Skip to main content
Version: v0.4.2

gbz80(7) — CPU opcode reference

This is the list of opcodes supported by rgbasm(1), including a short description, the number of bytes needed to encode them and the number of CPU cycles at 1MHz (or 2MHz in GBC dual speed mode) needed to complete them.

Note: All arithmetic/logic operations that use register A as destination can omit the destination as it is assumed to be register A by default. The following two lines have the same effect:

OR A,B
OR B

List of abbreviations used in this document.

r8
Any of the 8-bit registers (A, , C, , , H, ).
r16
Any of the general-purpose 16-bit registers (, , HL).
n8
8-bit integer constant.
n16
16-bit integer constant.
e8
8-bit offset ( to ).
u3
3-bit unsigned integer constant ( to ).
cc
Condition codes:
Execute if Z is set.
Execute if Z is not set.
Execute if C is set.
Execute if C is not set.
vec
One of the vectors (, , , , , , and ).

ADC A,r8
ADC A,[HL]
ADC A,n8
ADD A,r8
ADD A,[HL]
ADD A,n8
AND A,r8
AND A,[HL]
AND A,n8
CP A,r8
CP A,[HL]
CP A,n8
DEC r8
DEC [HL]
INC r8
INC [HL]
OR A,r8
OR A,[HL]
OR A,n8
SBC A,r8
SBC A,[HL]
SBC A,n8
SUB A,r8
SUB A,[HL]
SUB A,n8
XOR A,r8
XOR A,[HL]
XOR A,n8

ADD HL,r16
DEC r16
INC r16

BIT u3,r8
BIT u3,[HL]
RES u3,r8
RES u3,[HL]
SET u3,r8
SET u3,[HL]
SWAP r8
SWAP [HL]

RL r8
RL [HL]
RLA
RLC r8
RLC [HL]
RLCA
RR r8
RR [HL]
RRA
RRC r8
RRC [HL]
RRCA
SLA r8
SLA [HL]
SRA r8
SRA [HL]
SRL r8
SRL [HL]

LD r8,r8
LD r8,n8
LD r16,n16
LD [HL],r8
LD [HL],n8
LD r8,[HL]
LD [r16],A
LD [n16],A
LDH [n16],A
LDH [C],A
LD A,[r16]
LD A,[n16]
LDH A,[n16]
LDH A,[C]
LD [HLI],A
LD [HLD],A
LD A,[HLI]
LD A,[HLD]

CALL n16
CALL cc,n16
JP HL
JP n16
JP cc,n16
JR e8
JR cc,e8
RET cc
RET
RETI
RST vec

ADD HL,SP
ADD SP,e8
DEC SP
INC SP
LD SP,n16
LD [n16],SP
LD HL,SP+e8
LD SP,HL
POP AF
POP r16
PUSH AF
PUSH r16

CCF
CPL
DAA
DI
EI
HALT
NOP
SCF
STOP

Add the value in r8 plus the carry flag to A.

Cycles: 1

Bytes: 1

Flags:

Set if result is 0.
0
Set if overflow from bit 3.
Set if overflow from bit 7.

Add the byte pointed to by HL plus the carry flag to A.

Cycles: 2

Bytes: 1

Flags: See ADC A,r8

Add the value n8 plus the carry flag to A.

Cycles: 2

Bytes: 2

Flags: See ADC A,r8

Add the value in r8 to A.

Cycles: 1

Bytes: 1

Flags:

Set if result is 0.
0
Set if overflow from bit 3.
Set if overflow from bit 7.

Add the byte pointed to by HL to A.

Cycles: 2

Bytes: 1

Flags: See ADD A,r8

Add the value n8 to A.

Cycles: 2

Bytes: 2

Flags: See ADD A,r8

Add the value in r16 to HL.

Cycles: 2

Bytes: 1

Flags:

0
Set if overflow from bit 11.
Set if overflow from bit 15.

Add the value in SP to HL.

