diff options
Diffstat (limited to 'src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions')
20 files changed, 2748 insertions, 0 deletions
diff --git a/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_copy_f32.c b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_copy_f32.c new file mode 100644 index 000000000..619c1577a --- /dev/null +++ b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_copy_f32.c @@ -0,0 +1,130 @@ +/* ----------------------------------------------------------------------------
+* Copyright (C) 2010 ARM Limited. All rights reserved.
+*
+* $Date: 15. February 2012
+* $Revision: V1.1.0
+*
+* Project: CMSIS DSP Library
+* Title: arm_copy_f32.c
+*
+* Description: Copies the elements of a floating-point vector.
+*
+* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
+*
+* Version 1.1.0 2012/02/15
+* Updated with more optimizations, bug fixes and minor API changes.
+*
+* Version 1.0.10 2011/7/15
+* Big Endian support added and Merged M0 and M3/M4 Source code.
+*
+* Version 1.0.3 2010/11/29
+* Re-organized the CMSIS folders and updated documentation.
+*
+* Version 1.0.2 2010/11/11
+* Documentation updated.
+*
+* Version 1.0.1 2010/10/05
+* Production release and review comments incorporated.
+*
+* Version 1.0.0 2010/09/20
+* Production release and review comments incorporated.
+*
+* Version 0.0.7 2010/06/10
+* Misra-C changes done
+* ---------------------------------------------------------------------------- */
+
+#include "arm_math.h"
+
+/**
+ * @ingroup groupSupport
+ */
+
+/**
+ * @defgroup copy Vector Copy
+ *
+ * Copies sample by sample from source vector to destination vector.
+ *
+ * <pre>
+ * pDst[n] = pSrc[n]; 0 <= n < blockSize.
+ * </pre>
+ *
+ * There are separate functions for floating point, Q31, Q15, and Q7 data types.
+ */
+
+/**
+ * @addtogroup copy
+ * @{
+ */
+
+/**
+ * @brief Copies the elements of a floating-point vector.
+ * @param[in] *pSrc points to input vector
+ * @param[out] *pDst points to output vector
+ * @param[in] blockSize length of the input vector
+ * @return none.
+ *
+ */
+
+
+void arm_copy_f32(
+ float32_t * pSrc,
+ float32_t * pDst,
+ uint32_t blockSize)
+{
+ uint32_t blkCnt; /* loop counter */
+
+#ifndef ARM_MATH_CM0
+
+ /* Run the below code for Cortex-M4 and Cortex-M3 */
+ float32_t in1, in2, in3, in4;
+
+ /*loop Unrolling */
+ blkCnt = blockSize >> 2u;
+
+ /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
+ ** a second loop below computes the remaining 1 to 3 samples. */
+ while(blkCnt > 0u)
+ {
+ /* C = A */
+ /* Copy and then store the results in the destination buffer */
+ in1 = *pSrc++;
+ in2 = *pSrc++;
+ in3 = *pSrc++;
+ in4 = *pSrc++;
+
+ *pDst++ = in1;
+ *pDst++ = in2;
+ *pDst++ = in3;
+ *pDst++ = in4;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+ /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
+ ** No loop unrolling is used. */
+ blkCnt = blockSize % 0x4u;
+
+#else
+
+ /* Run the below code for Cortex-M0 */
+
+ /* Loop over blockSize number of values */
+ blkCnt = blockSize;
+
+#endif /* #ifndef ARM_MATH_CM0 */
+
+ while(blkCnt > 0u)
+ {
+ /* C = A */
+ /* Copy and then store the results in the destination buffer */
+ *pDst++ = *pSrc++;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+}
+
+/**
+ * @} end of BasicCopy group
+ */
diff --git a/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_copy_q15.c b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_copy_q15.c new file mode 100644 index 000000000..00be9dc0b --- /dev/null +++ b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_copy_q15.c @@ -0,0 +1,109 @@ +/* ----------------------------------------------------------------------
+* Copyright (C) 2010 ARM Limited. All rights reserved.
+*
+* $Date: 15. February 2012
+* $Revision: V1.1.0
+*
+* Project: CMSIS DSP Library
+* Title: arm_copy_q15.c
+*
+* Description: Copies the elements of a Q15 vector.
+*
+* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
+*
+* Version 1.1.0 2012/02/15
+* Updated with more optimizations, bug fixes and minor API changes.
+*
+* Version 1.0.10 2011/7/15
+* Big Endian support added and Merged M0 and M3/M4 Source code.
+*
+* Version 1.0.3 2010/11/29
+* Re-organized the CMSIS folders and updated documentation.
+*
+* Version 1.0.2 2010/11/11
+* Documentation updated.
+*
+* Version 1.0.1 2010/10/05
+* Production release and review comments incorporated.
+*
+* Version 1.0.0 2010/09/20
+* Production release and review comments incorporated.
+*
+* Version 0.0.7 2010/06/10
+* Misra-C changes done
+* -------------------------------------------------------------------- */
+
+#include "arm_math.h"
+
+/**
+ * @ingroup groupSupport
+ */
+
+/**
+ * @addtogroup copy
+ * @{
+ */
+/**
+ * @brief Copies the elements of a Q15 vector.
+ * @param[in] *pSrc points to input vector
+ * @param[out] *pDst points to output vector
+ * @param[in] blockSize length of the input vector
+ * @return none.
+ *
+ */
+
+void arm_copy_q15(
+ q15_t * pSrc,
+ q15_t * pDst,
+ uint32_t blockSize)
+{
+ uint32_t blkCnt; /* loop counter */
+
+#ifndef ARM_MATH_CM0
+
+ /* Run the below code for Cortex-M4 and Cortex-M3 */
+
+ /*loop Unrolling */
+ blkCnt = blockSize >> 2u;
+
+ /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
+ ** a second loop below computes the remaining 1 to 3 samples. */
+ while(blkCnt > 0u)
+ {
+ /* C = A */
+ /* Read two inputs */
+ *__SIMD32(pDst)++ = *__SIMD32(pSrc)++;
+ *__SIMD32(pDst)++ = *__SIMD32(pSrc)++;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+ /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
+ ** No loop unrolling is used. */
+ blkCnt = blockSize % 0x4u;
+
+
+#else
+
+ /* Run the below code for Cortex-M0 */
+
+ /* Loop over blockSize number of values */
+ blkCnt = blockSize;
+
+#endif /* #ifndef ARM_MATH_CM0 */
+
+ while(blkCnt > 0u)
+ {
+ /* C = A */
+ /* Copy and then store the value in the destination buffer */
+ *pDst++ = *pSrc++;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+}
+
+/**
+ * @} end of BasicCopy group
+ */
diff --git a/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_copy_q31.c b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_copy_q31.c new file mode 100644 index 000000000..7ab0849f6 --- /dev/null +++ b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_copy_q31.c @@ -0,0 +1,118 @@ +/* ----------------------------------------------------------------------
+* Copyright (C) 2010 ARM Limited. All rights reserved.
+*
+* $Date: 15. February 2012
+* $Revision: V1.1.0
+*
+* Project: CMSIS DSP Library
+* Title: arm_copy_q31.c
+*
+* Description: Copies the elements of a Q31 vector.
+*
+* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
+*
+* Version 1.1.0 2012/02/15
+* Updated with more optimizations, bug fixes and minor API changes.
+*
+* Version 1.0.10 2011/7/15
+* Big Endian support added and Merged M0 and M3/M4 Source code.
+*
+* Version 1.0.3 2010/11/29
+* Re-organized the CMSIS folders and updated documentation.
+*
+* Version 1.0.2 2010/11/11
+* Documentation updated.
+*
+* Version 1.0.1 2010/10/05
+* Production release and review comments incorporated.
+*
+* Version 1.0.0 2010/09/20
+* Production release and review comments incorporated.
+*
+* Version 0.0.7 2010/06/10
+* Misra-C changes done
+* -------------------------------------------------------------------- */
+
+#include "arm_math.h"
+
+/**
+ * @ingroup groupSupport
+ */
+
+/**
+ * @addtogroup copy
+ * @{
+ */
+
+/**
+ * @brief Copies the elements of a Q31 vector.
+ * @param[in] *pSrc points to input vector
+ * @param[out] *pDst points to output vector
+ * @param[in] blockSize length of the input vector
+ * @return none.
+ *
+ */
+
+void arm_copy_q31(
+ q31_t * pSrc,
+ q31_t * pDst,
+ uint32_t blockSize)
+{
+ uint32_t blkCnt; /* loop counter */
+
+
+#ifndef ARM_MATH_CM0
+
+ /* Run the below code for Cortex-M4 and Cortex-M3 */
+ q31_t in1, in2, in3, in4;
+
+ /*loop Unrolling */
+ blkCnt = blockSize >> 2u;
+
+ /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
+ ** a second loop below computes the remaining 1 to 3 samples. */
+ while(blkCnt > 0u)
+ {
+ /* C = A */
+ /* Copy and then store the values in the destination buffer */
+ in1 = *pSrc++;
+ in2 = *pSrc++;
+ in3 = *pSrc++;
+ in4 = *pSrc++;
+
+ *pDst++ = in1;
+ *pDst++ = in2;
+ *pDst++ = in3;
+ *pDst++ = in4;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+ /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
+ ** No loop unrolling is used. */
+ blkCnt = blockSize % 0x4u;
+
+#else
+
+ /* Run the below code for Cortex-M0 */
+
+ /* Loop over blockSize number of values */
+ blkCnt = blockSize;
+
+#endif /* #ifndef ARM_MATH_CM0 */
+
+ while(blkCnt > 0u)
+ {
+ /* C = A */
+ /* Copy and then store the value in the destination buffer */
+ *pDst++ = *pSrc++;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+}
+
+/**
+ * @} end of BasicCopy group
+ */
diff --git a/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_copy_q7.c b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_copy_q7.c new file mode 100644 index 000000000..425885753 --- /dev/null +++ b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_copy_q7.c @@ -0,0 +1,110 @@ +/* ----------------------------------------------------------------------
+* Copyright (C) 2010 ARM Limited. All rights reserved.
