| /cmsis-3.4.0/CMSIS/DSP/Source/BasicMathFunctions/ |
| D | arm_offset_q7.c | 58 q7_t offset, in arm_offset_q7() argument
74 vst1q(pDst, vqaddq(vecSrc, offset)); in arm_offset_q7()
93 vstrbq_p(pDst, vqaddq(vecSrc, offset), p0); in arm_offset_q7()
100 q7_t offset, in arm_offset_q7() argument
112 offset_packed = __PACKq7(offset, offset, offset, offset); in arm_offset_q7()
126 *pDst++ = (q7_t) __SSAT((q15_t) *pSrc++ + offset, 8); in arm_offset_q7()
127 *pDst++ = (q7_t) __SSAT((q15_t) *pSrc++ + offset, 8); in arm_offset_q7()
128 *pDst++ = (q7_t) __SSAT((q15_t) *pSrc++ + offset, 8); in arm_offset_q7()
129 *pDst++ = (q7_t) __SSAT((q15_t) *pSrc++ + offset, 8); in arm_offset_q7()
151 *pDst++ = (q7_t) __SSAT((q15_t) *pSrc++ + offset, 8); in arm_offset_q7()
|
| D | arm_offset_q15.c | 58 q15_t offset, in arm_offset_q15() argument
74 vst1q(pDst, vqaddq(vecSrc, offset)); in arm_offset_q15()
93 vstrhq_p(pDst, vqaddq(vecSrc, offset), p0); in arm_offset_q15()
101 q15_t offset, in arm_offset_q15() argument
113 offset_packed = __PKHBT(offset, offset, 16); in arm_offset_q15()
128 *pDst++ = (q15_t) __SSAT(((q31_t) *pSrc++ + offset), 16); in arm_offset_q15()
129 *pDst++ = (q15_t) __SSAT(((q31_t) *pSrc++ + offset), 16); in arm_offset_q15()
130 *pDst++ = (q15_t) __SSAT(((q31_t) *pSrc++ + offset), 16); in arm_offset_q15()
131 *pDst++ = (q15_t) __SSAT(((q31_t) *pSrc++ + offset), 16); in arm_offset_q15()
154 *pDst++ = (q15_t) __QADD16(*pSrc++, offset); in arm_offset_q15()
[all …]
|
| D | arm_offset_q31.c | 59 q31_t offset, in arm_offset_q31() argument
75 vst1q(pDst, vqaddq(vecSrc, offset)); in arm_offset_q31()
94 vstrwq_p(pDst, vqaddq(vecSrc, offset), p0); in arm_offset_q31()
101 q31_t offset, in arm_offset_q31() argument
117 *pDst++ = __QADD(*pSrc++, offset); in arm_offset_q31()
119 *pDst++ = __QADD(*pSrc++, offset); in arm_offset_q31()
121 *pDst++ = __QADD(*pSrc++, offset); in arm_offset_q31()
123 *pDst++ = __QADD(*pSrc++, offset); in arm_offset_q31()
145 *pDst++ = __QADD(*pSrc++, offset); in arm_offset_q31()
147 *pDst++ = (q31_t) clip_q63_to_q31((q63_t) * pSrc++ + offset); in arm_offset_q31()
|
| D | arm_offset_f32.c | 69 float32_t offset, in arm_offset_f32() argument
86 res = vaddq(vec1,offset); in arm_offset_f32()
104 vstrwq_p(pDst, vaddq(vec1, offset), p0); in arm_offset_f32()
113 float32_t offset, in arm_offset_f32() argument
132 res = vaddq_f32(vec1,vdupq_n_f32(offset)); in arm_offset_f32()
157 *pDst++ = (*pSrc++) + offset; in arm_offset_f32()
159 *pDst++ = (*pSrc++) + offset; in arm_offset_f32()
161 *pDst++ = (*pSrc++) + offset; in arm_offset_f32()
163 *pDst++ = (*pSrc++) + offset; in arm_offset_f32()
185 *pDst++ = (*pSrc++) + offset; in arm_offset_f32()
|
| D | arm_offset_f16.c | 69 float16_t offset, in arm_offset_f16() argument
