RA Flexible Software Package Documentation
Release v5.7.0
|
|
Middleware to implement Arm CMSIS DSP by using 32-bit Multiply-Accumulator (MACL).
The 32-bit Multiply-Accumulator (MACL) provides hardware acceleration for calculation of CMSIS DSP functions. After initializing this module, refer to the Arm documentation to use the file system: https://www.keil.com/pack/doc/CMSIS/DSP/html/index.html
The MACL module supports the following features:
No | MACL API | Feature |
---|---|---|
1 | arm_mult_q31 | Q31 vector multiplication |
2 | arm_scale_q31 | Multiplies a Q31 vector by a scalar |
3 | arm_mat_mult_q31 | Q31 matrix multiplication |
4 | arm_mat_vec_mult_q31 | Q31 matrix and vector multiplication |
5 | arm_mat_scale_q31 | Q31 matrix scaling |
6 | arm_biquad_cascade_df1_q31 | Processing function for the Q31 Biquad cascade filter |
7 | arm_conv_partial_q31 | Partial convolution of Q31 sequences |
8 | arm_conv_q31 | Convolution of Q31 sequences |
9 | arm_correlate_q31 | Correlation of Q31 sequences |
10 | arm_fir_decimate_q31 | Processing function for the Q31 FIR decimator |
11 | arm_fir_interpolate_q31 | Processing function for the Q31 FIR interpolator |
12 | arm_fir_q31 | Processing function for Q31 FIR filter |
13 | arm_fir_sparse_q31 | Processing function for the Q31 sparse FIR filter |
14 | arm_lms_norm_q31 | Processing function for Q31 normalized LMS filter |
15 | arm_lms_q31 | Processing function for Q31 LMS filter |
To enable hardware acceleration for Arm CMSIS DSP, the stack "Arm CMSIS6 DSP Acceleration" of CMSIS DSP must be filled with MACL (rm_cmsis_dsp). Otherwise the CMSIS DSP functions will perform the calculation by using software. This can be done at the Stack Configuration.
MACL module is required for using this middleware. Refer to the MCU hardware user's manual or datasheet to determine if it has MACL support.
About examples, how to use the CMSIS DSP APIs refer to the link: https://github.com/ARM-software/CMSIS-DSP/tree/main/Examples/ARM.