Floating point calculations require a dedicated FPU in the hardware, that alone can take up more power. Now fixed point math can be done purely in software using integer operations, later on saving power on microcontrollers without an FPU. fixed point has lower precision.




Floating pointFPUfixed pointfixed point