86 res = vaddq(vec1,offset); in arm_offset_f16()
104 vstrhq_p(pDst, vaddq(vec1, offset), p0); in arm_offset_f16()
114 float16_t offset, in arm_offset_f16() argument
131 *pDst++ = (*pSrc++) + offset; in arm_offset_f16()
133 *pDst++ = (*pSrc++) + offset; in arm_offset_f16()
135 *pDst++ = (*pSrc++) + offset; in arm_offset_f16()
137 *pDst++ = (*pSrc++) + offset; in arm_offset_f16()
158 *pDst++ = (*pSrc++) + offset; in arm_offset_f16()
|
| /cmsis-3.4.0/CMSIS/NN/Source/NNSupportFunctions/ |
| D | arm_q7_to_q15_with_offset.c | 42 void arm_q7_to_q15_with_offset(const q7_t *src, q15_t *dst, uint32_t block_size, q15_t offset) in arm_q7_to_q15_with_offset() argument
49 const int16x8_t source_offset = vdupq_n_s16(offset); in arm_q7_to_q15_with_offset()
76 const q31_t offset_q15x2 = __PKHBT(offset, offset, 16); in arm_q7_to_q15_with_offset()
105 *dst++ = (q15_t)*src++ + offset; in arm_q7_to_q15_with_offset()
|
| D | arm_nn_depthwise_conv_nt_t_padded_s8.c | 67 for (int i_loop_cnt = 0, offset = 0; i_loop_cnt < loop_count; in arm_nn_depthwise_conv_nt_t_padded_s8() local
68 num_ch_to_process -= 4, out += 4, offset += 4, i_loop_cnt++) in arm_nn_depthwise_conv_nt_t_padded_s8()
76 const int8_t *rhs_0 = rhs + offset; in arm_nn_depthwise_conv_nt_t_padded_s8()
77 const int8_t *lhs_0 = lhs + offset; in arm_nn_depthwise_conv_nt_t_padded_s8()
78 const int8_t *lhs_1 = lhs + row_x_col * num_ch + offset; in arm_nn_depthwise_conv_nt_t_padded_s8()
79 const int8_t *lhs_2 = lhs + (row_x_col * num_ch * 2) + offset; in arm_nn_depthwise_conv_nt_t_padded_s8()
80 const int8_t *lhs_3 = lhs + (row_x_col * num_ch * 3) + offset; in arm_nn_depthwise_conv_nt_t_padded_s8()
|
| D | arm_nn_depthwise_conv_nt_t_s8.c | 66 for (int i_loop_cnt = 0, offset = 0; i_loop_cnt < loop_count; in arm_nn_depthwise_conv_nt_t_s8() local
67 num_ch_to_process -= 4, offset += 4, out += 4, i_loop_cnt++) in arm_nn_depthwise_conv_nt_t_s8()
75 const int8_t *rhs_0 = rhs + offset; in arm_nn_depthwise_conv_nt_t_s8()
76 const int8_t *lhs_0 = lhs + offset; in arm_nn_depthwise_conv_nt_t_s8()
77 const int8_t *lhs_1 = lhs + row_x_col * num_ch + offset; in arm_nn_depthwise_conv_nt_t_s8()
78 const int8_t *lhs_2 = lhs + (row_x_col * num_ch * 2) + offset; in arm_nn_depthwise_conv_nt_t_s8()
79 const int8_t *lhs_3 = lhs + (row_x_col * num_ch * 3) + offset; in arm_nn_depthwise_conv_nt_t_s8()
|
| D | arm_nn_mat_mul_core_4x_s8.c | 48 const int32_t offset, in arm_nn_mat_mul_core_4x_s8() argument
60 const int8_t *ip_row_1 = row_base + offset; in arm_nn_mat_mul_core_4x_s8()
61 const int8_t *ip_row_2 = row_base + (2 * offset); in arm_nn_mat_mul_core_4x_s8()
62 const int8_t *ip_row_3 = row_base + (3 * offset); in arm_nn_mat_mul_core_4x_s8()
|