Cycles: 2

Bytes: 1

Flags: See ADD HL,r16

Add the signed value e8 to SP.

Cycles: 4

Bytes: 2

Flags:

0
0
Set if overflow from bit 3.
Set if overflow from bit 7.

Bitwise AND between the value in r8 and A.

Cycles: 1

Bytes: 1

Flags:

Set if result is 0.
0
1
0

Bitwise AND between the byte pointed to by HL and A.

Cycles: 2

Bytes: 1

Flags: See AND A,r8

Bitwise AND between the value in n8 and A.

Cycles: 2

Bytes: 2

Flags: See AND A,r8

Test bit u3 in register r8, set the zero flag if bit not set.

Cycles: 2

Bytes: 2

Flags:

Set if the selected bit is 0.
0
1

Test bit u3 in the byte pointed by HL, set the zero flag if bit not set.

Cycles: 3

Bytes: 2

Flags: See BIT u3,r8

Call address n16. This pushes the address of the instruction after the on the stack, such that RET can pop it later; then, it executes an implicit JP n16.

Cycles: 6

Bytes: 3

Flags: None affected.

Call address n16 if condition cc is met.

Cycles: 6 taken / 3 untaken

Bytes: 3

Flags: None affected.

Complement Carry Flag.

Cycles: 1

Bytes: 1

Flags:

0
0
Inverted.

Subtract the value in r8 from A and set flags accordingly, but don't store the result. This is useful for ComParing values.

Cycles: 1

Bytes: 1

Flags:

Set if result is 0.
1
Set if borrow from bit 4.
Set if borrow (i.e. if r8 > A).

Subtract the byte pointed to by HL from A and set flags accordingly, but don't store the result.

Cycles: 2

Bytes: 1

Flags: See CP A,r8

Subtract the value n8 from A and set flags accordingly, but don't store the result.

Cycles: 2

Bytes: 2

Flags: See CP A,r8

ComPLement accumulator (A = ).

Cycles: 1

Bytes: 1

Flags:

1
1

Decimal Adjust Accumulator to get a correct BCD representation after an arithmetic instruction.

Cycles: 1

Bytes: 1

Flags:

Set if result is 0.
0
Set or reset depending on the operation.

Decrement value in register r8 by 1.

Cycles: 1

Bytes: 1

Flags:

Set if result is 0.
1
Set if borrow from bit 4.

Decrement the byte pointed to by HL by 1.

Cycles: 3

Bytes: 1

Flags: See DEC r8

Decrement value in register r16 by 1.

Cycles: 2

Bytes: 1

Flags: None affected.

Decrement value in register SP by 1.

Cycles: 2

Bytes: 1

Flags: None affected.

Disable Interrupts by clearing the IME flag.

Cycles: 1

Bytes: 1

Flags: None affected.

Enable Interrupts by setting the IME flag. The flag is only set after the instruction following EI.

Cycles: 1

Bytes: 1

Flags: None affected.

Enter CPU low-power consumption mode until an interrupt occurs. The exact behavior of this instruction depends on the state of the IME flag.

set
The CPU enters low-power mode until after an interrupt is about to be serviced. The handler is executed normally, and the CPU resumes execution after the HALT when that returns.
not set
The behavior depends on whether an interrupt is pending (i.e. ‘[IE] & [IF]’ is non-zero).
None pending
As soon as an interrupt becomes pending, the CPU resumes execution. This is like the above, except that the handler is called.
Some pending
The CPU continues execution after the HALT, but the byte after it is read twice in a row (PC is not incremented, due to a hardware bug).

Cycles: -

Bytes: 1

Flags: None affected.

Increment value in register r8 by 1.

Cycles: 1

Bytes: 1

Flags:

Set if result is 0.
0
Set if overflow from bit 3.

Increment the byte pointed to by HL by 1.

Cycles: 3

Bytes: 1

Flags: See INC r8

Increment value in register r16 by 1.

Cycles: 2

Bytes: 1

Flags: None affected.

Increment value in register SP by 1.

