/hal_nxp-3.6.0/mcux/mcux-sdk/CMSIS/DSP/Source/TransformFunctions/ |
D | arm_cfft_f16.c | 123 float16_t *inA = pSrc + CMPLX_DIM * i * n1; in _arm_radix4_butterfly_f16_mve() local 124 float16_t *inB = inA + n2 * CMPLX_DIM; in _arm_radix4_butterfly_f16_mve() 138 vecA = vldrhq_f16(inA); in _arm_radix4_butterfly_f16_mve() 154 vst1q(inA, vecTmp0); in _arm_radix4_butterfly_f16_mve() 155 inA += 8; in _arm_radix4_butterfly_f16_mve() 196 vecA = vldrhq_f16(inA); in _arm_radix4_butterfly_f16_mve() 332 float16_t *inA = pSrc + CMPLX_DIM * i * n1; in _arm_radix4_butterfly_inverse_f16_mve() local 333 float16_t *inB = inA + n2 * CMPLX_DIM; in _arm_radix4_butterfly_inverse_f16_mve() 345 vecA = vldrhq_f16(inA); in _arm_radix4_butterfly_inverse_f16_mve() 361 vst1q(inA, vecTmp0); in _arm_radix4_butterfly_inverse_f16_mve() [all …]
|
D | arm_cfft_q31.c | 78 q31_t *inA = pSrc + CMPLX_DIM * i * n1; in _arm_radix4_butterfly_q31_mve() local 79 q31_t *inB = inA + n2 * CMPLX_DIM; in _arm_radix4_butterfly_q31_mve() 91 vecA = vldrwq_s32(inA); in _arm_radix4_butterfly_q31_mve() 107 vst1q(inA, vecTmp0); in _arm_radix4_butterfly_q31_mve() 108 inA += 4; in _arm_radix4_butterfly_q31_mve() 147 vecA = vldrwq_s32(inA); in _arm_radix4_butterfly_q31_mve() 328 q31_t *inA = pSrc + CMPLX_DIM * i * n1; in _arm_radix4_butterfly_inverse_q31_mve() local 329 q31_t *inB = inA + n2 * CMPLX_DIM; in _arm_radix4_butterfly_inverse_q31_mve() 341 vecA = vldrwq_s32(inA); in _arm_radix4_butterfly_inverse_q31_mve() 357 vst1q(inA, vecTmp0); in _arm_radix4_butterfly_inverse_q31_mve() [all …]
|
D | arm_cfft_q15.c | 75 q15_t *inA = pSrc + CMPLX_DIM * i * n1; in _arm_radix4_butterfly_q15_mve() local 76 q15_t *inB = inA + n2 * CMPLX_DIM; in _arm_radix4_butterfly_q15_mve() 88 vecA = vldrhq_s16(inA); in _arm_radix4_butterfly_q15_mve() 104 vst1q(inA, vecTmp0); in _arm_radix4_butterfly_q15_mve() 105 inA += 8; in _arm_radix4_butterfly_q15_mve() 145 vecA = vldrhq_s16(inA); in _arm_radix4_butterfly_q15_mve() 311 q15_t *inA = pSrc + CMPLX_DIM * i * n1; in _arm_radix4_butterfly_inverse_q15_mve() local 312 q15_t *inB = inA + n2 * CMPLX_DIM; in _arm_radix4_butterfly_inverse_q15_mve() 324 vecA = vldrhq_s16(inA); in _arm_radix4_butterfly_inverse_q15_mve() 340 vst1q(inA, vecTmp0); in _arm_radix4_butterfly_inverse_q15_mve() [all …]
|
D | arm_cfft_f32.c | 125 float32_t *inA = pSrc + CMPLX_DIM * i * n1; in _arm_radix4_butterfly_f32_mve() local 126 float32_t *inB = inA + n2 * CMPLX_DIM; in _arm_radix4_butterfly_f32_mve() 140 vecA = vldrwq_f32(inA); in _arm_radix4_butterfly_f32_mve() 156 vst1q(inA, vecTmp0); in _arm_radix4_butterfly_f32_mve() 157 inA += 4; in _arm_radix4_butterfly_f32_mve() 198 vecA = vldrwq_f32(inA); in _arm_radix4_butterfly_f32_mve() 334 float32_t *inA = pSrc + CMPLX_DIM * i * n1; in _arm_radix4_butterfly_inverse_f32_mve() local 335 float32_t *inB = inA + n2 * CMPLX_DIM; in _arm_radix4_butterfly_inverse_f32_mve() 347 vecA = vldrwq_f32(inA); in _arm_radix4_butterfly_inverse_f32_mve() 363 vst1q(inA, vecTmp0); in _arm_radix4_butterfly_inverse_f32_mve() [all …]
|
/hal_nxp-3.6.0/mcux/mcux-sdk/CMSIS/NN/Include/ |