+*
+* $Date: 15. February 2012
+* $Revision: V1.1.0
+*
+* Project: CMSIS DSP Library
+* Title: arm_copy_q7.c
+*
+* Description: Copies the elements of a Q7 vector.
+*
+* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
+*
+* Version 1.1.0 2012/02/15
+* Updated with more optimizations, bug fixes and minor API changes.
+*
+* Version 1.0.10 2011/7/15
+* Big Endian support added and Merged M0 and M3/M4 Source code.
+*
+* Version 1.0.3 2010/11/29
+* Re-organized the CMSIS folders and updated documentation.
+*
+* Version 1.0.2 2010/11/11
+* Documentation updated.
+*
+* Version 1.0.1 2010/10/05
+* Production release and review comments incorporated.
+*
+* Version 1.0.0 2010/09/20
+* Production release and review comments incorporated.
+*
+* Version 0.0.7 2010/06/10
+* Misra-C changes done
+* -------------------------------------------------------------------- */
+
+#include "arm_math.h"
+
+/**
+ * @ingroup groupSupport
+ */
+
+/**
+ * @addtogroup copy
+ * @{
+ */
+
+/**
+ * @brief Copies the elements of a Q7 vector.
+ * @param[in] *pSrc points to input vector
+ * @param[out] *pDst points to output vector
+ * @param[in] blockSize length of the input vector
+ * @return none.
+ *
+ */
+
+void arm_copy_q7(
+ q7_t * pSrc,
+ q7_t * pDst,
+ uint32_t blockSize)
+{
+ uint32_t blkCnt; /* loop counter */
+
+#ifndef ARM_MATH_CM0
+
+ /* Run the below code for Cortex-M4 and Cortex-M3 */
+
+ /*loop Unrolling */
+ blkCnt = blockSize >> 2u;
+
+ /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
+ ** a second loop below computes the remaining 1 to 3 samples. */
+ while(blkCnt > 0u)
+ {
+ /* C = A */
+ /* Copy and then store the results in the destination buffer */
+ /* 4 samples are copied and stored at a time using SIMD */
+ *__SIMD32(pDst)++ = *__SIMD32(pSrc)++;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+ /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
+ ** No loop unrolling is used. */
+ blkCnt = blockSize % 0x4u;
+
+#else
+
+ /* Run the below code for Cortex-M0 */
+
+ /* Loop over blockSize number of values */
+ blkCnt = blockSize;
+
+#endif /* #ifndef ARM_MATH_CM0 */
+
+
+ while(blkCnt > 0u)
+ {
+ /* C = A */
+ /* Copy and then store the results in the destination buffer */
+ *pDst++ = *pSrc++;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+}
+
+/**
+ * @} end of BasicCopy group
+ */
diff --git a/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_fill_f32.c b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_fill_f32.c new file mode 100644 index 000000000..439e60db0 --- /dev/null +++ b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_fill_f32.c @@ -0,0 +1,129 @@ +/* ----------------------------------------------------------------------
+* Copyright (C) 2010 ARM Limited. All rights reserved.
+*
+* $Date: 15. February 2012
+* $Revision: V1.1.0
+*
+* Project: CMSIS DSP Library
+* Title: arm_fill_f32.c
+*
+* Description: Fills a constant value into a floating-point vector.
+*
+* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
+*
+* Version 1.1.0 2012/02/15
+* Updated with more optimizations, bug fixes and minor API changes.
+*
+* Version 1.0.10 2011/7/15
+* Big Endian support added and Merged M0 and M3/M4 Source code.
+*
+* Version 1.0.3 2010/11/29
+* Re-organized the CMSIS folders and updated documentation.
+*
+* Version 1.0.2 2010/11/11
+* Documentation updated.
+*
+* Version 1.0.1 2010/10/05
+* Production release and review comments incorporated.
+*
+* Version 1.0.0 2010/09/20
+* Production release and review comments incorporated.
+*
+* Version 0.0.7 2010/06/10
+* Misra-C changes done
+* ---------------------------------------------------------------------------- */
+
+#include "arm_math.h"
+
+/**
+ * @ingroup groupSupport
+ */
+
+/**
+ * @defgroup Fill Vector Fill
+ *
+ * Fills the destination vector with a constant value.
+ *
+ * <pre>
+ * pDst[n] = value; 0 <= n < blockSize.
+ * </pre>
+ *
+ * There are separate functions for floating point, Q31, Q15, and Q7 data types.
+ */
+
+/**
+ * @addtogroup Fill
+ * @{
+ */
+
+/**
+ * @brief Fills a constant value into a floating-point vector.
+ * @param[in] value input value to be filled
+ * @param[out] *pDst points to output vector
+ * @param[in] blockSize length of the output vector
+ * @return none.
+ *
+ */
+
+
+void arm_fill_f32(
+ float32_t value,
+ float32_t * pDst,
+ uint32_t blockSize)
+{
+ uint32_t blkCnt; /* loop counter */
+
+#ifndef ARM_MATH_CM0
+
+ /* Run the below code for Cortex-M4 and Cortex-M3 */
+ float32_t in1 = value;
+ float32_t in2 = value;
+ float32_t in3 = value;
+ float32_t in4 = value;
+
+ /*loop Unrolling */
+ blkCnt = blockSize >> 2u;
+
+ /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
+ ** a second loop below computes the remaining 1 to 3 samples. */
+ while(blkCnt > 0u)
+ {
+ /* C = value */
+ /* Fill the value in the destination buffer */
+ *pDst++ = in1;
+ *pDst++ = in2;
+ *pDst++ = in3;
+ *pDst++ = in4;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+ /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
+ ** No loop unrolling is used. */
+ blkCnt = blockSize % 0x4u;
+
+#else
+
+ /* Run the below code for Cortex-M0 */
+
+ /* Loop over blockSize number of values */
+ blkCnt = blockSize;
+
+#endif /* #ifndef ARM_MATH_CM0 */
+
+
+ while(blkCnt > 0u)
+ {
+ /* C = value */
+ /* Fill the value in the destination buffer */
+ *pDst++ = value;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+}
+
+/**
+ * @} end of Fill group
+ */
diff --git a/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_fill_q15.c b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_fill_q15.c new file mode 100644 index 000000000..8f60d3b4e --- /dev/null +++ b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_fill_q15.c @@ -0,0 +1,115 @@ +/* ----------------------------------------------------------------------
+* Copyright (C) 2010 ARM Limited. All rights reserved.
+*
+* $Date: 15. February 2012
+* $Revision: V1.1.0
+*
+* Project: CMSIS DSP Library
+* Title: arm_fill_q15.c
+*
+* Description: Fills a constant value into a Q15 vector.
+*
+* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
+*
+* Version 1.1.0 2012/02/15
+* Updated with more optimizations, bug fixes and minor API changes.
+*
+* Version 1.0.10 2011/7/15
+* Big Endian support added and Merged M0 and M3/M4 Source code.
+*
+* Version 1.0.3 2010/11/29
+* Re-organized the CMSIS folders and updated documentation.
+*
+* Version 1.0.2 2010/11/11
+* Documentation updated.
+*
+* Version 1.0.1 2010/10/05
+* Production release and review comments incorporated.
+*
+* Version 1.0.0 2010/09/20
+* Production release and review comments incorporated.
+*
+* Version 0.0.7 2010/06/10
+* Misra-C changes done
+* -------------------------------------------------------------------- */
+
+#include "arm_math.h"
+
+/**
+ * @ingroup groupSupport
+ */
+
+/**
+ * @addtogroup Fill
+ * @{
+ */
+
+/**
+ * @brief Fills a constant value into a Q15 vector.
+ * @param[in] value input value to be filled
+ * @param[out] *pDst points to output vector
+ * @param[in] blockSize length of the output vector
+ * @return none.
+ *
+ */
+
+void arm_fill_q15(
+ q15_t value,
+ q15_t * pDst,
+ uint32_t blockSize)
+{
+ uint32_t blkCnt; /* loop counter */
+
+#ifndef ARM_MATH_CM0
+
+ /* Run the below code for Cortex-M4 and Cortex-M3 */
+
+ q31_t packedValue; /* value packed to 32 bits */
+
+
+ /*loop Unrolling */
+ blkCnt = blockSize >> 2u;
+
+ /* Packing two 16 bit values to 32 bit value in order to use SIMD */
+ packedValue = __PKHBT(value, value, 16u);
+
+ /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
+ ** a second loop below computes the remaining 1 to 3 samples. */
+ while(blkCnt > 0u)
+ {
+ /* C = value */
+ /* Fill the value in the destination buffer */
+ *__SIMD32(pDst)++ = packedValue;
+ *__SIMD32(pDst)++ = packedValue;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+ /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
+ ** No loop unrolling is used. */
+ blkCnt = blockSize % 0x4u;
+
+#else
+
+ /* Run the below code for Cortex-M0 */
+
+ /* Loop over blockSize number of values */
+ blkCnt = blockSize;
+
+#endif /* #ifndef ARM_MATH_CM0 */
+
+ while(blkCnt > 0u)
+ {
+ /* C = value */
+ /* Fill the value in the destination buffer */
+ *pDst++ = value;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+}
+
+/**
+ * @} end of Fill group
+ */
diff --git a/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_fill_q31.c b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_fill_q31.c new file mode 100644 index 000000000..35565c3cf --- /dev/null +++ b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_fill_q31.c @@ -0,0 +1,116 @@ +/* ----------------------------------------------------------------------
+* Copyright (C) 2010 ARM Limited. All rights reserved.
+*
+* $Date: 15. February 2012
+* $Revision: V1.1.0
+*
+* Project: CMSIS DSP Library
+* Title: arm_fill_q31.c
+*
+* Description: Fills a constant value into a Q31 vector.
+*
+* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
+*
+* Version 1.1.0 2012/02/15
+* Updated with more optimizations, bug fixes and minor API changes.
+*
+* Version 1.0.10 2011/7/15
+* Big Endian support added and Merged M0 and M3/M4 Source code.
+*
+* Version 1.0.3 2010/11/29
+* Re-organized the CMSIS folders and updated documentation.
+*
+* Version 1.0.2 2010/11/11
+* Documentation updated.