Cycles: 2

Bytes: 1

Flags: None affected.

Jump to address n16; effectively, store n16 into PC.

Cycles: 4

Bytes: 3

Flags: None affected.

Jump to address n16 if condition cc is met.

Cycles: 4 taken / 3 untaken

Bytes: 3

Flags: None affected.

Jump to address in HL; effectively, load PC with value in register HL.

Cycles: 1

Bytes: 1

Flags: None affected.

Relative Jump by adding e8 to the address of the instruction following the . To clarify, an operand of 0 is equivalent to no jumping.

Cycles: 3

Bytes: 2

Flags: None affected.

Relative Jump by adding e8 to the current address if condition cc is met.

Cycles: 3 taken / 2 untaken

Bytes: 2

Flags: None affected.

Load (copy) value in register on the right into register on the left.

Cycles: 1

Bytes: 1

Flags: None affected.

Load value n8 into register r8.

Cycles: 2

Bytes: 2

Flags: None affected.

Load value n16 into register r16.

Cycles: 3

Bytes: 3

Flags: None affected.

Store value in register r8 into byte pointed to by register HL.

Cycles: 2

Bytes: 1

Flags: None affected.

Store value n8 into byte pointed to by register HL.

Cycles: 3

Bytes: 2

Flags: None affected.

Load value into register r8 from byte pointed to by register HL.

Cycles: 2

Bytes: 1

Flags: None affected.

Store value in register A into byte pointed to by register r16.

Cycles: 2

Bytes: 1

Flags: None affected.

Store value in register A into byte at address n16.

Cycles: 4

Bytes: 3

Flags: None affected.

Store value in register A into byte at address n16, provided it is between $FF00 and $FFFF.

Cycles: 3

Bytes: 2

Flags: None affected.

This is sometimes written as ‘ldio [n16], a’, or ‘ld [$ff00+n8], a’.

Store value in register A into byte at address $FF00+C.

Cycles: 2

Bytes: 1

Flags: None affected.

This is sometimes written as ‘ldio [c], a’, or ‘ld [$ff00+c], a’.

Load value in register A from byte pointed to by register r16.

Cycles: 2

Bytes: 1

Flags: None affected.

Load value in register A from byte at address n16.

Cycles: 4

Bytes: 3

Flags: None affected.

Load value in register A from byte at address n16, provided it is between $FF00 and $FFFF.

Cycles: 3

Bytes: 2

Flags: None affected.

This is sometimes written as ‘ldio a, [n16]’, or ‘ld a, [$ff00+n8]’.

Load value in register A from byte at address $FF00+c.

Cycles: 2

Bytes: 1

Flags: None affected.

This is sometimes written as ‘ldio a, [c]’, or ‘ld a, [$ff00+c]’.

Store value in register A into byte pointed by HL and increment HL afterwards.

Cycles: 2

Bytes: 1

Flags: None affected.

Store value in register A into byte pointed by HL and decrement HL afterwards.

Cycles: 2

Bytes: 1

Flags: None affected.

Load value into register A from byte pointed by HL and decrement HL afterwards.

Cycles: 2

Bytes: 1

Flags: None affected.

Load value into register A from byte pointed by HL and increment HL afterwards.

Cycles: 2

Bytes: 1

Flags: None affected.

Load value n16 into register SP.

Cycles: 3

Bytes: 3

Flags: None affected.

Store SP & $FF at address n16 and SP >> 8 at address n16 + 1.

Cycles: 5

Bytes: 3

Flags: None affected.

Add the signed value e8 to SP and store the result in HL.

Cycles: 3

Bytes: 2

Flags:

0
0
Set if overflow from bit 3.
Set if overflow from bit 7.

Load register HL into register SP.

Cycles: 2

Bytes: 1

Flags: None affected.

No OPeration.

Cycles: 1

Bytes: 1

Flags: None affected.

Store into A the bitwise OR of the value in r8 and A.

