I was going to use this process(**broken link removed**), and impliment a long hand multiplcation / division routine for floating point numbers.... but your way seems a lot easier... and a faster
converting binary to hex implies you to do the followings;
1.multiply the individual numbers by that given base i.e base 16.
2.raise their bases by begining from the last i.e.from zero to where ever the base has ended.
3.addup all the multiples by so doing a result is gotting
Good luck.
So, I'm working on making a long hand multiplication routine for floating point numbers.
Is there a way that I can just reserve memory and have an indexable pointer to it?
Right now I'd like to use 17bit's of storage for each floating point number.
1 bit for sign
8 bits for the integer portion
8 bits for the fractional portion
That would allow me to have numbers between 1/128 and 255.9921875 ( + and - )
anyway... it's all pretty easy to do... I just need to figure out a way that I can index through the memory in a loop, instead of having to call my File reg's by name. I could do it even calling them by name... but it would be ugly and probably end up using more code space.
That is not floating point, that is known as fixed point.
To access your variables you use the FSR and INDF registers. Load FSR with the address of your variable and any read or writes to INDF will access the register pointed to by the FSR register.