/hal_nxp-3.5.0/mcux/mcux-sdk/CMSIS/DSP/Include/ |
D | arm_vec_math_f16.h | 47 __STATIC_INLINE f16x8_t vrecip_medprec_f16( in vrecip_medprec_f16() 48 f16x8_t x) in vrecip_medprec_f16() 51 f16x8_t b; in vrecip_medprec_f16() 53 f16x8_t ax = vabsq(x); in vrecip_medprec_f16() 81 __STATIC_INLINE f16x8_t vrecip_hiprec_f16( in vrecip_hiprec_f16() 82 f16x8_t x) in vrecip_hiprec_f16() 85 f16x8_t b; in vrecip_hiprec_f16() 87 f16x8_t ax = vabsq(x); in vrecip_hiprec_f16() 117 __STATIC_INLINE f16x8_t vdiv_f16( in vdiv_f16() 118 f16x8_t num, f16x8_t den) in vdiv_f16() [all …]
|
/hal_nxp-3.5.0/mcux/mcux-sdk/CMSIS/DSP/Source/TransformFunctions/ |
D | arm_cfft_f16.c | 93 f16x8_t vecTmp0, vecTmp1; in _arm_radix4_butterfly_f16_mve() 94 f16x8_t vecSum0, vecDiff0, vecSum1, vecDiff1; in _arm_radix4_butterfly_f16_mve() 95 f16x8_t vecA, vecB, vecC, vecD; in _arm_radix4_butterfly_f16_mve() 127 f16x8_t vecW; in _arm_radix4_butterfly_f16_mve() 215 vecA = (f16x8_t)vldrwq_gather_base_wb_f32(&vecScGathAddr, 64); in _arm_radix4_butterfly_f16_mve() 216 vecC = (f16x8_t)vldrwq_gather_base_f32(vecScGathAddr, 8); in _arm_radix4_butterfly_f16_mve() 224 vecB = (f16x8_t)vldrwq_gather_base_f32(vecScGathAddr, 4); in _arm_radix4_butterfly_f16_mve() 225 vecD = (f16x8_t)vldrwq_gather_base_f32(vecScGathAddr, 12); in _arm_radix4_butterfly_f16_mve() 231 vecA = (f16x8_t)vldrwq_gather_base_wb_f32(&vecScGathAddr, 64); in _arm_radix4_butterfly_f16_mve() 232 vecC = (f16x8_t)vldrwq_gather_base_f32(vecScGathAddr, 8); in _arm_radix4_butterfly_f16_mve() [all …]
|
/hal_nxp-3.5.0/mcux/mcux-sdk/CMSIS/DSP/Source/SVMFunctions/ |
D | arm_svm_rbf_predict_f16.c | 73 f16x8_t vSum = vdupq_n_f16(0); in arm_svm_rbf_predict_f16() 83 f16x8_t vecIn, acc0, acc1, acc2, acc3; in arm_svm_rbf_predict_f16() 112 f16x8_t vecA; in arm_svm_rbf_predict_f16() 113 f16x8_t vecDif; in arm_svm_rbf_predict_f16() 143 f16x8_t vecA; in arm_svm_rbf_predict_f16() 144 f16x8_t vecDif; in arm_svm_rbf_predict_f16() 165 f16x8_t vtmp = vuninitializedq_f16(); in arm_svm_rbf_predict_f16() 187 f16x8_t vecIn, acc0, acc1; in arm_svm_rbf_predict_f16() 209 f16x8_t vecA; in arm_svm_rbf_predict_f16() 210 f16x8_t vecDif; in arm_svm_rbf_predict_f16() [all …]
|
D | arm_svm_sigmoid_predict_f16.c | 73 f16x8_t vSum = vdupq_n_f16(0.0f); in arm_svm_sigmoid_predict_f16() 83 f16x8_t vecIn, acc0, acc1, acc2, acc3; in arm_svm_sigmoid_predict_f16() 112 f16x8_t vecA; in arm_svm_sigmoid_predict_f16() 138 f16x8_t vecA; in arm_svm_sigmoid_predict_f16() 153 f16x8_t vtmp = vuninitializedq_f16(); in arm_svm_sigmoid_predict_f16() 177 f16x8_t vecIn, acc0, acc1; in arm_svm_sigmoid_predict_f16() 199 f16x8_t vecA; in arm_svm_sigmoid_predict_f16() 219 f16x8_t vecA; in arm_svm_sigmoid_predict_f16() 230 f16x8_t vtmp = vuninitializedq_f16(); in arm_svm_sigmoid_predict_f16() 244 f16x8_t vecIn, acc0; in arm_svm_sigmoid_predict_f16() [all …]
|