Cycles: 1

Bytes: 1

Flags:

Set if result is 0.
0
0
0

Store into A the bitwise OR of the byte pointed to by HL and A.

Cycles: 2

Bytes: 1

Flags: See OR A,r8

Store into A the bitwise OR of n8 and A.

Cycles: 2

Bytes: 2

Flags: See OR A,r8

Pop register AF from the stack. This is roughly equivalent to the following imaginary instructions:

ld f, [sp] ; See below for individual flags
inc sp
ld a, [sp]
inc sp

Cycles: 3

Bytes: 1

Flags:

Set from bit 7 of the popped low byte.
Set from bit 6 of the popped low byte.
Set from bit 5 of the popped low byte.
Set from bit 4 of the popped low byte.

Pop register r16 from the stack. This is roughly equivalent to the following imaginary instructions:

ld LOW(r16), [sp] ; C, E or L
inc sp
ld HIGH(r16), [sp] ; B, D or H
inc sp

Cycles: 3

Bytes: 1

Flags: None affected.

Push register AF into the stack. This is roughly equivalent to the following imaginary instructions:

dec sp
ld [sp], a
dec sp
ld [sp], flag_Z << 7 | flag_N << 6 | flag_H << 5 | flag_C << 4

Cycles: 4

Bytes: 1

Flags: None affected.

Push register r16 into the stack. This is roughly equivalent to the following imaginary instructions:

dec sp
ld [sp], HIGH(r16) ; B, D or H
dec sp
ld [sp], LOW(r16) ; C, E or L

Cycles: 4

Bytes: 1

Flags: None affected.

Set bit u3 in register r8 to 0. Bit 0 is the rightmost one, bit 7 the leftmost one.

Cycles: 2

Bytes: 2

Flags: None affected.

Set bit u3 in the byte pointed by HL to 0. Bit 0 is the rightmost one, bit 7 the leftmost one.

Cycles: 4

Bytes: 2

Flags: None affected.

Return from subroutine. This is basically a POP PC (if such an instruction existed). See POP r16 for an explanation of how POP works.

Cycles: 4

Bytes: 1

Flags: None affected.

Return from subroutine if condition cc is met.

Cycles: 5 taken / 2 untaken

Bytes: 1

Flags: None affected.

Return from subroutine and enable interrupts. This is basically equivalent to executing EI then RET, meaning that IME is set right after this instruction.

Cycles: 4

Bytes: 1

Flags: None affected.

Rotate bits in register r8 left through carry.

C <- [7 <- 0] <- C

Cycles: 2

Bytes: 2

Flags:

Set if result is 0.
0
0
Set according to result.

Rotate byte pointed to by HL left through carry.

C <- [7 <- 0] <- C

Cycles: 4

Bytes: 2

Flags: See RL r8

Rotate register A left through carry.

C <- [7 <- 0] <- C

Cycles: 1

Bytes: 1

Flags:

0
0
0
Set according to result.

Rotate register r8 left.

C <- [7 <- 0] <- [7]

Cycles: 2

Bytes: 2

Flags:

Set if result is 0.
0
0
Set according to result.

Rotate byte pointed to by HL left.

C <- [7 <- 0] <- [7]

Cycles: 4

Bytes: 2

Flags: See RLC r8

Rotate register A left.

C <- [7 <- 0] <- [7]

Cycles: 1

Bytes: 1

Flags:

0
0
0
Set according to result.

Rotate register r8 right through carry.

C -> [7 -> 0] -> C

Cycles: 2

Bytes: 2

Flags:

Set if result is 0.
0
0
Set according to result.

Rotate byte pointed to by HL right through carry.

C -> [7 -> 0] -> C

Cycles: 4

Bytes: 2

Flags: See RR r8

Rotate register A right through carry.

C -> [7 -> 0] -> C

Cycles: 1

Bytes: 1

Flags:

0
0
0
Set according to result.

Rotate register r8 right.

[0] -> [7 -> 0] -> C

Cycles: 2

Bytes: 2

Flags:

Set if result is 0.
0
0
Set according to result.

