Hands-On RTOS with Microcontrollers

Hardware floating-point units

If your application is going to be crunching lots of floating point numbers, a hardware floating-point unit (FPU) can be extremely helpful. Until the past decade or so, floating point numbers were generally best avoided in most MCU-based embedded systems. The availability of faster processors started to change this. Now, FPUs are often implemented in hardware. Thanks to FPUs, many different applications can benefit from using floating point math, without incurring the CPU performance penalty commonly associated with software-based library implementations.

Single-precision (32-bit) FPUs are optional on Cortex-M4 processors, while Cortex-M7-based processors add optional hardware support for double-precision (64-bit) ...