Lines Matching +full:left +full:- +full:shifted
1 /* Copyright (C) 2010-2021 ARM Limited or its affiliates. All rights reserved.
2 * SPDX-License-Identifier: Apache-2.0
29 * Element-by-element multiplication of two vectors.
33 * There are separate functions for floating-point, Q7, Q15, and Q31 data types.
83 * @brief Floating-point vector multiplication.
100 * Element-by-element addition of two vectors.
104 * There are separate functions for floating-point, Q7, Q15, and Q31 data types.
109 * @brief Floating-point vector addition.
171 * Element-by-element subtraction of two vectors.
173 * dst[n] = src_a[n] - src_b[n], 0 <= n < block_size.
175 * There are separate functions for floating-point, Q7, Q15, and Q31 data types.
180 * @brief Floating-point vector subtraction.
242 * Multiply a vector by a scalar value. For floating-point data, the algorithm used is:
247 * In the fixed-point Q7, Q15, and Q31 functions, scale is represented by a fractional
249 * allows the gain of the scaling operation to exceed 1.0. The algorithm used with fixed-point data
255 * The overall scale factor applied to the fixed-point data is
259 * The functions support in-place computation allowing the source and destination pointers to
265 * @brief Multiplies a floating-point vector by a scalar.
279 * These are multiplied to yield a 2.14 intermediate result and this is shifted with saturation to
296 * These are multiplied to yield a 2.30 intermediate result and this is shifted with saturation to
313 * These are multiplied to yield a 2.62 intermediate result and this is shifted with saturation to
333 * Computes the absolute value of a vector on an element-by-element basis.
337 * The functions support in-place computation allowing the source and destination pointers to
338 * reference the same memory buffer. There are separate functions for floating-point, Q7, Q15, and
344 * @brief Floating-point vector absolute value.
357 * The Q7 value -1 (0x80) will be saturated to the maximum allowable positive value 0x7F.
370 * The Q15 value -1 (0x8000) will be saturated to the maximum allowable positive value 0x7FFF.
384 * The Q31 value -1 (0x80000000) will be saturated to the maximum allowable positive value
402 * Computes the dot product of two vectors. The vectors are multiplied element-by-element and then
405 * sum = src_a[0]*src_b[0] + src_a[1]*src_b[1] + ... + src_a[block_size-1]*src_b[block_size-1]
407 * There are separate functions for floating-point, Q7, Q15, and Q31 data types.
412 * @brief Dot product of floating-point vectors.
444 * to a 64-bit accumulator in 34.30 format. Nonsaturating additions are used and given that there
462 * to a 64-bit accumulator in 16.48 format. There are 15 guard bits in the accumulator and there
482 * Shifts the elements of a fixed-point vector by a specified number of bits.
487 * If <code>shift</code> is positive then the elements of the vector are shifted to the left.
488 * If <code>shift</code> is negative then the elements of the vector are shifted to the right.
490 * The functions support in-place computation allowing the source and destination pointers to
503 * @param[in] shift_bits number of bits to shift. A positive value shifts left; a negative value
519 * @param[in] shift_bits number of bits to shift. A positive value shifts left; a negative value
535 * @param[in] shift_bits number of bits to shift. A positive value shifts left; a negative value
555 * The functions support in-place computation allowing the source and destination pointers to
556 * reference the same memory buffer. There are separate functions for floating-point, Q7, Q15, and
563 * @brief Adds a constant offset to a floating-point vector.
627 * dst[n] = -src[n], 0 <= n < block_size.
629 * The functions support in-place computation allowing the source and destination pointers to
630 * reference the same memory buffer. There are separate functions for floating-point, Q7, Q15, and
637 * @brief Negates the elements of a floating-point vector.
650 * The Q7 value -1 (0x80) is saturated to the maximum allowable positive value 0x7F.
664 * The Q15 value -1 (0x8000) is saturated to the maximum allowable positive value 0x7FFF.
678 * The Q31 value -1 (0x80000000) is saturated to the maximum allowable positive value 0x7FFFFFFF.
702 * @brief Compute the logical bitwise AND of two fixed-point vectors.
712 * @brief Compute the logical bitwise AND of two fixed-point vectors.
722 * @brief Compute the logical bitwise AND of two fixed-point vectors.
746 * @brief Compute the logical bitwise OR of two fixed-point vectors.
756 * @brief Compute the logical bitwise OR of two fixed-point vectors.
766 * @brief Compute the logical bitwise OR of two fixed-point vectors.
790 * @brief Compute the logical bitwise NOT of a fixed-point vector.
799 * @brief Compute the logical bitwise NOT of a fixed-point vector.
808 * @brief Compute the logical bitwise NOT of a fixed-point vector.
831 * @brief Compute the logical bitwise XOR of two fixed-point vectors.
841 * @brief Compute the logical bitwise XOR of two fixed-point vectors.
851 * @brief Compute the logical bitwise XOR of two fixed-point vectors.
868 * Element-by-element clipping of a value.
872 * There are separate functions for floating-point, Q7, Q15, and Q31 data types.
877 * @brief Elementwise floating-point clipping
888 * @brief Elementwise fixed-point clipping
899 * @brief Elementwise fixed-point clipping
910 * @brief Elementwise fixed-point clipping