+*
+* Version 1.0.1 2010/10/05
+* Production release and review comments incorporated.
+*
+* Version 1.0.0 2010/09/20
+* Production release and review comments incorporated.
+*
+* Version 0.0.7 2010/06/10
+* Misra-C changes done
+* -------------------------------------------------------------------- */
+
+#include "arm_math.h"
+
+/**
+ * @ingroup groupSupport
+ */
+
+/**
+ * @addtogroup Fill
+ * @{
+ */
+
+/**
+ * @brief Fills a constant value into a Q31 vector.
+ * @param[in] value input value to be filled
+ * @param[out] *pDst points to output vector
+ * @param[in] blockSize length of the output vector
+ * @return none.
+ *
+ */
+
+void arm_fill_q31(
+ q31_t value,
+ q31_t * pDst,
+ uint32_t blockSize)
+{
+ uint32_t blkCnt; /* loop counter */
+
+
+#ifndef ARM_MATH_CM0
+
+ /* Run the below code for Cortex-M4 and Cortex-M3 */
+ q31_t in1 = value;
+ q31_t in2 = value;
+ q31_t in3 = value;
+ q31_t in4 = value;
+
+ /*loop Unrolling */
+ blkCnt = blockSize >> 2u;
+
+ /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
+ ** a second loop below computes the remaining 1 to 3 samples. */
+ while(blkCnt > 0u)
+ {
+ /* C = value */
+ /* Fill the value in the destination buffer */
+ *pDst++ = in1;
+ *pDst++ = in2;
+ *pDst++ = in3;
+ *pDst++ = in4;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+ /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
+ ** No loop unrolling is used. */
+ blkCnt = blockSize % 0x4u;
+
+#else
+
+ /* Run the below code for Cortex-M0 */
+
+ /* Loop over blockSize number of values */
+ blkCnt = blockSize;
+
+#endif /* #ifndef ARM_MATH_CM0 */
+
+ while(blkCnt > 0u)
+ {
+ /* C = value */
+ /* Fill the value in the destination buffer */
+ *pDst++ = value;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+}
+
+/**
+ * @} end of Fill group
+ */
diff --git a/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_fill_q7.c b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_fill_q7.c new file mode 100644 index 000000000..481d4c8f7 --- /dev/null +++ b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_fill_q7.c @@ -0,0 +1,113 @@ +/* ----------------------------------------------------------------------
+* Copyright (C) 2010 ARM Limited. All rights reserved.
+*
+* $Date: 15. February 2012
+* $Revision: V1.1.0
+*
+* Project: CMSIS DSP Library
+* Title: arm_fill_q7.c
+*
+* Description: Fills a constant value into a Q7 vector.
+*
+* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
+*
+* Version 1.1.0 2012/02/15
+* Updated with more optimizations, bug fixes and minor API changes.
+*
+* Version 1.0.10 2011/7/15
+* Big Endian support added and Merged M0 and M3/M4 Source code.
+*
+* Version 1.0.3 2010/11/29
+* Re-organized the CMSIS folders and updated documentation.
+*
+* Version 1.0.2 2010/11/11
+* Documentation updated.
+*
+* Version 1.0.1 2010/10/05
+* Production release and review comments incorporated.
+*
+* Version 1.0.0 2010/09/20
+* Production release and review comments incorporated.
+*
+* Version 0.0.7 2010/06/10
+* Misra-C changes done
+* -------------------------------------------------------------------- */
+
+#include "arm_math.h"
+
+/**
+ * @ingroup groupSupport
+ */
+
+/**
+ * @addtogroup Fill
+ * @{
+ */
+
+/**
+ * @brief Fills a constant value into a Q7 vector.
+ * @param[in] value input value to be filled
+ * @param[out] *pDst points to output vector
+ * @param[in] blockSize length of the output vector
+ * @return none.
+ *
+ */
+
+void arm_fill_q7(
+ q7_t value,
+ q7_t * pDst,
+ uint32_t blockSize)
+{
+ uint32_t blkCnt; /* loop counter */
+
+#ifndef ARM_MATH_CM0
+
+ /* Run the below code for Cortex-M4 and Cortex-M3 */
+
+ q31_t packedValue; /* value packed to 32 bits */
+
+ /*loop Unrolling */
+ blkCnt = blockSize >> 2u;
+
+ /* Packing four 8 bit values to 32 bit value in order to use SIMD */
+ packedValue = __PACKq7(value, value, value, value);
+
+ /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
+ ** a second loop below computes the remaining 1 to 3 samples. */
+ while(blkCnt > 0u)
+ {
+ /* C = value */
+ /* Fill the value in the destination buffer */
+ *__SIMD32(pDst)++ = packedValue;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+ /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
+ ** No loop unrolling is used. */
+ blkCnt = blockSize % 0x4u;
+
+#else
+
+ /* Run the below code for Cortex-M0 */
+
+ /* Loop over blockSize number of values */
+ blkCnt = blockSize;
+
+#endif /* #ifndef ARM_MATH_CM0 */
+
+ while(blkCnt > 0u)
+ {
+ /* C = value */
+ /* Fill the value in the destination buffer */
+ *pDst++ = value;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+}
+
+/**
+ * @} end of Fill group
+ */
diff --git a/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_float_to_q15.c b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_float_to_q15.c new file mode 100644 index 000000000..1537b93ab --- /dev/null +++ b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_float_to_q15.c @@ -0,0 +1,196 @@ +/* ----------------------------------------------------------------------------
+* Copyright (C) 2010 ARM Limited. All rights reserved.
+*
+* $Date: 15. February 2012
+* $Revision: V1.1.0
+*
+* Project: CMSIS DSP Library
+* Title: arm_float_to_q15.c
+*
+* Description: Converts the elements of the floating-point vector to Q15 vector.
+*
+* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
+*
+* Version 1.1.0 2012/02/15
+* Updated with more optimizations, bug fixes and minor API changes.
+*
+* Version 1.0.10 2011/7/15
+* Big Endian support added and Merged M0 and M3/M4 Source code.
+*
+* Version 1.0.3 2010/11/29
+* Re-organized the CMSIS folders and updated documentation.
+*
+* Version 1.0.2 2010/11/11
+* Documentation updated.
+*
+* Version 1.0.1 2010/10/05
+* Production release and review comments incorporated.
+*
+* Version 1.0.0 2010/09/20
+* Production release and review comments incorporated.
+* ---------------------------------------------------------------------------- */
+
+#include "arm_math.h"
+
+/**
+ * @ingroup groupSupport
+ */
+
+/**
+ * @addtogroup float_to_x
+ * @{
+ */
+
+/**
+ * @brief Converts the elements of the floating-point vector to Q15 vector.
+ * @param[in] *pSrc points to the floating-point input vector
+ * @param[out] *pDst points to the Q15 output vector
+ * @param[in] blockSize length of the input vector
+ * @return none.
+ *
+ * \par Description:
+ * \par
+ * The equation used for the conversion process is:
+ * <pre>
+ * pDst[n] = (q15_t)(pSrc[n] * 32768); 0 <= n < blockSize.
+ * </pre>
+ * \par Scaling and Overflow Behavior:
+ * \par
+ * The function uses saturating arithmetic.
+ * Results outside of the allowable Q15 range [0x8000 0x7FFF] will be saturated.
+ * \note
+ * In order to apply rounding, the library should be rebuilt with the ROUNDING macro
+ * defined in the preprocessor section of project options.
+ *
+ */
+
+
+void arm_float_to_q15(
+ float32_t * pSrc,
+ q15_t * pDst,
+ uint32_t blockSize)
+{
+ float32_t *pIn = pSrc; /* Src pointer */
+ uint32_t blkCnt; /* loop counter */
+
+#ifdef ARM_MATH_ROUNDING
+
+ float32_t in;
+
+#endif /* #ifdef ARM_MATH_ROUNDING */
+
+#ifndef ARM_MATH_CM0
+
+ /* Run the below code for Cortex-M4 and Cortex-M3 */
+
+ /*loop Unrolling */
+ blkCnt = blockSize >> 2u;
+
+ /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
+ ** a second loop below computes the remaining 1 to 3 samples. */
+ while(blkCnt > 0u)
+ {
+
+#ifdef ARM_MATH_ROUNDING
+ /* C = A * 32768 */
+ /* convert from float to q15 and then store the results in the destination buffer */
+ in = *pIn++;
+ in = (in * 32768.0f);
+ in += in > 0 ? 0.5 : -0.5;
+ *pDst++ = (q15_t) (__SSAT((q31_t) (in), 16));
+
+ in = *pIn++;
+ in = (in * 32768.0f);
+ in += in > 0 ? 0.5 : -0.5;
+ *pDst++ = (q15_t) (__SSAT((q31_t) (in), 16));
+
+ in = *pIn++;
+ in = (in * 32768.0f);
+ in += in > 0 ? 0.5 : -0.5;
+ *pDst++ = (q15_t) (__SSAT((q31_t) (in), 16));
+
+ in = *pIn++;
+ in = (in * 32768.0f);
+ in += in > 0 ? 0.5 : -0.5;
+ *pDst++ = (q15_t) (__SSAT((q31_t) (in), 16));
+
+#else
+
+ /* C = A * 32768 */
+ /* convert from float to q15 and then store the results in the destination buffer */
+ *pDst++ = (q15_t) __SSAT((q31_t) (*pIn++ * 32768.0f), 16);
+ *pDst++ = (q15_t) __SSAT((q31_t) (*pIn++ * 32768.0f), 16);
+ *pDst++ = (q15_t) __SSAT((q31_t) (*pIn++ * 32768.0f), 16);
+ *pDst++ = (q15_t) __SSAT((q31_t) (*pIn++ * 32768.0f), 16);
+
+#endif /* #ifdef ARM_MATH_ROUNDING */
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+ /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
+ ** No loop unrolling is used. */
+ blkCnt = blockSize % 0x4u;
+
+ while(blkCnt > 0u)
+ {
+
+#ifdef ARM_MATH_ROUNDING
+ /* C = A * 32768 */
+ /* convert from float to q15 and then store the results in the destination buffer */
+ in = *pIn++;
+ in = (in * 32768.0f);
+ in += in > 0 ? 0.5 : -0.5;
+ *pDst++ = (q15_t) (__SSAT((q31_t) (in), 16));
+
+#else
+
+ /* C = A * 32768 */
+ /* convert from float to q15 and then store the results in the destination buffer */
+ *pDst++ = (q15_t) __SSAT((q31_t) (*pIn++ * 32768.0f), 16);
+
+#endif /* #ifdef ARM_MATH_ROUNDING */
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+
+#else
+
+ /* Run the below code for Cortex-M0 */
+
+ /* Loop over blockSize number of values */
+ blkCnt = blockSize;
+
+ while(blkCnt > 0u)
+ {
+
+#ifdef ARM_MATH_ROUNDING
+ /* C = A * 32768 */
+ /* convert from float to q15 and then store the results in the destination buffer */
+ in = *pIn++;
+ in = (in * 32768.0f);
+ in += in > 0 ? 0.5f : -0.5f;
+ *pDst++ = (q15_t) (__SSAT((q31_t) (in), 16));
+
+#else
+
+ /* C = A * 32768 */
+ /* convert from float to q15 and then store the results in the destination buffer */
+ *pDst++ = (q15_t) __SSAT((q31_t) (*pIn++ * 32768.0f), 16);
+
+#endif /* #ifdef ARM_MATH_ROUNDING */
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+#endif /* #ifndef ARM_MATH_CM0 */
+
+}
+
+/**
+ * @} end of float_to_x group
+ */
diff --git a/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_float_to_q31.c b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_float_to_q31.c new file mode 100644 index 000000000..3ab52b9ff --- /dev/null +++ b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_float_to_q31.c @@ -0,0 +1,203 @@ +/* ----------------------------------------------------------------------------
+* Copyright (C) 2010 ARM Limited. All rights reserved.
