ARM Math

1. Introduction 

Microcontrollers generally have microprocessors having low computation capacity. Even some of them do not have Floating-Point Processing Unit (FPU). So it can be very expensive to do even common mathematics like floating point calculation, trigonometric calculation, matrix calculation etc. So, to do such calculation efficienty, we use Cortex Microcontroller Software Interface Standard (CMSIS) arm maths library. If you are more interested about floating point and floating point library, search for IEE 754 and Qfplib.

In STM32 project, arm_math header and source files are inside Drivers > CMSIS > DSP. Cortex M3 microcontrollers do not have FPU so there is linkable math library inside Drivers > CMSIS > Lib.

2. Linking Arm Math Library for ARM Cortex-M3 

You can link the library by adding the GCC library path Drivers/CMSIS/Lib/GCC and the linker flag “libarm_cortexM3l_math.a in the Makefile or CMakeList.txt.

# libraries
LIBS = -lc -lm -lnosys -larm_cortexM3l_math.a
LIBDIR = -LDrivers/CMSIS/Lib/GCC/

# Link directories setup
target_link_directories(${CMAKE_PROJECT_NAME} PRIVATE
    # Add user defined library search paths
    Drivers/CMSIS/Lib/GCC
)

# Add linked libraries
target_link_libraries(${CMAKE_PROJECT_NAME}
    stm32cubemx

    # Add user defined libraries
    arm_cortexM3l_math
)

3. Using Arm Math Library 

Add the DSP include path.

# C includes
C_INCLUDES =  \
...
-IDrivers/CMSIS/DSP/Include

# Add include paths
target_include_directories(${CMAKE_PROJECT_NAME} PRIVATE
   # Add user defined include paths
   Drivers/CMSIS/DSP/Include
)

To use the arm math library, you need to include the header file arm_math.h in your source file.

#include "arm_math.h"

Let’s suppose you want to calculate sine and cosine of an angle. To do so, you need to add the respective sources in the Makefile or CMakeLists.txt.

# C sources
C_SOURCES = \
...
Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_sin_f32.c \
Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_cos_f32.c

# Add sources to executable
target_sources(${CMAKE_PROJECT_NAME} PRIVATE
   # Add user sources here
   Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_sin_f32.c
   Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_cos_f32.c
)

Don’t forget to define ARM_MATH version in the Makefile or CMakeLists.txt.

# C defines
C_DEFS =  \
...
-DARM_MATH_CM3 # For Cortex M3, -DARM_MATH_CM4 for Cortex M4

# Add project symbols (macros)
target_compile_definitions(${CMAKE_PROJECT_NAME} PRIVATE
   # Add user definitions
   ARM_MATH_CM3 # For Cortex M3, ARM_MATH_CM4 for Cortex M4
)

Now you can use the functions arm_sin_f32 and arm_cos_f32 to calculate the sine and cosine of an angle.

float32_t angle = 0.5;
float32_t sine = arm_sin_f32(angle);
float32_t cosine = arm_cos_f32(angle);

Issues 

Error

Drivers/CMSIS/Include/core_cm4.h:105:8: error: #error “Compiler generates FPU instructions for a device without an FPU (check __FPU_PRESENT)”: 105 | #error “Compiler generates FPU instructions for a device without an FPU (check __FPU_PRESENT)”

You may get such error for Cortex M4 or others having FPU. To solve this, define __FPU_PRESENT=1U in the Makefile or CMakeList.

# C defines
C_DEFS =  \
...
-D__FPU_PRESENT=1U # For having FPU

# Add project symbols (macros)
target_compile_definitions(${CMAKE_PROJECT_NAME} PRIVATE
   # Add user definitions
   __FPU_PRESENT=1U # For having FPU
)

Now let’s learn to use matrix function

ARM Math

1. Introduction

2. Linking Arm Math Library for ARM Cortex-M3

3. Using Arm Math Library

Issues

1. Introduction 

2. Linking Arm Math Library for ARM Cortex-M3 

3. Using Arm Math Library 

Issues 