA ; transfer X to A TYA ; transfer Y to A
You can also transfer the Stack Pointer (SP) to/from the X register:
TSX ; transfer SP to X TXS ; tranfer X to SP
It is not possible to directly transfer the Status Register (SR) to a general-purpose register, but you can you transfer it via the stack (e.g., by pushing SR to the stack with with PHP and then popping the stack to the accumulator with PLA).
Arithmetic and Bitwise Operations
The 6502 has rudimentry addition and subtraction instructions, which operate on the accumulator (A):
ADC ; add with carry SBC ; subtract with carry
There are also increment and decrement instructions for the X and Y registers and for memory:
DEC ; decrement memory DEX ; decrement X register DEY ; decrement Y register INC ; increment memory INX ; increment X register INY ; increment Y register
The 6502 also has instructions for left and right bit-shifts and rotations (which can act as multiply-by-2 and divide-by-2):
ASL ; arithmetic shift left ROL ; rotate left LSR ; logical shift right ROR ; rotate right
There are also instructions for bitwise operations such as exclusive-OR, OR, and AND. Exclusive-OR with #$FF is equivalent to a NOT operation, and these operations can be combined to produce other logical operations such as NOR and NAND.
AND ; bitwise AND (with accumulator) EOR ; bitwise exclusive-OR (with accumulator) ORA ; bitwise OR (with accumulator)
For full details on all of the arithmetic and bitwise instructions, see the 6502 Math page.
Test and Comparison Operations
The A, X, and Y registers can be directly compared with immediate or memory values:
CMP ; compare (accumulator) CPX ; compare (X register) CPY ; compare (Y register)
These operations are performed by subtraction. The appropriate condition flags are set, and the result of the subtraction is discarded. Conditional branch instructions can be used to alter program flow based on the results of the comparisons.
There is another test instruction available:
BIT ; bit test
This instruction places bit 7 of the operand into the N flag and bit 6 of the operand into the V flag. The operand is then ANDed with the accumulator, and the Z flag is set if the result is zero. The result of the AND is discarded. In this way, you can test the value of bit 7, bit 6, or any arbitrary bits (using the operand).
Note that in addition to these instructions, many other instructions (such as register loads) affect condition flags.
Failing to clear the carry flag before addition or to set the carry flag before subtraction is the cause of many bugs in 6502 programs. The carry flag also affects the rotate instructions. Be sure to set or clear this flag with the
SEC or CLC instructions when needed!Program Flow
Unconditional Jump
An unconditional jump is like a "Goto" -- it sets the address of the next instruction to be executed (by modifying the Program Counter (PC)):
JMP ; jump to address
Jump to SubRoutine
A jump to a subroutine is also unconditional, but the current value of the Program Counter (PC) is placed on the stack so that when the subroutine (aka procedure, function, or method) is finished, execution can resume at the instruction after the jump to subroutine:
JSR ; jump to subroutine (pushes PC on stack, loads operand into PC) RTS ; return from subroutine (pops PC from stack)
Conditional Branch
A conditional branch is like a jump, except that it is only performed if a certain condition is met:
BCC ; branch on carry clear (C flag is clear) BCS ; branch on carry set (C flag is set) BEQ ; branch if equal (Z flag is set) BMI ; branch if minus (N flag is set) BNE ; branch if not equal (Z flag is clear) BPL ; branch if plus (N flag is clear) BVC ; branch if overflow clear (V flag is clear) BVS ; branch if overflow set (V flag is set)
Note that the operand for conditional branch instructions is a relative offset - a signed 8-bit value (in the range -128 to +127) that is added to the current PC. When writing assembler (or viewing disassembled code), the operand is written as an absolute address or label, but the actual machine language code uses the relative addressing mode. For this reason, a branch that is too far will not assemble and will produce an error message.