D | arm_svm_polynomial_predict_f16.c | 73 f16x8_t vSum = vdupq_n_f16(0.0f); in arm_svm_polynomial_predict_f16() 83 f16x8_t vecIn, acc0, acc1, acc2, acc3; in arm_svm_polynomial_predict_f16() 112 f16x8_t vecA; in arm_svm_polynomial_predict_f16() 138 f16x8_t vecA; in arm_svm_polynomial_predict_f16() 153 f16x8_t vtmp = vuninitializedq_f16(); in arm_svm_polynomial_predict_f16() 178 f16x8_t vecIn, acc0, acc1; in arm_svm_polynomial_predict_f16() 200 f16x8_t vecA; in arm_svm_polynomial_predict_f16() 220 f16x8_t vecA; in arm_svm_polynomial_predict_f16() 231 f16x8_t vtmp = vuninitializedq_f16(); in arm_svm_polynomial_predict_f16() 246 f16x8_t vecIn, acc0; in arm_svm_polynomial_predict_f16() [all …]
|
D | arm_svm_linear_predict_f16.c | 80 f16x8_t vecIn, acc0, acc1, acc2, acc3; in arm_svm_linear_predict_f16() 109 f16x8_t vecA; in arm_svm_linear_predict_f16() 135 f16x8_t vecA; in arm_svm_linear_predict_f16() 169 f16x8_t vecIn, acc0, acc1; in arm_svm_linear_predict_f16() 191 f16x8_t vecA; in arm_svm_linear_predict_f16() 211 f16x8_t vecA; in arm_svm_linear_predict_f16() 232 f16x8_t vecIn, acc0; in arm_svm_linear_predict_f16() 252 f16x8_t vecA; in arm_svm_linear_predict_f16() 269 f16x8_t vecA; in arm_svm_linear_predict_f16()
|
/hal_nxp-3.5.0/mcux/mcux-sdk/CMSIS/DSP/Source/MatrixFunctions/ |
D | arm_mat_vec_mult_f16.c | 79 f16x8_t vecIn, acc0, acc1, acc2, acc3; in arm_mat_vec_mult_f16() 109 f16x8_t vecA; in arm_mat_vec_mult_f16() 136 f16x8_t vecA; in arm_mat_vec_mult_f16() 169 f16x8_t vecIn, acc0, acc1; in arm_mat_vec_mult_f16() 192 f16x8_t vecA; in arm_mat_vec_mult_f16() 213 f16x8_t vecA; in arm_mat_vec_mult_f16() 233 f16x8_t vecIn, acc0; in arm_mat_vec_mult_f16() 254 f16x8_t vecA; in arm_mat_vec_mult_f16() 272 f16x8_t vecA; in arm_mat_vec_mult_f16()
|
D | arm_mat_inverse_f16.c | 217 f16x8_t vecA, vecB; in arm_mat_inverse_f16() 331 f16x8_t invIn = vdupq_n_f16(1.0f16 / in); in arm_mat_inverse_f16() 334 f16x8_t vecA; in arm_mat_inverse_f16() 337 *(f16x8_t *) pTmpA = *(f16x8_t *) pTmpA * invIn; in arm_mat_inverse_f16() 368 *(f16x8_t *) pTmpA = *(f16x8_t *) pTmpA *invIn; in arm_mat_inverse_f16() 433 f16x8_t tmpV = vdupq_n_f16(in); in arm_mat_inverse_f16() 438 f16x8_t vec1, vec2; in arm_mat_inverse_f16() 461 f16x8_t vec1, vec2; in arm_mat_inverse_f16() 474 f16x8_t vec1, vec2; in arm_mat_inverse_f16() 498 f16x8_t vec1, vec2; in arm_mat_inverse_f16()
|
D | arm_mat_cmplx_mult_f16.c | 76 f16x8_t acc0, acc1; in arm_mat_cmplx_mult_f16_2x2_mve() 77 f16x8_t vecB, vecA0, vecA1; in arm_mat_cmplx_mult_f16_2x2_mve() 78 f16x8_t vecTmp; in arm_mat_cmplx_mult_f16_2x2_mve() 117 vecTmp = (f16x8_t) vrev64q_s32((int32x4_t) acc0); in arm_mat_cmplx_mult_f16_2x2_mve() 124 vecTmp = (f16x8_t) vrev64q_s32((int32x4_t) acc1); in arm_mat_cmplx_mult_f16_2x2_mve() 151 f16x8_t acc0, acc1, acc2; in arm_mat_cmplx_mult_f16_3x3_mve() 152 f16x8_t vecB, vecA0, vecA1, vecA2; in arm_mat_cmplx_mult_f16_3x3_mve() 248 f16x8_t acc0, acc1, acc2, acc3; in arm_mat_cmplx_mult_f16_4x4_mve() 249 f16x8_t vecB, vecA; in arm_mat_cmplx_mult_f16_4x4_mve() 480 f16x8_t acc0, acc1, acc2, acc3; in arm_mat_cmplx_mult_f16() [all …]
