# Multiplication And Division

Fixed multiplication and division can converted to bit shifting to optimise code. The bit shifting operations can be found by doing a Taylor’s expansion of the equation. The error in the result depends on how many bit shifting operations you wish to perform. The below example uses bit shifting instead of fixed-number multiplication to reduces the number of execution cycles, and shows the difference and error between the result and the actual value.

# Powers And Square Roots

pow(), a function provided with most IDE’s, is a slow function. If raising to the power of an integer, multiply the variable by itself rather than use the pow() function.

There is a brilliant article on square root optimisation: http://www.azillionmonkeys.com/qed/sqroot.html

# Fixed Point

mbedded.ninja has an open-source fixed-point library (MFixedPoint) that is hosted on GitLab.

# Trigonometry

Fast and accurate sin/cos approximations can be found here.

The CORDIC algorithm can be used, which is a simple and efficient algorithm for calculating trigonometry and hyperbolic functions.

Posted: March 18th, 2013 at 9:09 pm
Last Updated on: November 26th, 2017 at 1:46 pm