Is there a 64 bit maths library?

[Vaughn]

I am using a Pic18F67K22 in a project that requires the solving of a matrix for a regression during the calibration of an instrument. This involves many iterations of subtraction, multiplication and division. I have used the standard library as well as the replacement library . It appears that the very small errors caused by the floating point arithmetic grow through the repetitive operations in solving the matrix until , in some conditions, they become unacceptably large. I have checked the calculations using a spreadsheet as well as an online regression calculator.( Xuru.org ) which agree with each other. I have tried many ways of improving the accuracy such as multiplying by 1000 before an operation and then dividing the answer by 1000 so as to reduce the number of decimal places but nothing has improved the accuracy. I can only think that 64 bit maths may solve this problem. Is there a 64 bit maths library available or is there possibly a maths co processor that I could use. Speed is not an issue here because this part of the program will only be used during calibration of the instrument which may be once or twice a year.

[filip.grujcic]

Hello,

I'm afraid there are no official libraries for what you are looking for.
I apologize for the inconvenience.

Regards,