+*
+* $Date: 15. February 2012
+* $Revision: V1.1.0
+*
+* Project: CMSIS DSP Library
+* Title: arm_float_to_q31.c
+*
+* Description: Converts the elements of the floating-point vector to Q31 vector.
+*
+* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
+*
+* Version 1.1.0 2012/02/15
+* Updated with more optimizations, bug fixes and minor API changes.
+*
+* Version 1.0.10 2011/7/15
+* Big Endian support added and Merged M0 and M3/M4 Source code.
+*
+* Version 1.0.3 2010/11/29
+* Re-organized the CMSIS folders and updated documentation.
+*
+* Version 1.0.2 2010/11/11
+* Documentation updated.
+*
+* Version 1.0.1 2010/10/05
+* Production release and review comments incorporated.
+*
+* Version 1.0.0 2010/09/20
+* Production release and review comments incorporated.
+* ---------------------------------------------------------------------------- */
+
+#include "arm_math.h"
+
+/**
+ * @ingroup groupSupport
+ */
+
+/**
+ * @defgroup float_to_x Convert 32-bit floating point value
+ */
+
+/**
+ * @addtogroup float_to_x
+ * @{
+ */
+
+/**
+ * @brief Converts the elements of the floating-point vector to Q31 vector.
+ * @param[in] *pSrc points to the floating-point input vector
+ * @param[out] *pDst points to the Q31 output vector
+ * @param[in] blockSize length of the input vector
+ * @return none.
+ *
+ *\par Description:
+ * \par
+ * The equation used for the conversion process is:
+ *
+ * <pre>
+ * pDst[n] = (q31_t)(pSrc[n] * 2147483648); 0 <= n < blockSize.
+ * </pre>
+ * <b>Scaling and Overflow Behavior:</b>
+ * \par
+ * The function uses saturating arithmetic.
+ * Results outside of the allowable Q31 range[0x80000000 0x7FFFFFFF] will be saturated.
+ *
+ * \note In order to apply rounding, the library should be rebuilt with the ROUNDING macro
+ * defined in the preprocessor section of project options.
+ */
+
+
+void arm_float_to_q31(
+ float32_t * pSrc,
+ q31_t * pDst,
+ uint32_t blockSize)
+{
+ float32_t *pIn = pSrc; /* Src pointer */
+ uint32_t blkCnt; /* loop counter */
+
+#ifdef ARM_MATH_ROUNDING
+
+ float32_t in;
+
+#endif /* #ifdef ARM_MATH_ROUNDING */
+
+#ifndef ARM_MATH_CM0
+
+ /* Run the below code for Cortex-M4 and Cortex-M3 */
+
+ /*loop Unrolling */
+ blkCnt = blockSize >> 2u;
+
+ /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
+ ** a second loop below computes the remaining 1 to 3 samples. */
+ while(blkCnt > 0u)
+ {
+
+#ifdef ARM_MATH_ROUNDING
+
+ /* C = A * 32768 */
+ /* convert from float to Q31 and then store the results in the destination buffer */
+ in = *pIn++;
+ in = (in * 2147483648.0f);
+ in += in > 0 ? 0.5 : -0.5;
+ *pDst++ = clip_q63_to_q31((q63_t) (in));
+
+ in = *pIn++;
+ in = (in * 2147483648.0f);
+ in += in > 0 ? 0.5 : -0.5;
+ *pDst++ = clip_q63_to_q31((q63_t) (in));
+
+ in = *pIn++;
+ in = (in * 2147483648.0f);
+ in += in > 0 ? 0.5 : -0.5;
+ *pDst++ = clip_q63_to_q31((q63_t) (in));
+
+ in = *pIn++;
+ in = (in * 2147483648.0f);
+ in += in > 0 ? 0.5 : -0.5;
+ *pDst++ = clip_q63_to_q31((q63_t) (in));
+
+#else
+
+ /* C = A * 2147483648 */
+ /* convert from float to Q31 and then store the results in the destination buffer */
+ *pDst++ = clip_q63_to_q31((q63_t) (*pIn++ * 2147483648.0f));
+ *pDst++ = clip_q63_to_q31((q63_t) (*pIn++ * 2147483648.0f));
+ *pDst++ = clip_q63_to_q31((q63_t) (*pIn++ * 2147483648.0f));
+ *pDst++ = clip_q63_to_q31((q63_t) (*pIn++ * 2147483648.0f));
+
+#endif /* #ifdef ARM_MATH_ROUNDING */
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+ /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
+ ** No loop unrolling is used. */
+ blkCnt = blockSize % 0x4u;
+
+ while(blkCnt > 0u)
+ {
+
+#ifdef ARM_MATH_ROUNDING
+
+ /* C = A * 2147483648 */
+ /* convert from float to Q31 and then store the results in the destination buffer */
+ in = *pIn++;
+ in = (in * 2147483648.0f);
+ in += in > 0 ? 0.5 : -0.5;
+ *pDst++ = clip_q63_to_q31((q63_t) (in));
+
+#else
+
+ /* C = A * 2147483648 */
+ /* convert from float to Q31 and then store the results in the destination buffer */
+ *pDst++ = clip_q63_to_q31((q63_t) (*pIn++ * 2147483648.0f));
+
+#endif /* #ifdef ARM_MATH_ROUNDING */
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+
+#else
+
+ /* Run the below code for Cortex-M0 */
+
+ /* Loop over blockSize number of values */
+ blkCnt = blockSize;
+
+ while(blkCnt > 0u)
+ {
+
+#ifdef ARM_MATH_ROUNDING
+
+ /* C = A * 2147483648 */
+ /* convert from float to Q31 and then store the results in the destination buffer */
+ in = *pIn++;
+ in = (in * 2147483648.0f);
+ in += in > 0 ? 0.5f : -0.5f;
+ *pDst++ = clip_q63_to_q31((q63_t) (in));
+
+#else
+
+ /* C = A * 2147483648 */
+ /* convert from float to Q31 and then store the results in the destination buffer */
+ *pDst++ = clip_q63_to_q31((q63_t) (*pIn++ * 2147483648.0f));
+
+#endif /* #ifdef ARM_MATH_ROUNDING */
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+#endif /* #ifndef ARM_MATH_CM0 */
+
+}
+
+/**
+ * @} end of float_to_x group
+ */
diff --git a/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_float_to_q7.c b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_float_to_q7.c new file mode 100644 index 000000000..dea14c6a1 --- /dev/null +++ b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_float_to_q7.c @@ -0,0 +1,195 @@ +/* ----------------------------------------------------------------------------
+* Copyright (C) 2010 ARM Limited. All rights reserved.
+*
+* $Date: 15. February 2012
+* $Revision: V1.1.0
+*
+* Project: CMSIS DSP Library
+* Title: arm_float_to_q7.c
+*
+* Description: Converts the elements of the floating-point vector to Q7 vector.
+*
+* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
+*
+* Version 1.1.0 2012/02/15
+* Updated with more optimizations, bug fixes and minor API changes.
+*
+* Version 1.0.10 2011/7/15
+* Big Endian support added and Merged M0 and M3/M4 Source code.
+*
+* Version 1.0.3 2010/11/29
+* Re-organized the CMSIS folders and updated documentation.
+*
+* Version 1.0.2 2010/11/11
+* Documentation updated.
+*
+* Version 1.0.1 2010/10/05
+* Production release and review comments incorporated.
+*
+* Version 1.0.0 2010/09/20
+* Production release and review comments incorporated.
+* ---------------------------------------------------------------------------- */
+
+#include "arm_math.h"
+
+/**
+ * @ingroup groupSupport
+ */
+
+/**
+ * @addtogroup float_to_x
+ * @{
+ */
+
+/**
+ * @brief Converts the elements of the floating-point vector to Q7 vector.
+ * @param[in] *pSrc points to the floating-point input vector
+ * @param[out] *pDst points to the Q7 output vector
+ * @param[in] blockSize length of the input vector
+ * @return none.
+ *
+ *\par Description:
+ * \par
+ * The equation used for the conversion process is:
+ * <pre>
+ * pDst[n] = (q7_t)(pSrc[n] * 128); 0 <= n < blockSize.
+ * </pre>
+ * \par Scaling and Overflow Behavior:
+ * \par
+ * The function uses saturating arithmetic.
+ * Results outside of the allowable Q7 range [0x80 0x7F] will be saturated.
+ * \note
+ * In order to apply rounding, the library should be rebuilt with the ROUNDING macro
+ * defined in the preprocessor section of project options.
