User Tools

Site Tools


base:practical_memory_move_routines
no way to compare when less than two revisions

Differences

This shows you the differences between two versions of the page.


base:practical_memory_move_routines [2015-04-17 04:33] (current) – created - external edit 127.0.0.1
Line 1: Line 1:
 +====== Practical Memory Move Routines ======
 +by Bruce Clark
 +
 +(Info file taken from [[http://www.6502.org|www.6502.org]])
 +
 +Here are some reasonably fast general-purpose routines for moving blocks of memory. You simply specify the address to move from, the address to move to, and the size of the block. When SIZE is zero, no bytes are moved. SIZEL and SIZEH do not need to be consecutive memory locations, or even on the zero page for that matter. These routines only take one additional cycle if SIZEL is not on the zero page. Likewise, they only take one additional cycle if SIZEH is not on the zero page. Note that this adds only to the total number of cycles, not to the number of cycles per byte, since neither SIZEL nor SIZEH is inside a loop anywhere.
 +
 +These routines are intended to be both flexible and practical, without being excessively lengthy or excessively slow. To that end, they can be placed in ROM or in RAM.
 +
 +There are three routines moving memory upward (i.e. to a higher address), each of which is tailored to a slightly different set of input parameters.
 +<code>
 +; Move memory down
 +;
 +; FROM = source start address
 +;   TO = destination start address
 +; SIZE = number of bytes to move
 +;
 +MOVEDOWN LDY #0
 +         LDX SIZEH
 +         BEQ MD2
 +MD1      LDA (FROM),Y ; move a page at a time
 +         STA (TO),Y
 +         INY
 +         BNE MD1
 +         INC FROM+1
 +         INC TO+1
 +         DEX
 +         BNE MD1
 +MD2      LDX SIZEL
 +         BEQ MD4
 +MD3      LDA (FROM),Y ; move the remaining bytes
 +         STA (TO),Y
 +         INY
 +         DEX
 +         BNE MD3
 +MD4      RTS
 +
 +; Move memory up
 +;
 +; FROM = source start address
 +;   TO = destination start address
 +; SIZE = number of bytes to move
 +;
 +MOVEUP   LDX SIZEH    ; the last byte must be moved first
 +         CLC          ; start at the final pages of FROM and TO
 +         TXA
 +         ADC FROM+1
 +         STA FROM+1
 +         CLC
 +         TXA
 +         ADC TO+1
 +         STA TO+1
 +         INX          ; allows the use of BNE after the DEX below
 +         LDY SIZEL
 +         BEQ MU3
 +         DEY          ; move bytes on the last page first
 +         BEQ MU2
 +MU1      LDA (FROM),Y
 +         STA (TO),Y
 +         DEY
 +         BNE MU1
 +MU2      LDA (FROM),Y ; handle Y = 0 separately
 +         STA (TO),Y
 +MU3      DEY
 +         DEC FROM+1   ; move the next page (if any)
 +         DEC TO+1
 +         DEX
 +         BNE MU1
 +         RTS
 +
 +; Move memory up
 +;
 +; FROM = 1 + source end address
 +; TO   = 1 + destination end address
 +; SIZE = number of bytes to move
 +;
 +MOVEUP   LDY #$FF
 +         LDX SIZEH
 +         BEQ MU3
 +MU1      DEC FROM+1
 +         DEC TO+1
 +MU2      LDA (FROM),Y ; move a page at a time
 +         STA (TO),Y
 +         DEY
 +         BNE MU2
 +         LDA (FROM),Y ; handle Y = 0 separately
 +         STA (TO),Y
 +         DEY
 +         DEX
 +         BNE MU1
 +MU3      LDX SIZEL
 +         BEQ MU5
 +         DEC FROM+1
 +         DEC TO+1
 +MU4      LDA (FROM),Y ; move the remaining bytes
 +         STA (TO),Y
 +         DEY
 +         DEX
 +         BNE MU4
 +MU5      RTS
 +
 +; Move memory up
 +;
 +; FROM = source end address
 +; TO   = destination end address
 +; SIZE = number of bytes to move
 +;
 +MOVEUP   LDY #0
 +         LDX SIZEH
 +         BEQ MU3
 +MU1      LDA (FROM),Y ; handle Y = 0 separately
 +         STA (TO),Y
 +         DEY
 +         DEC FROM+1
 +         DEC TO+1
 +MU2      LDA (FROM),Y ; move a page at a time
 +         STA (TO),Y
 +         DEY
 +         BNE MU2
 +         DEX
 +         BNE MU1
 +MU3      LDX SIZEL
 +         BEQ MU5
 +         LDA (FROM),Y ; handle Y = 0 separately
 +         STA (TO),Y
 +         DEY
 +         DEX
 +         BEQ MU5
 +         DEC FROM+1
 +         DEC TO+1
 +MU4      LDA (FROM),Y ; move the remaining bytes
 +         STA (TO),Y
 +         DEY
 +         DEX
 +         BNE MU4
 +MU5      RTS
 +</code>
 +Even more speed can be gained by using self-modifying code, i.e. replacing the (ZeroPage),Y addressing mode with the Absolute,Y addressing mode. This will take 2 fewer cycles per byte. There will be some additional cycles from the added instructions that self-modify the code, but the self-modification occurs only once, and therefore adds these cycles to total number of cycles, rather than the number of cycles per byte moved. As always, the instructions that are self-modified can't be located in ROM.
  
base/practical_memory_move_routines.txt · Last modified: 2015-04-17 04:33 by 127.0.0.1