This commit updates the matrix Binary F32 test patterns and
implementations for the CMSIS-DSP 1.9.0.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit updates the matrix Binary Q31 test patterns and
implementations for the CMSIS-DSP 1.9.0.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit updates the matrix Binary Q15 test patterns and
implementations for the CMSIS-DSP 1.9.0.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds the filtering DECIM F32 test patterns and
implementations for the CMSIS-DSP 1.9.0.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds the filtering DECIM Q31 test patterns and
implementations for the CMSIS-DSP 1.9.0.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds the filtering DECIM Q15 test patterns and
implementations for the CMSIS-DSP 1.9.0.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit makes the tests for the partial convolution functions
(`arm_conv_partial*`) optional because these functions have computation
precision-related issues (tested with CMSIS-DSP 1.9.0, which is part of
the CMSIS 5.8.0 release).
The partial convolution functions (especially the fast variants) can
return very low-precision results (as low as 14dB SNR) under certain
configurations.
This issue will be further investigated and reported to the upstream
CMSIS developers.
Once this issue is fixed upstream, revert this commit in order to make
the partial convolution tests run all the time.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds the filtering MISC F16 test patterns and
implementations for the CMSIS-DSP 1.9.0.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit updates the filtering MISC F32 test patterns and
implementations for the CMSIS-DSP 1.9.0.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit updates the filtering MISC Q31 test patterns and
implementations for the CMSIS-DSP 1.9.0.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit updates the filtering MISC Q15 test patterns and
implementations for the CMSIS-DSP 1.9.0.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit updates the filtering MISC Q7 test patterns and
implementations for the CMSIS-DSP 1.9.0.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds the filtering FIR F16 test patterns and
implementations for the CMSIS-DSP 1.9.0.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds the filtering BIQUAD F16 test patterns and
implementations for the CMSIS-DSP 1.9.0.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit updates the basicmath F32 test patterns and implementations
for the CMSIS-DSP 1.9.0.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit updates the basicmath Q31 test patterns and implementations
for the CMSIS-DSP 1.9.0.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit updates the basicmath Q15 test patterns and implementations
for the CMSIS-DSP 1.9.0.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds the common test and validation routines for the half-
precision floating point data.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
Move to CMake 3.20.0.
At the Toolchain WG it was decided to move to CMake 3.20.0.
The main reason for increasing CMake version is better toolchain
support.
Better toolchain support is added in the following CMake versions:
- armclang, CMake 3.15
- Intel oneAPI, CMake 3.20
- IAR, CMake 3.15 and 3.20
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
The CMSIS-DSP matrix binary_q15 tests should not be excluded on the
mps2_an521 board, which is one of the integration platforms.
This test was likely excluded because of the computation precision-
related failures; but, this failure is no longer observed, so there is
no reason to exclude it anymore.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds new testcases for the CMSIS-DSP transform test that
enable testing with hardware FPU.
Note that the common properties are relocated to each testcase because
the twister does not support specifying extra filter and tags.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds a new testcase for the CMSIS-DSP svm test that
enables testing with hardware FPU.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds a new testcase for the CMSIS-DSP support test that
enables testing with hardware FPU.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds a new testcase for the CMSIS-DSP statistics test that
enables testing with hardware FPU.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds new testcases for the CMSIS-DSP matrix test that
enable testing with hardware FPU.
Note that the common properties are relocated to each testcase because
the twister does not support specifying extra filters and tags.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds new testcases for the CMSIS-DSP filtering test that
enable testing with hardware FPU.
Note that the common properties are relocated to each testcase because
the twister does not support specifying extra filters and tags.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds a new testcase for the CMSIS-DSP fastmath test that
enables testing with hardware FPU.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds a new testcase for the CMSIS-DSP distance test that
enables testing with hardware FPU.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds a new testcase for the CMSIS-DSP complexmath test that
enables testing with hardware FPU.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds a new testcase for the CMSIS-DSP bayes test that
enables testing with hardware FPU.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds a new testcase for the CMSIS-DSP basicmath test that
enables testing with hardware FPU.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
CMSIS_DSP and net_socket tests are either too
slow or too large for this platform, also drop
bluetooth for test
Signed-off-by: Hake Huang <hake.huang@oss.nxp.com>
some tests configured with CONFIG_NEWLIB_LIBC=y,
it's better to add a filter filter: TOOLCHAIN_HAS_NEWLIB == 1
in those tests yaml file.
Signed-off-by: Watson Zeng <zhiwei@synopsys.com>
tests/lib/cmsis_dsp/transform/ with config:
* libraries.cmsis_dsp.transform.cf64
96k is sufficient if tests steps are reordered
(decreasing malloc size)
Tested on nucleo_l476rg
Signed-off-by: Alexandre Bourdiol <alexandre.bourdiol@st.com>
tests/lib/cmsis_dsp/transform/ with config:
* libraries.cmsis_dsp.transform.cf64
Needs more than 64k RAM.
Signed-off-by: Alexandre Bourdiol <alexandre.bourdiol@st.com>
tests/lib/cmsis_dsp/matrix/ with config:
* libraries.cmsis_dsp.matrix.binary_q15
128k is sufficient if tests steps are reordered,
otherwise system run out of memory.
Biggest chunk first for better malloc chunk reusability.
Tested on nucleo_f207zg
Signed-off-by: Alexandre Bourdiol <alexandre.bourdiol@st.com>
Stephanos Ioannidis