+ */
+
+
+void arm_float_to_q7(
+ float32_t * pSrc,
+ q7_t * pDst,
+ uint32_t blockSize)
+{
+ float32_t *pIn = pSrc; /* Src pointer */
+ uint32_t blkCnt; /* loop counter */
+
+#ifdef ARM_MATH_ROUNDING
+
+ float32_t in;
+
+#endif /* #ifdef ARM_MATH_ROUNDING */
+
+#ifndef ARM_MATH_CM0
+
+ /* Run the below code for Cortex-M4 and Cortex-M3 */
+
+ /*loop Unrolling */
+ blkCnt = blockSize >> 2u;
+
+ /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
+ ** a second loop below computes the remaining 1 to 3 samples. */
+ while(blkCnt > 0u)
+ {
+
+#ifdef ARM_MATH_ROUNDING
+ /* C = A * 128 */
+ /* convert from float to q7 and then store the results in the destination buffer */
+ in = *pIn++;
+ in = (in * 128);
+ in += in > 0 ? 0.5 : -0.5;
+ *pDst++ = (q7_t) (__SSAT((q15_t) (in), 8));
+
+ in = *pIn++;
+ in = (in * 128);
+ in += in > 0 ? 0.5 : -0.5;
+ *pDst++ = (q7_t) (__SSAT((q15_t) (in), 8));
+
+ in = *pIn++;
+ in = (in * 128);
+ in += in > 0 ? 0.5 : -0.5;
+ *pDst++ = (q7_t) (__SSAT((q15_t) (in), 8));
+
+ in = *pIn++;
+ in = (in * 128);
+ in += in > 0 ? 0.5 : -0.5;
+ *pDst++ = (q7_t) (__SSAT((q15_t) (in), 8));
+
+#else
+
+ /* C = A * 128 */
+ /* convert from float to q7 and then store the results in the destination buffer */
+ *pDst++ = __SSAT((q31_t) (*pIn++ * 128.0f), 8);
+ *pDst++ = __SSAT((q31_t) (*pIn++ * 128.0f), 8);
+ *pDst++ = __SSAT((q31_t) (*pIn++ * 128.0f), 8);
+ *pDst++ = __SSAT((q31_t) (*pIn++ * 128.0f), 8);
+
+#endif /* #ifdef ARM_MATH_ROUNDING */
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+ /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
+ ** No loop unrolling is used. */
+ blkCnt = blockSize % 0x4u;
+
+ while(blkCnt > 0u)
+ {
+
+#ifdef ARM_MATH_ROUNDING
+ /* C = A * 128 */
+ /* convert from float to q7 and then store the results in the destination buffer */
+ in = *pIn++;
+ in = (in * 128);
+ in += in > 0 ? 0.5 : -0.5;
+ *pDst++ = (q7_t) (__SSAT((q15_t) (in), 8));
+
+#else
+
+ /* C = A * 128 */
+ /* convert from float to q7 and then store the results in the destination buffer */
+ *pDst++ = __SSAT((q31_t) (*pIn++ * 128.0f), 8);
+
+#endif /* #ifdef ARM_MATH_ROUNDING */
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+
+#else
+
+ /* Run the below code for Cortex-M0 */
+
+
+ /* Loop over blockSize number of values */
+ blkCnt = blockSize;
+
+ while(blkCnt > 0u)
+ {
+#ifdef ARM_MATH_ROUNDING
+ /* C = A * 128 */
+ /* convert from float to q7 and then store the results in the destination buffer */
+ in = *pIn++;
+ in = (in * 128.0f);
+ in += in > 0 ? 0.5f : -0.5f;
+ *pDst++ = (q7_t) (__SSAT((q31_t) (in), 8));
+
+#else
+
+ /* C = A * 128 */
+ /* convert from float to q7 and then store the results in the destination buffer */
+ *pDst++ = (q7_t) __SSAT((q31_t) (*pIn++ * 128.0f), 8);
+
+#endif /* #ifdef ARM_MATH_ROUNDING */
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+#endif /* #ifndef ARM_MATH_CM0 */
+
+}
+
+/**
+ * @} end of float_to_x group
+ */
diff --git a/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q15_to_float.c b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q15_to_float.c new file mode 100644 index 000000000..ce51e9a39 --- /dev/null +++ b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q15_to_float.c @@ -0,0 +1,126 @@ +/* ----------------------------------------------------------------------------
+* Copyright (C) 2010 ARM Limited. All rights reserved.
+*
+* $Date: 15. February 2012
+* $Revision: V1.1.0
+*
+* Project: CMSIS DSP Library
+* Title: arm_q15_to_float.c
+*
+* Description: Converts the elements of the Q15 vector to floating-point vector.
+*
+* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
+*
+* Version 1.1.0 2012/02/15
+* Updated with more optimizations, bug fixes and minor API changes.
+*
+* Version 1.0.10 2011/7/15
+* Big Endian support added and Merged M0 and M3/M4 Source code.
+*
+* Version 1.0.3 2010/11/29
+* Re-organized the CMSIS folders and updated documentation.
+*
+* Version 1.0.2 2010/11/11
+* Documentation updated.
+*
+* Version 1.0.1 2010/10/05
+* Production release and review comments incorporated.
+*
+* Version 1.0.0 2010/09/20
+* Production release and review comments incorporated.
+* ---------------------------------------------------------------------------- */
+
+#include "arm_math.h"
+
+/**
+ * @ingroup groupSupport
+ */
+
+/**
+ * @defgroup q15_to_x Convert 16-bit Integer value
+ */
+
+/**
+ * @addtogroup q15_to_x
+ * @{
+ */
+
+
+
+
+/**
+ * @brief Converts the elements of the Q15 vector to floating-point vector.
+ * @param[in] *pSrc points to the Q15 input vector
+ * @param[out] *pDst points to the floating-point output vector
+ * @param[in] blockSize length of the input vector
+ * @return none.
+ *
+ * \par Description:
+ *
+ * The equation used for the conversion process is:
+ *
+ * <pre>
+ * pDst[n] = (float32_t) pSrc[n] / 32768; 0 <= n < blockSize.
+ * </pre>
+ *
+ */
+
+
+void arm_q15_to_float(
+ q15_t * pSrc,
+ float32_t * pDst,
+ uint32_t blockSize)
+{
+ q15_t *pIn = pSrc; /* Src pointer */
+ uint32_t blkCnt; /* loop counter */
+
+
+#ifndef ARM_MATH_CM0
+
+ /* Run the below code for Cortex-M4 and Cortex-M3 */
+
+ /*loop Unrolling */
+ blkCnt = blockSize >> 2u;
+
+ /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
+ ** a second loop below computes the remaining 1 to 3 samples. */
+ while(blkCnt > 0u)
+ {
+ /* C = (float32_t) A / 32768 */
+ /* convert from q15 to float and then store the results in the destination buffer */
+ *pDst++ = ((float32_t) * pIn++ / 32768.0f);
+ *pDst++ = ((float32_t) * pIn++ / 32768.0f);
+ *pDst++ = ((float32_t) * pIn++ / 32768.0f);
+ *pDst++ = ((float32_t) * pIn++ / 32768.0f);
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+ /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
+ ** No loop unrolling is used. */
+ blkCnt = blockSize % 0x4u;
+
+#else
+
+ /* Run the below code for Cortex-M0 */
+
+ /* Loop over blockSize number of values */
+ blkCnt = blockSize;
+
+#endif /* #ifndef ARM_MATH_CM0 */
+
+ while(blkCnt > 0u)
+ {
+ /* C = (float32_t) A / 32768 */
+ /* convert from q15 to float and then store the results in the destination buffer */
+ *pDst++ = ((float32_t) * pIn++ / 32768.0f);
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+}
+
+/**
+ * @} end of q15_to_x group
+ */
diff --git a/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q15_to_q31.c b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q15_to_q31.c new file mode 100644 index 000000000..a0f66c29e --- /dev/null +++ b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q15_to_q31.c @@ -0,0 +1,148 @@ +/* ----------------------------------------------------------------------------
+* Copyright (C) 2010 ARM Limited. All rights reserved.
+*
+* $Date: 15. February 2012
+* $Revision: V1.1.0
+*
+* Project: CMSIS DSP Library
+* Title: arm_q15_to_q31.c
+*
+* Description: Converts the elements of the Q15 vector to Q31 vector.
+*
+* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
+*
+* Version 1.1.0 2012/02/15
+* Updated with more optimizations, bug fixes and minor API changes.
+*
+* Version 1.0.10 2011/7/15
+* Big Endian support added and Merged M0 and M3/M4 Source code.
+*
+* Version 1.0.3 2010/11/29
+* Re-organized the CMSIS folders and updated documentation.
+*
+* Version 1.0.2 2010/11/11
+* Documentation updated.
+*
+* Version 1.0.1 2010/10/05
+* Production release and review comments incorporated.
+*
+* Version 1.0.0 2010/09/20
+* Production release and review comments incorporated.
+* ---------------------------------------------------------------------------- */
+
+#include "arm_math.h"
+
+/**
+ * @ingroup groupSupport
+ */
+
+/**
+ * @addtogroup q15_to_x
+ * @{
+ */
+
+/**
+ * @brief Converts the elements of the Q15 vector to Q31 vector.
+ * @param[in] *pSrc points to the Q15 input vector
+ * @param[out] *pDst points to the Q31 output vector
+ * @param[in] blockSize length of the input vector
+ * @return none.
+ *
+ * \par Description:
+ *
+ * The equation used for the conversion process is:
+ *
+ * <pre>
+ * pDst[n] = (q31_t) pSrc[n] << 16; 0 <= n < blockSize.
