According to the C standard q15_t * and const q15_t * are not compatible
types which, among other things, imply that an object of type const
q15_t * can't be modified by writing to a q15_t ** or vice versa.
Programs doing this are undefined.
Because of this rule all programs using the functions read_q15x2_ia,
read_q15x2_da, read_q7x4_ia, or read_q7x4_da for reading data from an
array of constant elements will be undefined. To solve this it is not
enough to change the type of the function since this will give problems
when reading data from an array of non-const elements. To get a defined
solution I needed to switch from functions to macros to allow the
increment to be done in the original type of the pointer.
Solve most of f16 issues. But there are still some remaining
build issues with gcc10q4.
2 functions are reverting to scalar version when build with gcc on M55.
(Since Helium versions of those functions are not building).
Longer test patterns for BasicMaths, ComplexMaths and Statistics.
Corrected saturation bug in MVE code for arm_abs_q7,
arm_cmplx_conj_q15 and arm_cmplx_conj_q31.
Added a new header of vectorial inline functions for Neon and Helium.
This header is used in the implementation of some functions.
Corrected used of float (float functions rather than double ones).
Added test patterns for functions in BasicMaths.
Corrected github issue 622 (clarified comment for lms_norm_f32).
Added support for q63_t to the test framework.
(For our internal use. In short term, we won't give support about it).
CMSIS-DSP: Update to cmake build for the testing framework
CMSIS-NN:Implementation of arm_fully_connected_s8
Use API and quantization compatible with TF Lite.
Compilation symbols have been introduced to allow to include only the
FFT tables which are useful in a given application.
The README.md in DSP is giving more details.
- new version V1.5.5
- reworked DSP library source files
added macro ARM_MATH_LOOPUNROLL
removed macro UNALIGNED_SUPPORT_DISABLE
relpaced arm_bitreversal2.S with C version
added const-correctness
replaced SIMD pointer construct with memcpy solution
- reworked DSP library documentation
- moved DSP libraries to folder ./DSP/Lib
- ARM DSP Libraries are built with ARMCLANG
- Added DSP Libraries Source variant