Understanding assembler logics

[Walle]

Ok, so I'm learning assembler. What I'm having a hard time grasping is how to do basic arithmetics.

Addition and Subtraction is no problems ofc, but how to do multiplication and division is harder to get. Multiplication/division by any number that is a power of 2 is easy to go ofc, but what about multiplying a number by say 3?
To multiply by one I would only shift left one bit, right? To multiply by 4 I would shift left 2 bits, but how do I multiply by 3?

What I really would like is an application that take C-style operations and output assembler-like code. Ie x=x+2 would output something like:

movlw 0x02
addwf x,1

A=B+C would output something like:

movf B,0
addwf C,1
movwf A

you get the picture, right?

Are there any tools like this? Can anyone point me to some other good resources on learning the logics behind arithmetic operations in assembler?

/Chris

[TkTech]

movfw? movf? I know x86 nasm style asm, and these aren't any ops I know. Is this for a MIPS machine or a different architecture? ASM is very low level, you need to specify the compiler/architecture your using.

ALL C compilers that I know of offer a flag to output to asm, since a C compiler will compile to proper asm before its compiled into machine code. For GCC, this flag is -S.

Why are you not simply using the DIV/MUL-series of instructions?

[Walle]

Sorry, I forgot to mention that I'm programming MCU's, mostly 8-bit MCU's from Microchip. The instruction set is very limited, with a total of 35 ops to play with. The only arithmetic ops available are addition, subtraction, complement, increment, decrement, and the left/right shifts. Then ofc the logic ops.

Ofc! Thank you! The C-compiler (Hitech) must have an asm-output as well as binary, I will look into that!

[G_Morgan]

For integer multiplication you can always write a loop that does the following.

int mul(int first, int second) {
  int output = 0;

  while(second > 0) {
    output += first;
    second--;
  }

  return output;
}

[appstateguy]

You probably won't gain too much my looking at the assembly of C output since there are multiply and divide instructions for x86.

Say you have your values you wish to multiply together in two registers (R1 & R2), you will simply need to add the value within R1 R2 times.
I.E. 2 * 3 = 2 + 2 + 2

Since i don't know MCU I'll use pseudocode

//R1 * R2 -> R3

MOV 0 to R3 //clears R3
TEST R1 //set condition codes for R1 -- ADD 0 to R1 would also work
Branch on zero to End:
TEST R2 //set condition codes for R2 -- ADD 0 to R2 would also work
Branch on zero to End:
MUL:
ADD R1 to R3
ADD -1 to R2
Branch on zero flag to End // done
Branch Unconditionally to MUL: //continue adding
End:
//R3 now contains the result

Integer division would be similar, but a little more complicated. Hope that helps.