+ * </pre>
+ *
+ */
+
+
+void arm_q15_to_q31(
+ q15_t * pSrc,
+ q31_t * pDst,
+ uint32_t blockSize)
+{
+ q15_t *pIn = pSrc; /* Src pointer */
+ uint32_t blkCnt; /* loop counter */
+
+#ifndef ARM_MATH_CM0
+
+ /* Run the below code for Cortex-M4 and Cortex-M3 */
+ q31_t in1, in2;
+ q31_t out1, out2, out3, out4;
+
+ /*loop Unrolling */
+ blkCnt = blockSize >> 2u;
+
+ /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
+ ** a second loop below computes the remaining 1 to 3 samples. */
+ while(blkCnt > 0u)
+ {
+ /* C = (q31_t)A << 16 */
+ /* convert from q15 to q31 and then store the results in the destination buffer */
+ in1 = *__SIMD32(pIn)++;
+ in2 = *__SIMD32(pIn)++;
+
+#ifndef ARM_MATH_BIG_ENDIAN
+
+ /* extract lower 16 bits to 32 bit result */
+ out1 = in1 << 16u;
+ /* extract upper 16 bits to 32 bit result */
+ out2 = in1 & 0xFFFF0000;
+ /* extract lower 16 bits to 32 bit result */
+ out3 = in2 << 16u;
+ /* extract upper 16 bits to 32 bit result */
+ out4 = in2 & 0xFFFF0000;
+
+#else
+
+ /* extract upper 16 bits to 32 bit result */
+ out1 = in1 & 0xFFFF0000;
+ /* extract lower 16 bits to 32 bit result */
+ out2 = in1 << 16u;
+ /* extract upper 16 bits to 32 bit result */
+ out3 = in2 & 0xFFFF0000;
+ /* extract lower 16 bits to 32 bit result */
+ out4 = in2 << 16u;
+
+#endif // #ifndef ARM_MATH_BIG_ENDIAN
+
+ *pDst++ = out1;
+ *pDst++ = out2;
+ *pDst++ = out3;
+ *pDst++ = out4;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+ /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
+ ** No loop unrolling is used. */
+ blkCnt = blockSize % 0x4u;
+
+#else
+
+ /* Run the below code for Cortex-M0 */
+
+ /* Loop over blockSize number of values */
+ blkCnt = blockSize;
+
+#endif /* #ifndef ARM_MATH_CM0 */
+
+ while(blkCnt > 0u)
+ {
+ /* C = (q31_t)A << 16 */
+ /* convert from q15 to q31 and then store the results in the destination buffer */
+ *pDst++ = (q31_t) * pIn++ << 16;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+}
+
+/**
+ * @} end of q15_to_x group
+ */
diff --git a/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q15_to_q7.c b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q15_to_q7.c new file mode 100644 index 000000000..87fe63d9e --- /dev/null +++ b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q15_to_q7.c @@ -0,0 +1,146 @@ +/* ----------------------------------------------------------------------------
+* Copyright (C) 2010 ARM Limited. All rights reserved.
+*
+* $Date: 15. February 2012
+* $Revision: V1.1.0
+*
+* Project: CMSIS DSP Library
+* Title: arm_q15_to_q7.c
+*
+* Description: Converts the elements of the Q15 vector to Q7 vector.
+*
+* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
+*
+* Version 1.1.0 2012/02/15
+* Updated with more optimizations, bug fixes and minor API changes.
+*
+* Version 1.0.10 2011/7/15
+* Big Endian support added and Merged M0 and M3/M4 Source code.
+*
+* Version 1.0.3 2010/11/29
+* Re-organized the CMSIS folders and updated documentation.
+*
+* Version 1.0.2 2010/11/11
+* Documentation updated.
+*
+* Version 1.0.1 2010/10/05
+* Production release and review comments incorporated.
+*
+* Version 1.0.0 2010/09/20
+* Production release and review comments incorporated.
+* ---------------------------------------------------------------------------- */
+
+#include "arm_math.h"
+
+/**
+ * @ingroup groupSupport
+ */
+
+/**
+ * @addtogroup q15_to_x
+ * @{
+ */
+
+
+/**
+ * @brief Converts the elements of the Q15 vector to Q7 vector.
+ * @param[in] *pSrc points to the Q15 input vector
+ * @param[out] *pDst points to the Q7 output vector
+ * @param[in] blockSize length of the input vector
+ * @return none.
+ *
+ * \par Description:
+ *
+ * The equation used for the conversion process is:
+ *
+ * <pre>
+ * pDst[n] = (q7_t) pSrc[n] >> 8; 0 <= n < blockSize.
+ * </pre>
+ *
+ */
+
+
+void arm_q15_to_q7(
+ q15_t * pSrc,
+ q7_t * pDst,
+ uint32_t blockSize)
+{
+ q15_t *pIn = pSrc; /* Src pointer */
+ uint32_t blkCnt; /* loop counter */
+
+#ifndef ARM_MATH_CM0
+
+ /* Run the below code for Cortex-M4 and Cortex-M3 */
+ q31_t in1, in2;
+ q31_t out1, out2;
+
+ /*loop Unrolling */
+ blkCnt = blockSize >> 2u;
+
+ /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
+ ** a second loop below computes the remaining 1 to 3 samples. */
+ while(blkCnt > 0u)
+ {
+ /* C = (q7_t) A >> 8 */
+ /* convert from q15 to q7 and then store the results in the destination buffer */
+ in1 = *__SIMD32(pIn)++;
+ in2 = *__SIMD32(pIn)++;
+
+#ifndef ARM_MATH_BIG_ENDIAN
+
+ out1 = __PKHTB(in2, in1, 16);
+ out2 = __PKHBT(in2, in1, 16);
+
+#else
+
+ out1 = __PKHTB(in1, in2, 16);
+ out2 = __PKHBT(in1, in2, 16);
+
+#endif // #ifndef ARM_MATH_BIG_ENDIAN
+
+ /* rotate packed value by 24 */
+ out2 = ((uint32_t) out2 << 8) | ((uint32_t) out2 >> 24);
+
+ /* anding with 0xff00ff00 to get two 8 bit values */
+ out1 = out1 & 0xFF00FF00;
+ /* anding with 0x00ff00ff to get two 8 bit values */
+ out2 = out2 & 0x00FF00FF;
+
+ /* oring two values(contains two 8 bit values) to get four packed 8 bit values */
+ out1 = out1 | out2;
+
+ /* store 4 samples at a time to destiantion buffer */
+ *__SIMD32(pDst)++ = out1;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+ /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
+ ** No loop unrolling is used. */
+ blkCnt = blockSize % 0x4u;
+
+#else
+
+ /* Run the below code for Cortex-M0 */
+
+ /* Loop over blockSize number of values */
+ blkCnt = blockSize;
+
+#endif /* #ifndef ARM_MATH_CM0 */
+
+ while(blkCnt > 0u)
+ {
+ /* C = (q7_t) A >> 8 */
+ /* convert from q15 to q7 and then store the results in the destination buffer */
+ *pDst++ = (q7_t) (*pIn++ >> 8);
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+}
+
+/**
+ * @} end of q15_to_x group
+ */
diff --git a/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q31_to_float.c b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q31_to_float.c new file mode 100644 index 000000000..94deec9a1 --- /dev/null +++ b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q31_to_float.c @@ -0,0 +1,123 @@ +/* ----------------------------------------------------------------------------
+* Copyright (C) 2010 ARM Limited. All rights reserved.
+*
+* $Date: 15. February 2012
+* $Revision: V1.1.0
+*
+* Project: CMSIS DSP Library
+* Title: arm_q31_to_float.c
+*
+* Description: Converts the elements of the Q31 vector to floating-point vector.
+*
+* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
+*
+* Version 1.1.0 2012/02/15
+* Updated with more optimizations, bug fixes and minor API changes.
+*
+* Version 1.0.10 2011/7/15
+* Big Endian support added and Merged M0 and M3/M4 Source code.
+*
+* Version 1.0.3 2010/11/29
+* Re-organized the CMSIS folders and updated documentation.
+*
+* Version 1.0.2 2010/11/11
+* Documentation updated.
+*
+* Version 1.0.1 2010/10/05
+* Production release and review comments incorporated.
+*
+* Version 1.0.0 2010/09/20
+* Production release and review comments incorporated.
+* ---------------------------------------------------------------------------- */
+
+#include "arm_math.h"
+
+/**
+ * @ingroup groupSupport
+ */
+
+/**
+ * @defgroup q31_to_x Convert 32-bit Integer value
+ */
+
+/**
+ * @addtogroup q31_to_x
+ * @{
+ */
+
+/**
+ * @brief Converts the elements of the Q31 vector to floating-point vector.
+ * @param[in] *pSrc points to the Q31 input vector
+ * @param[out] *pDst points to the floating-point output vector
+ * @param[in] blockSize length of the input vector
+ * @return none.
+ *
+ * \par Description:
+ *
+ * The equation used for the conversion process is:
+ *
+ * <pre>
+ * pDst[n] = (float32_t) pSrc[n] / 2147483648; 0 <= n < blockSize.
+ * </pre>
+ *
+ */
+
+
+void arm_q31_to_float(
+ q31_t * pSrc,
+ float32_t * pDst,
+ uint32_t blockSize)
+{
+ q31_t *pIn = pSrc; /* Src pointer */
+ uint32_t blkCnt; /* loop counter */
+
+
+#ifndef ARM_MATH_CM0
+
+ /* Run the below code for Cortex-M4 and Cortex-M3 */
+
+ /*loop Unrolling */
+ blkCnt = blockSize >> 2u;
+
+ /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
+ ** a second loop below computes the remaining 1 to 3 samples. */
+ while(blkCnt > 0u)
+ {
+ /* C = (float32_t) A / 2147483648 */
+ /* convert from q31 to float and then store the results in the destination buffer */
+ *pDst++ = ((float32_t) * pIn++ / 2147483648.0f);
+ *pDst++ = ((float32_t) * pIn++ / 2147483648.0f);
+ *pDst++ = ((float32_t) * pIn++ / 2147483648.0f);
+ *pDst++ = ((float32_t) * pIn++ / 2147483648.0f);
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+ /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
+ ** No loop unrolling is used. */
+ blkCnt = blockSize % 0x4u;
+
+#else
+
+ /* Run the below code for Cortex-M0 */
+
+ /* Loop over blockSize number of values */
+ blkCnt = blockSize;
+
+#endif /* #ifndef ARM_MATH_CM0 */
+
+ while(blkCnt > 0u)
+ {
+ /* C = (float32_t) A / 2147483648 */
+ /* convert from q31 to float and then store the results in the destination buffer */
+ *pDst++ = ((float32_t) * pIn++ / 2147483648.0f);
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+}
+
+/**
+ * @} end of q31_to_x group
+ */
diff --git a/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q31_to_q15.c b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q31_to_q15.c new file mode 100644 index 000000000..ba79f85a1 --- /dev/null +++ b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q31_to_q15.c @@ -0,0 +1,137 @@ +/* ----------------------------------------------------------------------------
+* Copyright (C) 2010 ARM Limited. All rights reserved.
