Arnon,

Here is a way you can multiply the equivalent of 0.034145355224609375 times whatever number. If the number you are multiplying by is over 0x3fff, you will overrun the accumulator. If you don't need as much accuracy, you can change the shiftFactor to 16, and things get easier in assembly by just using the upper 16 bits of the multiplication result.

<>< Lance.

typedef unsigned int uint16;
typedef unsigned long uint32;
int shiftFactor = 18;
uint32 MagicConstant = 8951; // .034146 * (1 << BitShift)


uint16 MultiplyBy0_034146( uint16 value )
{
// This will round the result.
return ( (uint32)value * MagicConstant + (1<<(shiftFactor-1))) >> shiftFactor;
}