| /cmsis-3.4.0/CMSIS/DSP/Source/FilteringFunctions/ |
| D | arm_levinson_durbin_f16.c | 135 uint16x8_t offset,offsetInc,vecTmp; in arm_levinson_durbin_f16() local
138 offset = vld1q(orgOffsetArray); in arm_levinson_durbin_f16()
141 offset = vaddq_m_u16(offset,offset,vecTmp,LANE4567_MASK); in arm_levinson_durbin_f16()
162 vecA = vldrhq_gather_shifted_offset_f16(a,offset); in arm_levinson_durbin_f16()
172 vstrhq_scatter_shifted_offset_f16(a, offset, tmp); in arm_levinson_durbin_f16()
174 offset = vaddq(offset,offsetInc); in arm_levinson_durbin_f16()
|
| D | arm_levinson_durbin_f32.c | 135 uint32x4_t offset,offsetInc,vecTmp; in arm_levinson_durbin_f32() local
138 offset = vld1q(orgOffsetArray); in arm_levinson_durbin_f32()
141 offset = vaddq_m_u32(offset,offset,vecTmp,LANE23_MASK); in arm_levinson_durbin_f32()
162 vecA = vldrwq_gather_shifted_offset_f32(a,offset); in arm_levinson_durbin_f32()
172 vstrwq_scatter_shifted_offset_f32(a, offset, tmp); in arm_levinson_durbin_f32()
174 offset = vaddq(offset,offsetInc); in arm_levinson_durbin_f32()
|
| D | arm_levinson_durbin_q31.c | 202 uint32x4_t offset,offsetInc,vecTmp; in arm_levinson_durbin_q31() local
205 offset = vld1q(orgOffsetArray); in arm_levinson_durbin_q31()
208 offset = vaddq_m_u32(offset,offset,vecTmp,LANE23_MASK); in arm_levinson_durbin_q31()
234 vecA = vldrwq_gather_shifted_offset_s32(a,offset); in arm_levinson_durbin_q31()
244 vstrwq_scatter_shifted_offset_s32(a, offset, tmp); in arm_levinson_durbin_q31()
246 offset = vaddq(offset,offsetInc); in arm_levinson_durbin_q31()
|
| /cmsis-3.4.0/CMSIS/DSP/Source/TransformFunctions/ |
| D | arm_rfft_q31.c | 158 uint32x4_t offset = { 2, 3, 0, 1 }; in arm_split_rfft_q31() local
161 offset = offset + (2 * fftLen - 4); in arm_split_rfft_q31()
178 q31x4_t in2 = vldrwq_gather_shifted_offset_s32(pSrc, offset); in arm_split_rfft_q31()
190 offset -= 4; in arm_split_rfft_q31()
314 uint32x4_t offset = { 2, 3, 0, 1 }; in arm_split_rifft_q31() local
318 offset = offset + (2 * fftLen - 2); in arm_split_rifft_q31()
334 q31x4_t in2 = vldrwq_gather_shifted_offset_s32(pSrc, offset); in arm_split_rifft_q31()
350 offset -= 4; in arm_split_rifft_q31()
|
| D | arm_rfft_q15.c | 384 uint16x8_t offset = { 6, 7, 4, 5, 2, 3, 0, 1 }; in arm_split_rifft_q15() local
394 offset = vaddq_n_u16(offset, (2 * fftLen - 6)); in arm_split_rifft_q15()
411 q15x8_t in2 = vldrhq_gather_shifted_offset_s16(pSrc, offset); in arm_split_rifft_q15()
423 offset -= 8; in arm_split_rifft_q15()
|
| /cmsis-3.4.0/CMSIS/NN/Source/ConvolutionFunctions/ |
| D | arm_nn_mat_mult_s8.c | 76 const int16x8_t offset = vdupq_m_n_s16(vuninitializedq_s16(), col_offset, p); in arm_nn_mat_mult_s8() local
84 c0 = vaddq_m_s16(vuninitializedq_s16(), c0, offset, p); in arm_nn_mat_mult_s8()
