Fixed point operations in C

[kaushiksangam]

I want to emulate an algorithm in C before actually writing to hardware.

This algorithm invovles lot of multiplications and additions of real numbers. Hence, I want to use fixed point operations(to represent real numbers) in C to emulate the algorithm.

Can someone please tell me or give me an example of how to perform fixed point operations in C.

 

[jagan326]

hey can u give me the details of this topic

 

[kaushiksangam]

It is about the implementation of a Brain State in a Box (BSB) algorithm. Before implementinng this in hardware, i would want to emulate the same in C.

Already have implemented algorithm in C through floating point oprations.

I can also send you the code of the implementation.

 

[neuralc]

Why dont use floating point?
Which up are you using?

NeuralC

 

[alvarorahul]

u can use an array of two integers for each number then, for example
0.004676 can be stored as
var1=4676 and var2= -3 u can u short int for the var2 coz im assuming that ur powers would not vary too much

 

[neuralc]

This is floating point....

 

[lambtron]

Can someone please tell me or give me an example of how to perform fixed point operations in C.

Have a look at this: **broken link removed**

 

[louisnells]

Hi,
Fixed point operations in C is not that difficult to implement.
Lets say we are storing numbers in 8.24 format. So we can use a 32 bit un-signed integer for that. We are going to store the integer 8 bits in the MSB side and the fractional 24 bits on the LSB side (as we always represent in digital form - Little Endian). Then may be you should write a small library for fixed operations. Then implement operations like multiplication as bit-wise operations.

 

[rme_elektroda]

Probably useful link

**broken link removed**