Rotate byte pointed to by HL right.

[0] -> [7 -> 0] -> C

Cycles: 4

Bytes: 2

Flags: See RRC r8

Rotate register A right.

[0] -> [7 -> 0] -> C

Cycles: 1

Bytes: 1

Flags:

0
0
0
Set according to result.

Call address vec. This is a shorter and faster equivalent to CALL for suitable values of vec.

Cycles: 4

Bytes: 1

Flags: None affected.

Subtract the value in r8 and the carry flag from A.

Cycles: 1

Bytes: 1

Flags:

Set if result is 0.
1
Set if borrow from bit 4.
Set if borrow (i.e. if (r8 + carry) > A).

Subtract the byte pointed to by HL and the carry flag from A.

Cycles: 2

Bytes: 1

Flags: See SBC A,r8

Subtract the value n8 and the carry flag from A.

Cycles: 2

Bytes: 2

Flags: See SBC A,r8

Set Carry Flag.

Cycles: 1

Bytes: 1

Flags:

0
0
1

Set bit u3 in register r8 to 1. Bit 0 is the rightmost one, bit 7 the leftmost one.

Cycles: 2

Bytes: 2

Flags: None affected.

Set bit u3 in the byte pointed by HL to 1. Bit 0 is the rightmost one, bit 7 the leftmost one.

Cycles: 4

Bytes: 2

Flags: None affected.

Shift Left Arithmetic register r8.

C <- [7 <- 0] <- 0

Cycles: 2

Bytes: 2

Flags:

Set if result is 0.
0
0
Set according to result.

Shift Left Arithmetic byte pointed to by HL.

C <- [7 <- 0] <- 0

Cycles: 4

Bytes: 2

Flags: See SLA r8

Shift Right Arithmetic register r8.

[7] -> [7 -> 0] -> C

Cycles: 2

Bytes: 2

Flags:

Set if result is 0.
0
0
Set according to result.

Shift Right Arithmetic byte pointed to by HL.

[7] -> [7 -> 0] -> C

Cycles: 4

Bytes: 2

Flags: See SRA r8

Shift Right Logic register r8.

0 -> [7 -> 0] -> C

Cycles: 2

Bytes: 2

Flags:

Set if result is 0.
0
0
Set according to result.

Shift Right Logic byte pointed to by HL.

0 -> [7 -> 0] -> C

Cycles: 4

Bytes: 2

Flags: See SRA r8

Enter CPU very low power mode. Also used to switch between double and normal speed CPU modes in GBC.

Cycles: -

Bytes: 2

Flags: None affected.

Subtract the value in r8 from A.

Cycles: 1

Bytes: 1

Flags:

Set if result is 0.
1
Set if borrow from bit 4.
Set if borrow (set if r8 > A).

Subtract the byte pointed to by HL from A.

Cycles: 2

Bytes: 1

Flags: See SUB A,r8

Subtract the value n8 from A.

Cycles: 2

Bytes: 2

Flags: See SUB A,r8

Swap upper 4 bits in register r8 and the lower 4 ones.

Cycles: 2

Bytes: 2

Flags:

Set if result is 0.
0
0
0

Swap upper 4 bits in the byte pointed by HL and the lower 4 ones.

Cycles: 4

Bytes: 2

Flags: See SWAP r8

Bitwise XOR between the value in r8 and A.

Cycles: 1

Bytes: 1

Flags:

Set if result is 0.
0
0
0

Bitwise XOR between the byte pointed to by HL and A.

Cycles: 2

Bytes: 1

Flags: See XOR A,r8

Bitwise XOR between the value in n8 and A.

Cycles: 2

Bytes: 2

Flags: See XOR A,r8

rgbasm(1), rgbds(7)

rgbds was originally written by Carsten Sørensen as part of the ASMotor package, and was later packaged in RGBDS by Justin Lloyd. It is now maintained by a number of contributors at https://github.com/gbdev/rgbds.