+*
+* $Date: 15. February 2012
+* $Revision: V1.1.0
+*
+* Project: CMSIS DSP Library
+* Title: arm_q31_to_q15.c
+*
+* Description: Converts the elements of the Q31 vector to Q15 vector.
+*
+* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
+*
+* Version 1.1.0 2012/02/15
+* Updated with more optimizations, bug fixes and minor API changes.
+*
+* Version 1.0.10 2011/7/15
+* Big Endian support added and Merged M0 and M3/M4 Source code.
+*
+* Version 1.0.3 2010/11/29
+* Re-organized the CMSIS folders and updated documentation.
+*
+* Version 1.0.2 2010/11/11
+* Documentation updated.
+*
+* Version 1.0.1 2010/10/05
+* Production release and review comments incorporated.
+*
+* Version 1.0.0 2010/09/20
+* Production release and review comments incorporated.
+* ---------------------------------------------------------------------------- */
+
+#include "arm_math.h"
+
+/**
+ * @ingroup groupSupport
+ */
+
+/**
+ * @addtogroup q31_to_x
+ * @{
+ */
+
+/**
+ * @brief Converts the elements of the Q31 vector to Q15 vector.
+ * @param[in] *pSrc points to the Q31 input vector
+ * @param[out] *pDst points to the Q15 output vector
+ * @param[in] blockSize length of the input vector
+ * @return none.
+ *
+ * \par Description:
+ *
+ * The equation used for the conversion process is:
+ *
+ * <pre>
+ * pDst[n] = (q15_t) pSrc[n] >> 16; 0 <= n < blockSize.
+ * </pre>
+ *
+ */
+
+
+void arm_q31_to_q15(
+ q31_t * pSrc,
+ q15_t * pDst,
+ uint32_t blockSize)
+{
+ q31_t *pIn = pSrc; /* Src pointer */
+ uint32_t blkCnt; /* loop counter */
+
+#ifndef ARM_MATH_CM0
+
+ /* Run the below code for Cortex-M4 and Cortex-M3 */
+ q31_t in1, in2, in3, in4;
+ q31_t out1, out2;
+
+ /*loop Unrolling */
+ blkCnt = blockSize >> 2u;
+
+ /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
+ ** a second loop below computes the remaining 1 to 3 samples. */
+ while(blkCnt > 0u)
+ {
+ /* C = (q15_t) A >> 16 */
+ /* convert from q31 to q15 and then store the results in the destination buffer */
+ in1 = *pIn++;
+ in2 = *pIn++;
+ in3 = *pIn++;
+ in4 = *pIn++;
+
+ /* pack two higher 16-bit values from two 32-bit values */
+#ifndef ARM_MATH_BIG_ENDIAN
+
+ out1 = __PKHTB(in2, in1, 16);
+ out2 = __PKHTB(in4, in3, 16);
+
+#else
+
+ out1 = __PKHTB(in1, in2, 16);
+ out2 = __PKHTB(in3, in4, 16);
+
+#endif // #ifdef ARM_MATH_BIG_ENDIAN
+
+ *__SIMD32(pDst)++ = out1;
+ *__SIMD32(pDst)++ = out2;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+ /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
+ ** No loop unrolling is used. */
+ blkCnt = blockSize % 0x4u;
+
+#else
+
+ /* Run the below code for Cortex-M0 */
+
+ /* Loop over blockSize number of values */
+ blkCnt = blockSize;
+
+#endif /* #ifndef ARM_MATH_CM0 */
+
+ while(blkCnt > 0u)
+ {
+ /* C = (q15_t) A >> 16 */
+ /* convert from q31 to q15 and then store the results in the destination buffer */
+ *pDst++ = (q15_t) (*pIn++ >> 16);
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+}
+
+/**
+ * @} end of q31_to_x group
+ */
diff --git a/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q31_to_q7.c b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q31_to_q7.c new file mode 100644 index 000000000..afc8fdbed --- /dev/null +++ b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q31_to_q7.c @@ -0,0 +1,128 @@ +/* ----------------------------------------------------------------------------
+* Copyright (C) 2010 ARM Limited. All rights reserved.
+*
+* $Date: 15. February 2012
+* $Revision: V1.1.0
+*
+* Project: CMSIS DSP Library
+* Title: arm_q31_to_q7.c
+*
+* Description: Converts the elements of the Q31 vector to Q7 vector.
+*
+* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
+*
+* Version 1.1.0 2012/02/15
+* Updated with more optimizations, bug fixes and minor API changes.
+*
+* Version 1.0.10 2011/7/15
+* Big Endian support added and Merged M0 and M3/M4 Source code.
+*
+* Version 1.0.3 2010/11/29
+* Re-organized the CMSIS folders and updated documentation.
+*
+* Version 1.0.2 2010/11/11
+* Documentation updated.
+*
+* Version 1.0.1 2010/10/05
+* Production release and review comments incorporated.
+*
+* Version 1.0.0 2010/09/20
+* Production release and review comments incorporated.
+* ---------------------------------------------------------------------------- */
+
+#include "arm_math.h"
+
+/**
+ * @ingroup groupSupport
+ */
+
+/**
+ * @addtogroup q31_to_x
+ * @{
+ */
+
+/**
+ * @brief Converts the elements of the Q31 vector to Q7 vector.
+ * @param[in] *pSrc points to the Q31 input vector
+ * @param[out] *pDst points to the Q7 output vector
+ * @param[in] blockSize length of the input vector
+ * @return none.
+ *
+ * \par Description:
+ *
+ * The equation used for the conversion process is:
+ *
+ * <pre>
+ * pDst[n] = (q7_t) pSrc[n] >> 24; 0 <= n < blockSize.
+ * </pre>
+ *
+ */
+
+
+void arm_q31_to_q7(
+ q31_t * pSrc,
+ q7_t * pDst,
+ uint32_t blockSize)
+{
+ q31_t *pIn = pSrc; /* Src pointer */
+ uint32_t blkCnt; /* loop counter */
+
+#ifndef ARM_MATH_CM0
+
+ /* Run the below code for Cortex-M4 and Cortex-M3 */
+ q31_t in1, in2, in3, in4;
+ q7_t out1, out2, out3, out4;
+
+ /*loop Unrolling */
+ blkCnt = blockSize >> 2u;
+
+ /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
+ ** a second loop below computes the remaining 1 to 3 samples. */
+ while(blkCnt > 0u)
+ {
+ /* C = (q7_t) A >> 24 */
+ /* convert from q31 to q7 and then store the results in the destination buffer */
+ in1 = *pIn++;
+ in2 = *pIn++;
+ in3 = *pIn++;
+ in4 = *pIn++;
+
+ out1 = (q7_t) (in1 >> 24);
+ out2 = (q7_t) (in2 >> 24);
+ out3 = (q7_t) (in3 >> 24);
+ out4 = (q7_t) (in4 >> 24);
+
+ *__SIMD32(pDst)++ = __PACKq7(out1, out2, out3, out4);
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+ /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
+ ** No loop unrolling is used. */
+ blkCnt = blockSize % 0x4u;
+
+#else
+
+ /* Run the below code for Cortex-M0 */
+
+ /* Loop over blockSize number of values */
+ blkCnt = blockSize;
+
+#endif /* #ifndef ARM_MATH_CM0 */
+
+ while(blkCnt > 0u)
+ {
+ /* C = (q7_t) A >> 24 */
+ /* convert from q31 to q7 and then store the results in the destination buffer */
+ *pDst++ = (q7_t) (*pIn++ >> 24);
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+}
+
+/**
+ * @} end of q31_to_x group
+ */
diff --git a/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q7_to_float.c b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q7_to_float.c new file mode 100644 index 000000000..be4ac5fe3 --- /dev/null +++ b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q7_to_float.c @@ -0,0 +1,123 @@ +/* ----------------------------------------------------------------------------
+* Copyright (C) 2010 ARM Limited. All rights reserved.
+*
+* $Date: 15. February 2012
+* $Revision: V1.1.0
+*
+* Project: CMSIS DSP Library
+* Title: arm_q7_to_float.c
+*
+* Description: Converts the elements of the Q7 vector to floating-point vector.
+*
+* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
+*
+* Version 1.1.0 2012/02/15
+* Updated with more optimizations, bug fixes and minor API changes.
+*
+* Version 1.0.10 2011/7/15
+* Big Endian support added and Merged M0 and M3/M4 Source code.
+*
+* Version 1.0.3 2010/11/29
+* Re-organized the CMSIS folders and updated documentation.
+*
+* Version 1.0.2 2010/11/11
+* Documentation updated.
+*
+* Version 1.0.1 2010/10/05
+* Production release and review comments incorporated.
+*
+* Version 1.0.0 2010/09/20
+* Production release and review comments incorporated.
+* ---------------------------------------------------------------------------- */
+
+#include "arm_math.h"
+
+/**
+ * @ingroup groupSupport
+ */
+
+/**
+ * @defgroup q7_to_x Convert 8-bit Integer value
+ */
+
+/**
+ * @addtogroup q7_to_x
+ * @{
+ */
+
+/**
+ * @brief Converts the elements of the Q7 vector to floating-point vector.
+ * @param[in] *pSrc points to the Q7 input vector
+ * @param[out] *pDst points to the floating-point output vector
+ * @param[in] blockSize length of the input vector
+ * @return none.
+ *
+ * \par Description:
+ *
+ * The equation used for the conversion process is:
+ *
+ * <pre>
+ * pDst[n] = (float32_t) pSrc[n] / 128; 0 <= n < blockSize.