88 c1 = vaddq_m_s16(vuninitializedq_s16(), c1, offset, p); in arm_nn_mat_mult_s8()
92 c2 = vaddq_m_s16(vuninitializedq_s16(), c2, offset, p); in arm_nn_mat_mult_s8()
96 c3 = vaddq_m_s16(vuninitializedq_s16(), c3, offset, p); in arm_nn_mat_mult_s8()
136 const int16x8_t offset = vdupq_m_n_s16(vuninitializedq_s16(), col_offset, p); in arm_nn_mat_mult_s8() local
144 c0 = vaddq_m_s16(vuninitializedq_s16(), c0, offset, p); in arm_nn_mat_mult_s8()
|
| D | arm_depthwise_conv_s8_opt.c | 170 …for (int i_loop_cnt = 0, offset = 0; i_loop_cnt < loop_count; num_ch_to_process -= 4, offset += 4,… in arm_depthwise_conv_s8_opt() local
172 const int8_t *col_0 = lhs_buffer + (kernel_size * input_ch * i_buf) + offset; in arm_depthwise_conv_s8_opt()
173 const int8_t *row_0 = kernel + offset; in arm_depthwise_conv_s8_opt()
174 int32x4_t out_0 = vldrwq_s32(&bias[offset]); in arm_depthwise_conv_s8_opt()
188 const int32x4_t mult = vldrwq_s32(&output_mult[offset]); in arm_depthwise_conv_s8_opt()
189 const int32x4_t shift = vldrwq_s32(&output_shift[offset]); in arm_depthwise_conv_s8_opt()
|
| D | arm_nn_depthwise_conv_s8_core.c | 69 const int32_t offset = idx * 4; in arm_nn_depthwise_conv_s8_core() local
70 const int8_t *row_0 = row + offset; in arm_nn_depthwise_conv_s8_core()
71 const int16_t *col_0 = col + offset; in arm_nn_depthwise_conv_s8_core()
72 const int16_t *col_1 = col + kernel_size * num_ch + offset; in arm_nn_depthwise_conv_s8_core()
|
| /cmsis-3.4.0/CMSIS/DSP/Include/dsp/ |
| D | basic_math_functions.h | 438 float32_t offset,
453 q7_t offset,
467 q15_t offset,
481 q31_t offset,
|
| D | basic_math_functions_f16.h | 105 float16_t offset,
|
| /cmsis-3.4.0/CMSIS/NN/Include/ |
| D | arm_nnsupportfunctions.h | 138 void arm_q7_to_q15_with_offset(const q7_t *src, q15_t *dst, uint32_t block_size, q15_t offset);
153 …rm_q7_to_q15_reordered_with_offset(const q7_t *src, q15_t *dst, uint32_t block_size, q15_t offset);
286 const int32_t offset,
610 read_and_pad_reordered_with_offset(const q7_t *source, q31_t *out1, q31_t *out2, q31_t offset) in read_and_pad_reordered_with_offset() argument
621 *out1 = __QADD16(*out1, offset); in read_and_pad_reordered_with_offset()
622 *out2 = __QADD16(*out2, offset); in read_and_pad_reordered_with_offset()
|
| /cmsis-3.4.0/CMSIS/Core_A/Include/ |
| D | core_ca.h | 2483 uint32_t offset; in MMU_TTSection() local
2487 offset = base_address >> 20; in MMU_TTSection()
2491 ttb = ttb + offset; in MMU_TTSection()
2514 uint32_t offset, offset2; in MMU_TTPage4k() local
2518 offset = base_address >> 20; in MMU_TTPage4k()
2522 ttb += offset; in MMU_TTPage4k()
2549 uint32_t offset, offset2; in MMU_TTPage64k() local
2554 offset = base_address >> 20; in MMU_TTPage64k()
2558 ttb += offset; in MMU_TTPage64k()
|
