Shifting bits in an arbitrary number of bytes

coco2 · **Posted:** Sat Jun 21, 2014 1:09 am

I found it, I had to dec ecx immediately before the jump condition test, because of the zero flag I'm assuming. Well encryption is already 80% faster with just these simple changes, I don't know if I need to work on more optimisations since RSA is falling out of use to ECC these days.

wilbert · **Posted:** Fri Oct 03, 2014 1:44 pm

A faster implementation of the two procedures ...

Code:

Procedure ShiftLeftBuffer(*data, length.l)
  !xor eax, eax
  !mov ecx, [p.v_length]
  !dec ecx
  !jc slb_exit
  CompilerIf #PB_Compiler_Processor = #PB_Processor_x64
    !mov rdx, [p.p_data]
    !slb_loop:  
    !mov ah, [rdx + rcx]
    !shl eax, 1
    !mov [rdx + rcx], ah
  CompilerElse
    !mov edx, [p.p_data]
    !slb_loop:  
    !mov ah, [edx + ecx]
    !shl eax, 1
    !mov [edx + ecx], ah
  CompilerEndIf
  !shr eax, 9
  !sub ecx, 1
  !jnc slb_loop
  !slb_exit:
  !shr eax, 7
  ProcedureReturn
EndProcedure

Procedure ShiftRightBuffer(*data, length.l)
  !xor eax, eax
  !mov ecx, [p.v_length]
  !and ecx, ecx
  !jz srb_exit
  CompilerIf #PB_Compiler_Processor = #PB_Processor_x64
    !mov rdx, [p.p_data]
    !srb_loop:  
    !mov ah, [rdx]
    !shr eax, 1
    !mov [rdx], ah
    !inc rdx
  CompilerElse
    !mov edx, [p.p_data]
    !srb_loop:  
    !mov ah, [edx]
    !shr eax, 1
    !mov [edx], ah
    !inc edx
  CompilerEndIf  
  !shl eax, 9
  !dec ecx
  !jnz srb_loop
  !srb_exit:
  !shr eax, 16
  !and eax, 1
  ProcedureReturn
EndProcedure

coco2 · **Posted:** Thu Dec 04, 2014 12:56 pm

Thanks I have been doing some testing and noticed it doesn't help with my large number module any more, so I will have to optimise some more. I've noticed that all the time is spent on this:

Code:

      While i
        i=i-1
        If j
          j=j-1
          sum = PeekA(*h1\num+i) + PeekA(*h2\num+j) + carry
        Else
          sum = PeekA(*h1\num+i) + carry
        EndIf
        If sum > $FF : carry = 1 : Else : carry = 0 : EndIf
        PokeA(*h1\num+i, sum)
      Wend

so looks like I'll have to delve into assembly again

wilbert · **Posted:** Thu Dec 04, 2014 6:10 pm

coco2 wrote:

Thanks I have been doing some testing and noticed it doesn't help with my large number module any more, so I will have to optimise some more. I've noticed that all the time is spent on this:

You could already speed this up a lot without using asm.
I wanted to try something with you huge.pbi code to improve the speed but your code depends on cc2debug.pbi and it looks like you didn't post that code.

coco2 · **Posted:** Thu Dec 04, 2014 10:07 pm

This one?

Code:

Macro DebugOut(DebugText)
  Debug #PB_Compiler_Module+"::"+#PB_Compiler_Procedure+": "+DebugText
EndMacro

Macro DebugPrintN (DebugText)
  PrintN(#PB_Compiler_Module+"::"+#PB_Compiler_Procedure+": "+DebugText)
EndMacro

PureBasic Forum

Shifting bits in an arbitrary number of bytes

Who is online