+ * </pre>
+ *
+ */
+
+
+void arm_q7_to_float(
+ q7_t * pSrc,
+ float32_t * pDst,
+ uint32_t blockSize)
+{
+ q7_t *pIn = pSrc; /* Src pointer */
+ uint32_t blkCnt; /* loop counter */
+
+
+#ifndef ARM_MATH_CM0
+
+ /* Run the below code for Cortex-M4 and Cortex-M3 */
+
+ /*loop Unrolling */
+ blkCnt = blockSize >> 2u;
+
+ /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
+ ** a second loop below computes the remaining 1 to 3 samples. */
+ while(blkCnt > 0u)
+ {
+ /* C = (float32_t) A / 128 */
+ /* convert from q7 to float and then store the results in the destination buffer */
+ *pDst++ = ((float32_t) * pIn++ / 128.0f);
+ *pDst++ = ((float32_t) * pIn++ / 128.0f);
+ *pDst++ = ((float32_t) * pIn++ / 128.0f);
+ *pDst++ = ((float32_t) * pIn++ / 128.0f);
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+ /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
+ ** No loop unrolling is used. */
+ blkCnt = blockSize % 0x4u;
+
+#else
+
+ /* Run the below code for Cortex-M0 */
+
+ /* Loop over blockSize number of values */
+ blkCnt = blockSize;
+
+#endif /* #ifndef ARM_MATH_CM0 */
+
+ while(blkCnt > 0u)
+ {
+ /* C = (float32_t) A / 128 */
+ /* convert from q7 to float and then store the results in the destination buffer */
+ *pDst++ = ((float32_t) * pIn++ / 128.0f);
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+}
+
+/**
+ * @} end of q7_to_x group
+ */
diff --git a/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q7_to_q15.c b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q7_to_q15.c new file mode 100644 index 000000000..019ca4815 --- /dev/null +++ b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q7_to_q15.c @@ -0,0 +1,149 @@ +/* ----------------------------------------------------------------------------
+* Copyright (C) 2010 ARM Limited. All rights reserved.
+*
+* $Date: 15. February 2012
+* $Revision: V1.1.0
+*
+* Project: CMSIS DSP Library
+* Title: arm_q7_to_q15.c
+*
+* Description: Converts the elements of the Q7 vector to Q15 vector.
+*
+* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
+*
+* Version 1.1.0 2012/02/15
+* Updated with more optimizations, bug fixes and minor API changes.
+*
+* Version 1.0.10 2011/7/15
+* Big Endian support added and Merged M0 and M3/M4 Source code.
+*
+* Version 1.0.3 2010/11/29
+* Re-organized the CMSIS folders and updated documentation.
+*
+* Version 1.0.2 2010/11/11
+* Documentation updated.
+*
+* Version 1.0.1 2010/10/05
+* Production release and review comments incorporated.
+*
+* Version 1.0.0 2010/09/20
+* Production release and review comments incorporated.
+* ---------------------------------------------------------------------------- */
+
+#include "arm_math.h"
+
+/**
+ * @ingroup groupSupport
+ */
+
+/**
+ * @addtogroup q7_to_x
+ * @{
+ */
+
+
+
+
+/**
+ * @brief Converts the elements of the Q7 vector to Q15 vector.
+ * @param[in] *pSrc points to the Q7 input vector
+ * @param[out] *pDst points to the Q15 output vector
+ * @param[in] blockSize length of the input vector
+ * @return none.
+ *
+ * \par Description:
+ *
+ * The equation used for the conversion process is:
+ *
+ * <pre>
+ * pDst[n] = (q15_t) pSrc[n] << 8; 0 <= n < blockSize.
+ * </pre>
+ *
+ */
+
+
+void arm_q7_to_q15(
+ q7_t * pSrc,
+ q15_t * pDst,
+ uint32_t blockSize)
+{
+ q7_t *pIn = pSrc; /* Src pointer */
+ uint32_t blkCnt; /* loop counter */
+
+#ifndef ARM_MATH_CM0
+ q31_t in;
+ q31_t in1, in2;
+ q31_t out1, out2;
+
+ /* Run the below code for Cortex-M4 and Cortex-M3 */
+
+ /*loop Unrolling */
+ blkCnt = blockSize >> 2u;
+
+ /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
+ ** a second loop below computes the remaining 1 to 3 samples. */
+ while(blkCnt > 0u)
+ {
+ /* C = (q15_t) A << 8 */
+ /* convert from q7 to q15 and then store the results in the destination buffer */
+ in = *__SIMD32(pIn)++;
+
+ /* rotatate in by 8 and extend two q7_t values to q15_t values */
+ in1 = __SXTB16(__ROR(in, 8));
+
+ /* extend remainig two q7_t values to q15_t values */
+ in2 = __SXTB16(in);
+
+ in1 = in1 << 8u;
+ in2 = in2 << 8u;
+
+ in1 = in1 & 0xFF00FF00;
+ in2 = in2 & 0xFF00FF00;
+
+#ifndef ARM_MATH_BIG_ENDIAN
+
+ out2 = __PKHTB(in1, in2, 16);
+ out1 = __PKHBT(in2, in1, 16);
+
+#else
+
+ out1 = __PKHTB(in1, in2, 16);
+ out2 = __PKHBT(in2, in1, 16);
+
+#endif
+
+ *__SIMD32(pDst)++ = out1;
+ *__SIMD32(pDst)++ = out2;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+ /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
+ ** No loop unrolling is used. */
+ blkCnt = blockSize % 0x4u;
+
+#else
+
+ /* Run the below code for Cortex-M0 */
+
+ /* Loop over blockSize number of values */
+ blkCnt = blockSize;
+
+#endif /* #ifndef ARM_MATH_CM0 */
+
+ while(blkCnt > 0u)
+ {
+ /* C = (q15_t) A << 8 */
+ /* convert from q7 to q15 and then store the results in the destination buffer */
+ *pDst++ = (q15_t) * pIn++ << 8;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+}
+
+/**
+ * @} end of q7_to_x group
+ */
diff --git a/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q7_to_q31.c b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q7_to_q31.c new file mode 100644 index 000000000..bbbcc6f4a --- /dev/null +++ b/src/modules/mathlib/CMSIS/DSP_Lib/Source/SupportFunctions/arm_q7_to_q31.c @@ -0,0 +1,134 @@ +/* ----------------------------------------------------------------------------
+* Copyright (C) 2010 ARM Limited. All rights reserved.
+*
+* $Date: 15. February 2012
+* $Revision: V1.1.0
+*
+* Project: CMSIS DSP Library
+* Title: arm_q7_to_q31.c
+*
+* Description: Converts the elements of the Q7 vector to Q31 vector.
+*
+* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
+*
+* Version 1.1.0 2012/02/15
+* Updated with more optimizations, bug fixes and minor API changes.
+*
+* Version 1.0.10 2011/7/15
+* Big Endian support added and Merged M0 and M3/M4 Source code.
+*
+* Version 1.0.3 2010/11/29
+* Re-organized the CMSIS folders and updated documentation.
+*
+* Version 1.0.2 2010/11/11
+* Documentation updated.
+*
+* Version 1.0.1 2010/10/05
+* Production release and review comments incorporated.
+*
+* Version 1.0.0 2010/09/20
+* Production release and review comments incorporated.
+* ---------------------------------------------------------------------------- */
+
+#include "arm_math.h"
+
+/**
+ * @ingroup groupSupport
+ */
+
+/**
+ * @addtogroup q7_to_x
+ * @{
+ */
+
+/**
+ * @brief Converts the elements of the Q7 vector to Q31 vector.
+ * @param[in] *pSrc points to the Q7 input vector
+ * @param[out] *pDst points to the Q31 output vector
+ * @param[in] blockSize length of the input vector
+ * @return none.
+ *
+ * \par Description:
+ *
+ * The equation used for the conversion process is:
+ *
+ * <pre>
+ * pDst[n] = (q31_t) pSrc[n] << 24; 0 <= n < blockSize.
+ * </pre>
+ *
+ */
+
+
+void arm_q7_to_q31(
+ q7_t * pSrc,
+ q31_t * pDst,
+ uint32_t blockSize)
+{
+ q7_t *pIn = pSrc; /* Src pointer */
+ uint32_t blkCnt; /* loop counter */
+
+#ifndef ARM_MATH_CM0
+
+ q31_t in;
+
+ /* Run the below code for Cortex-M4 and Cortex-M3 */
+
+ /*loop Unrolling */
+ blkCnt = blockSize >> 2u;
+
+ /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
+ ** a second loop below computes the remaining 1 to 3 samples. */
+ while(blkCnt > 0u)
+ {
+ /* C = (q31_t) A << 24 */
+ /* convert from q7 to q31 and then store the results in the destination buffer */
+ in = *__SIMD32(pIn)++;
+
+#ifndef ARM_MATH_BIG_ENDIAN
+
+ *pDst++ = (__ROR(in, 8)) & 0xFF000000;
+ *pDst++ = (__ROR(in, 16)) & 0xFF000000;
+ *pDst++ = (__ROR(in, 24)) & 0xFF000000;
+ *pDst++ = (in & 0xFF000000);
+
+#else
+
+ *pDst++ = (in & 0xFF000000);
+ *pDst++ = (__ROR(in, 24)) & 0xFF000000;
+ *pDst++ = (__ROR(in, 16)) & 0xFF000000;
+ *pDst++ = (__ROR(in, 8)) & 0xFF000000;
+
+#endif // #ifndef ARM_MATH_BIG_ENDIAN
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+ /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
+ ** No loop unrolling is used. */
+ blkCnt = blockSize % 0x4u;
+
+#else
+
+ /* Run the below code for Cortex-M0 */
+
+ /* Loop over blockSize number of values */
+ blkCnt = blockSize;
+
+#endif /* #ifndef ARM_MATH_CM0 */
+
+ while(blkCnt > 0u)
+ {
+ /* C = (q31_t) A << 24 */
+ /* convert from q7 to q31 and then store the results in the destination buffer */
+ *pDst++ = (q31_t) * pIn++ << 24;
+
+ /* Decrement the loop counter */
+ blkCnt--;
+ }
+
+}
+
+/**
+ * @} end of q7_to_x group
+ */
|