Division Algorithm

Bagaturia david

"Interested?" Yes I am. About your problem: Use integer division for your non integer Number. Ex: 121/11 = 11 if 1.21/11 = (121/11)*0.01 or 14/3 = 4+2/3 = 4+(20/3)*0.1 = 4+(6+2/3)*0.1 = 4+(6+(20/3)*0.1)*0.1= =...=4+0.6+0.06+0.006+0.0006+..... it is idea and math.I hope you like this.:) Can you send my division algorithm?

Are Jay

I'll apply the above to my class and I'll email you my source or I'll email you a link to it.
Class Diagram::LongMath Addition Add(sourceA As String, sourceB As String) As String Subtraction Subtract(sourceA As String, sourceB As String) As String Multiplication Multiply(sourceA As String, sourceB As String) As String Division Divide(divisor As String, dividend As String) As String This may take me a few days with the holidays here and all.

djlove

Hi, I was playing around with a BigInteger class recently and so have come across the same problem. The fastest way I could think of is below, if you say x / y = z (with remainder = r). 1. Find the leading 64 bits of your big number (x) 2. Find the leading 32 bits of your divisor (y). 3. Divide (1) by (2) to give you the leading 32 bits of z. 4. Multiply (3) by y and take this number from x Continually applying steps 1-4 reduces the size of x until you are within +/- 1 of the correct value of z. This is fast because all operations are O(n) or less if you code them properly! Hope this is of some use & good luck! Russ

Are Jay

I've posted my source code to my LongMath dll. link to my profile where I've posted it.[^] After re-reading your prior post I'm not sure if this is what your looking for. If not, I'm sorry.:-O

Dan2178

Hi, I can't help specifically with c++ code to do long division .. but I wrote code in vb to do arbitrary precision floating point math. The division method that I used turned out to be suprisingly fast and much easier to implement than addition or subtraction with signed floating point numbers. The method is called the Newton-Raphson method or just Newton's method. I won't go into it here, just search for it, you'll find numerous examples. I don't know how well it would would with integers only since the method uses multiplication, addition and subtraction iteratively but principal is sound anyway. The guts of the divide method I wrote is this: Do While Compare <> Guess Compare = Guess Guess = Left(Multiply(Subtract("2", Multiply(y, Guess)), Guess), Precision) Loop If your multiply, add and subtract methods can handle large numbers (this one uses strings) and you provide code to prevent the length of the result with each iteration from getting too long then sometime like may work. Each iteration will double the accuracy. With another 50 or so lines of code this method can divide numbers thousands of digits long in milliseconds. I also read somewhere where the Newton method is the fastest for divison.