CMSIS-DSP
About
CMSIS-DSP is an optimized compute library for embedded systems (DSP is in the name for legacy reasons).
It provides optimized compute kernels for Cortex-M and for Cortex-A.
Different variants are available according to the core and most of the functions are using a vectorized version when the Helium or Neon extension is available.
This repository contains the CMSIS-DSP library and several other projects:
- Test framework for bare metal Cortex-M or Cortex-A
- Examples for bare metal Cortex-M
- ComputeGraph
- PythonWrapper
You don't need any of the other projects to build and use CMSIS-DSP library. Building the other projects may require installation of other libraries (CMSIS), other tools (Arm Virtual Hardware) or CMSIS build tools.
CMSIS-DSP Kernels
Kernels provided by CMSIS-DSP (list not exhaustive):
- Basic mathematics (real, complex, quaternion, linear algebra, fast math functions)
- DSP (filtering)
- Transforms (FFT, MFCC, DCT)
- Statistics
- Classical ML (Support Vector Machine, Distance functions for clustering ...)
Kernels are provided with several datatypes : f64, f32, f16, q31, q15, q7.
Python wrapper
A PythonWrapper is also available and can be installed with:
pip install cmsisdsp
With this wrapper you can design your algorithm in Python using an API as close as possible to the C API. The wrapper is compatible with NumPy. The wrapper is supporting fixed point arithmetic. This wrapper works in google colab.
The goal is to make it easier to move from a design to a final implementation in C.
Compute Graph
CMSIS-DSP is also providing an experimental static scheduler for compute graph to describe streaming solutions:
- You define your compute graph in Python
- A static and deterministic schedule (computed by the Python script) is generated
- The static schedule can be run on the device with low overhead
The Python scripts for the static scheduler generator are part of the CMSIS-DSP Python wrapper.
The header files are part of the CMSIS-DSP pack (version 1.10.2 and above).
The Compute Graph makes it easier to implement a streaming solution : connecting different compute kernels each consuming and producing different amount of data.
Support / Contact
For any questions or to reach the CMSIS-DSP team, please create a new issue in https://github.com/ARM-software/CMSIS-DSP/issues
Table of content
Building for speed
CMSIS-DSP is used when you need performance. As consequence CMSIS-DSP should be compiled with the options giving the best performance:
Options to use
-Ofast
must be used for best performances.- When using Helium it is strongly advised to use
-Ofast
GCC
is currently not giving good performances when targeting Helium. You should use the Arm compiler
When float are used, then the fpu should be selected to ensure that the compiler is not using a software float emulation.
When building with Helium support, it will be automatically detected by CMSIS-DSP. For Neon, it is not the case and you must enable the option -DARM_MATH_NEON
for the C compilation. With cmake
this option is controlled with -DNEON=ON
.
-DLOOPUNROLL=ON
can also be used when compiling with cmake- It corresponds to the C options
-DARM_MATH_LOOPUNROLL
Compilers are doing unrolling. So this option may not be needed but it is highly dependent on the compiler. With some compilers, this option is needed to get better performances.
Speed of memory is important. If you can map the data and the constant tables used by CMSIS-DSP in DTCM
memory then it is better. If you have a cache, enable it.
Options to avoid
-fno-builtin
-ffreestanding
because it enables previous options
The library is doing some type punning to process word 32 from memory as a pair of q15
or a quadruple of q7
. Those type manipulations are done through memcpy
functions. Most compilers should be able to optimize out those function calls when the length to copy is small (4 bytes).
This optimization will not occur when -fno-builtin
is used and it will have a very bad impact on the performances.
Some compiler may also require the use of option -munaligned-access
to specify that unaligned accesses are used.
Half float support
f16
data type (half float) has been added to the library. It is useful only if your Cortex has some half float hardware acceleration (for instance with Helium extension). If you don't need f16
, you should disable it since it may cause compilation problems. Just define -DDISABLEFLOAT16
when building.
How to build
You can build CMSIS-DSP with the open CMSIS-Pack, or cmake, or Makefile and it is also easy to build if you use any other build tool.
How to build with MDK or Open CMSIS-Pack
The standard way to build is by using the CMSIS pack technology. CMSIS-DSP is available as a pack.
This pack technology is supported by some IDE like Keil MDK or Keil studio.
You can also use those packs using the Open CMSIS-Pack technology and from command line on any platform.
You should first install the tools from https://github.com/Open-CMSIS-Pack/devtools/tree/main/tools
You can get the CMSIS-Toolbox which is containing the package installer, cmsis build and cmsis project manager. Here is some documentation:
- Documentation about CMSIS Build
- Documentation about CMSIS Pack
- Documentation about CMSIS Project manager
Once you have installed the tools, you'll need to download the pack index using the cpackget
tool.
Then, you'll need to convert a solution file into .cprj
. For instance, for the CMSIS-DSP Examples, you can go to:
Examples/cmsis_build
and then type
csolution convert -s examples.csolution_ac6.yml
This command processes the examples.csolution_ac6.yml
describing how to build the examples for several platforms. It will generate lots of .cprj
files that can be built with cbuild
.
If you want to build the FFT
example for the Corstone-300
virtual hardware platform, you could just do:
cbuild "fftbin.Release+VHT-Corstone-300.cprj"
How to build with Make
There is an example Makefile
in Source
.