D | arm_nnsupportfunctions.h | 573 q31_t inA = arm_nn_read_q7x4_ia(&source); in read_and_pad() local 574 q31_t inAbuf1 = __SXTB16(__ROR((uint32_t)inA, 8)); in read_and_pad() 575 q31_t inAbuf2 = __SXTB16(inA); in read_and_pad() 594 q31_t inA = arm_nn_read_q7x4_ia(&source); in read_and_pad_reordered() local 596 *out2 = __SXTB16(__ROR((uint32_t)inA, 8)); in read_and_pad_reordered() 597 *out1 = __SXTB16(inA); in read_and_pad_reordered() 599 *out1 = __SXTB16(__ROR((uint32_t)inA, 8)); in read_and_pad_reordered() 600 *out2 = __SXTB16(inA); in read_and_pad_reordered() 612 q31_t inA = arm_nn_read_q7x4_ia(&source); in read_and_pad_reordered_with_offset() local 615 *out2 = __SXTB16(__ROR((uint32_t)inA, 8)); in read_and_pad_reordered_with_offset() [all …]
|
/hal_nxp-3.6.0/mcux/mcux-sdk/CMSIS/NN/Source/FullyConnectedFunctions/ |
D | arm_fully_connected_q15_opt.c | 291 q15_t inA = *pA++; in arm_fully_connected_q15_opt() local 293 sum += inA * inB; in arm_fully_connected_q15_opt() 295 sum2 += inA * inB; in arm_fully_connected_q15_opt() 297 sum3 += inA * inB; in arm_fully_connected_q15_opt() 299 sum4 += inA * inB; in arm_fully_connected_q15_opt() 319 q15_t inA = *pA++; in arm_fully_connected_q15_opt() local 321 ip_out += inA * inB; in arm_fully_connected_q15_opt()
|
D | arm_fully_connected_mat_q7_vec_q15_opt.c | 370 q15_t inA = *pA++; in arm_fully_connected_mat_q7_vec_q15_opt() local 372 sum += inA * inB; in arm_fully_connected_mat_q7_vec_q15_opt() 374 sum2 += inA * inB; in arm_fully_connected_mat_q7_vec_q15_opt() 376 sum3 += inA * inB; in arm_fully_connected_mat_q7_vec_q15_opt() 378 sum4 += inA * inB; in arm_fully_connected_mat_q7_vec_q15_opt() 400 q15_t inA = *pA++; in arm_fully_connected_mat_q7_vec_q15_opt() local 402 ip_out += inA * inB; in arm_fully_connected_mat_q7_vec_q15_opt()
|
D | arm_fully_connected_q7_opt.c | 446 q7_t inA = *pA++; in arm_fully_connected_q7_opt() 448 sum += inA * inB; in arm_fully_connected_q7_opt() 450 sum2 += inA * inB; in arm_fully_connected_q7_opt() 452 sum3 += inA * inB; in arm_fully_connected_q7_opt() 454 sum4 += inA * inB; in arm_fully_connected_q7_opt() 477 q7_t inA = *pA++; in arm_fully_connected_q7_opt() 479 ip_out += inA * inB; in arm_fully_connected_q7_opt()
|
/hal_nxp-3.6.0/mcux/mcux-sdk/CMSIS/NN/Source/ConvolutionFunctions/ |
D | arm_depthwise_separable_conv_HWC_q7.c | 326 union arm_nnword inA, inB; in arm_depthwise_separable_conv_HWC_q7() local 327 inA.word = arm_nn_read_q7x4(pA); in arm_depthwise_separable_conv_HWC_q7() 331 sum += inA.bytes[0] * inB.bytes[0]; in arm_depthwise_separable_conv_HWC_q7() 332 sum2 += inA.bytes[1] * inB.bytes[1]; in arm_depthwise_separable_conv_HWC_q7() 333 sum3 += inA.bytes[2] * inB.bytes[2]; in arm_depthwise_separable_conv_HWC_q7() 334 sum4 += inA.bytes[3] * inB.bytes[3]; in arm_depthwise_separable_conv_HWC_q7()
|
D | arm_depthwise_separable_conv_HWC_q7_nonsquare.c | 330 union arm_nnword inA, inB; in arm_depthwise_separable_conv_HWC_q7_nonsquare() local 331 inA.word = arm_nn_read_q7x4(pA); in arm_depthwise_separable_conv_HWC_q7_nonsquare() 335 sum += inA.bytes[0] * inB.bytes[0]; in arm_depthwise_separable_conv_HWC_q7_nonsquare() 336 sum2 += inA.bytes[1] * inB.bytes[1]; in arm_depthwise_separable_conv_HWC_q7_nonsquare() 337 sum3 += inA.bytes[2] * inB.bytes[2]; in arm_depthwise_separable_conv_HWC_q7_nonsquare() 338 sum4 += inA.bytes[3] * inB.bytes[3]; in arm_depthwise_separable_conv_HWC_q7_nonsquare()
|