CPU Registers and Flags

From GbdevWiki
Revision as of 05:55, 28 June 2009 by Xzakox (Talk | contribs) (New page: ===Registers=== 16bit Hi Lo Name/Function AF A - Accumulator & Flags BC B C BC DE D E DE HL H L HL SP - - Stack Pointer PC ...)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Registers

 16bit Hi   Lo   Name/Function
 AF    A    -    Accumulator & Flags
 BC    B    C    BC
 DE    D    E    DE
 HL    H    L    HL
 SP    -    -    Stack Pointer
 PC    -    -    Program Counter/Pointer

As shown above, most registers can be accessed either as one 16bit register, or as two separate 8bit registers.

The Flag Register (lower 8bit of AF register)

 Bit  Name  Set Clr  Expl.
 7    zf    Z   NZ   Zero Flag
 6    n     -   -    Add/Sub-Flag (BCD)
 5    h     -   -    Half Carry Flag (BCD)
 4    cy    C   NC   Carry Flag
 3-0  -     -   -    Not used (always zero)

Conatins the result from the recent instruction which has affected flags.

The Zero Flag (Z)

This bit becomes set (1) if the result of an operation has been zero (0). Used for conditional jumps.

The Carry Flag (C, or Cy)

Becomes set when the result of an addition became bigger than FFh (8bit) or FFFFh (16bit). Or when the result of a subtraction or comparision became less than zero (much as for Z80 and 80x86 CPUs, but unlike as for 65XX and ARM CPUs). Also the flag becomes set when a rotate/shift operation has shifted-out a "1"-bit. Used for conditional jumps, and for instructions such like ADC, SBC, RL, RLA, etc.

The BCD Flags (N, H)

These flags are (rarely) used for the DAA instruction only, N Indicates whether the previous instruction has been an addition or subtraction, and H indicates carry for lower 4bits of the result, also for DAA, the C flag must indicate carry for upper 8bits. After adding/subtracting two BCD numbers, DAA is intended to convert the result into BCD format; BCD numbers are ranged from 00h to 99h rather than 00h to FFh. Because C and H flags must contain carry-outs for each digit, DAA cannot be used for 16bit operations (which have 4 digits), or for INC/DEC operations (which do not affect C-flag).