In each source folder (like BasicMathFunctions
), you'll see files with no _datatype
suffix (like BasicMathFunctions.c
and BasicMathFunctionsF16.c
).
Those files are all you need in your makefile. They are including all other C files from the source folders.
Then, for the includes you'll need to add the paths: Include
, PrivateInclude
and, since there is a dependency to CMSIS Core, Core/Include
from CMSIS_5/CMSIS
.
If you are building for Cortex-A
and want to use Neon, you'll also need to include ComputeLibrary/Include
and the source file in ComputeLibrary/Source
.
How to build with cmake
Create a CMakeLists.txt
and inside add a project.
Add CMSIS-DSP as a subdirectory. The variable CMSISDSP
is the path to the CMSIS-DSP repository in below example.
cmake_minimum_required (VERSION 3.14)
# Define the project
project (testcmsisdsp VERSION 0.1)
add_subdirectory(${CMSISDSP}/Source bin_dsp)
CMSIS-DSP is dependent on the CMSIS Core includes. So, you should define CMSISCORE
on the cmake command line. The path used by CMSIS-DSP will be ${CMSISCORE}/Include
.
You should also set the compilation options to use to build the library.
If you build for Helium, you should use any of the option MVEF
, MVEI
or HELIUM
.
If you build for Neon, use NEON
and/or NEONEXPERIMENTAL
.
Launching the build
Once cmake has generated the makefiles, you can use a GNU Make to build.
make VERBOSE=1
How to build with any other build system
You need the following folders:
- Source
- Include
- PrivateInclude
- ComputeLibrary (only if you target Neon)
In Source
subfolders, you may either build all of the source file with a datatype suffix (like _f32.c
), or just compile the files without a datatype suffix. For instance for BasicMathFunctions
, you can build all the C files except BasicMathFunctions.c
and BasicMathFunctionsF16.c
, or you can just build those two files (they are including all of the other C files of the folder).
f16
files are not mandatory. You can build with -DDISABLEFLOAT16
How to build for aarch64
The intrinsics defined in Core_A/Include
are not available on recent Cortex-A processors.
But you can still build for those Cortex-A cores and benefit from the Neon intrinsics.
You need to build with -D__GNUC_PYTHON__
on the compiler command line. This flag was introduced for building the Python wrapper and is disabling the use of CMSIS Core includes.
When this flag is enabled, CMSIS-DSP is defining a few macros used in the library for compiler portability:
#define __ALIGNED(x) __attribute__((aligned(x)))
#define __STATIC_FORCEINLINE static inline __attribute__((always_inline))
#define __STATIC_INLINE static inline
If the compiler you are using is requiring different definitions, you can add them to arm_math_types.h
in the Include
folder of the library. MSVC and XCode are already supported and in those case, you don't need to define -D__GNUC_PYTHON__
Then, you need to define -DARM_MATH_NEON
For cmake the equivalent options are:
-DHOST=ON
-DNEON=ON
cmake is automatically including the ComputeLibrary
folder. If you are using a different build, you need to include this folder too to build with Neon support.
Code size
Previous versions of the library were using compilation directives to control the code size. It was too complex and not available in case CMSIS-DSP is only delivered as a static library.
Now, the library relies again on the linker to do the code size optimization. But, this implies some constraints on the code you write and new functions had to be introduced.
If you know the size of your FFT in advance, use initializations functions like arm_cfft_init_64_f32
instead of using the generic initialization functions arm_cfft_init_f32
. Using the generic function will prevent the linker from being able to deduce which functions and tables must be kept for the FFT and everything will be included.
There are similar functions for RFFT, MFCC ...
If the flag ARM_DSP_CONFIG_TABLES
is still set, you'll now get a compilation error to remind you that this flag no more have any effect on code size and that you may have to rework the initializations.
Folders and files
The only folders required to build and use CMSIS-DSP Library are:
- Source
- Include
- PrivateInclude
- ComputeLibrary (only when using Neon)
Other folders are part of different projects, tests or examples.
Folders
-
cmsisdsp
- Required to build the CMSIS-DSP PythonWrapper for the Python repository
- It contains all Python packages
-
ComputeLibrary:
- Some kernels required when building CMSIS-DSP with Neon acceleration
-
Examples:
- Examples of use of CMSIS-DSP on bare metal Cortex-M
- Require the use of CMSIS Build tools
-
Include:
- Include files for CMSIS-DSP
-
PrivateInclude:
- Some include needed to build CMSIS-DSP
-
PythonWrapper:
- C code for the CMSIS-DSP PythonWrapper
- Examples for the PythonWrapper
-
Scripts:
- Debugging scripts
- Script to generate some coefficient tables used by CMSIS-DSP
-
Compute Graph:
- Not needed to build CMSIS-DSP. This project is relying on CMSIS-DSP library
- Examples for the Compute Graph
- C++ templates for the Compute Graph
- Default (and optional) nodes
-
Source:
- CMSIS-DSP source
-
Testing:
- CMSIS-DSP Test framework for bare metal Cortex-M and Cortex-A
- Require the use of CMSIS build tools
Files
Some files are needed to generate the PythonWrapper:
- PythonWrapper_README.md
- LICENSE.txt
- MANIFEST.in
- pyproject.toml
- setup.py
And we have a script to make it easier to customize the build:
- cmsisdspconfig.py:
- Web browser UI to generate build configurations (temporary until the CMSIS-DSP configuration is reworked to be simpler and more maintainable)