|
D | arm_mat_mult_f16.c | 64 f16x8_t vecInA, vecInB, vecDst; in arm_mat_mult_f16_2x2_mve() 133 f16x8_t vecInA, vecInB, vecDst; in arm_mat_mult_f16_3x3_mve() 231 f16x8_t vecInA, vecInB, vecDst0, vecDst1; in arm_mat_mult_f16_4x4_mve() 428 f16x8_t vecMac0, vecMac1, vecMac2, vecMac3; in arm_mat_mult_f16() 429 f16x8_t vecInB; in arm_mat_mult_f16() 459 vecInB = *(f16x8_t *)pInB0; /* vldrhq_f16(pInB0, 0); */ in arm_mat_mult_f16() 544 f16x8_t vecInB; in arm_mat_mult_f16() 545 f16x8_t vecMac0; in arm_mat_mult_f16() 563 vecInB = *(f16x8_t *)pInB0; /* vldrhq_f16(pInB0, 0); */ in arm_mat_mult_f16()
|
/hal_nxp-3.5.0/mcux/mcux-sdk/CMSIS/DSP/Source/StatisticsFunctions/ |
D | arm_kullback_leibler_f16.c | 84 f16x8_t vSum = vdupq_n_f16(0.0f); in arm_kullback_leibler_f16() 88 f16x8_t vecA = vld1q(pSrcA); in arm_kullback_leibler_f16() 89 f16x8_t vecB = vld1q(pSrcB); in arm_kullback_leibler_f16() 90 f16x8_t vRatio; in arm_kullback_leibler_f16()
|
D | arm_logsumexp_f16.c | 88 f16x8_t vSum = vdupq_n_f16(0.0f16); in arm_logsumexp_f16() 92 f16x8_t vecIn = vld1q(pIn); in arm_logsumexp_f16() 93 f16x8_t vecExp; in arm_logsumexp_f16()
|
/hal_nxp-3.5.0/mcux/mcux-sdk/CMSIS/DSP/Source/BasicMathFunctions/ |
D | arm_abs_f16.c | 74 f16x8_t vec1; in arm_abs_f16() 75 f16x8_t res; in arm_abs_f16() 122 f16x8_t vec1; in arm_abs_f16() 123 f16x8_t res; in arm_abs_f16()
|
D | arm_add_f16.c | 73 f16x8_t vec1; in arm_add_f16() 74 f16x8_t vec2; in arm_add_f16() 75 f16x8_t res; in arm_add_f16()
|
D | arm_mult_f16.c | 74 f16x8_t vec1; in arm_mult_f16() 75 f16x8_t vec2; in arm_mult_f16() 76 f16x8_t res; in arm_mult_f16()
|
D | arm_sub_f16.c | 72 f16x8_t vec1; in arm_sub_f16() 73 f16x8_t vec2; in arm_sub_f16() 74 f16x8_t res; in arm_sub_f16()
|
/hal_nxp-3.5.0/mcux/mcux-sdk/CMSIS/DSP/Source/ComplexMathFunctions/ |
D | arm_cmplx_mag_f16.c | 86 f16x8_t sum; in arm_cmplx_mag_f16() 93 f16x8_t sumHalf, invSqrt; in arm_cmplx_mag_f16() 114 INVSQRT_NEWTON_MVE_F16(invSqrt, sumHalf, (f16x8_t) newtonStartVec); in arm_cmplx_mag_f16() 140 f16x8_t sumHalf, invSqrt; in arm_cmplx_mag_f16() 156 INVSQRT_NEWTON_MVE_F16(invSqrt, sumHalf, (f16x8_t) newtonStartVec); in arm_cmplx_mag_f16()
|
D | arm_cmplx_conj_f16.c | 82 f16x8_t vecSrc; in arm_cmplx_conj_f16() 83 f16x8_t vecSign; in arm_cmplx_conj_f16() 88 vecSign = *(f16x8_t *) cmplx_conj_sign; in arm_cmplx_conj_f16()
|
D | arm_cmplx_mult_real_f16.c | 87 f16x8_t rVec; in arm_cmplx_mult_real_f16() 88 f16x8_t cmplxVec; in arm_cmplx_mult_real_f16() 89 f16x8_t dstVec; in arm_cmplx_mult_real_f16()
|
/hal_nxp-3.5.0/mcux/mcux-sdk/CMSIS/DSP/Source/BayesFunctions/ |
D | arm_gaussian_naive_bayes_predict_f16.c | 72 f16x8_t vsigma; in arm_gaussian_naive_bayes_predict_f16() 74 f16x8_t vacc1, vacc2; in arm_gaussian_naive_bayes_predict_f16() 92 f16x8_t vinvSigma, vtmp; in arm_gaussian_naive_bayes_predict_f16() 114 f16x8_t vinvSigma, vtmp; in arm_gaussian_naive_bayes_predict_f16()
|
/hal_nxp-3.5.0/mcux/mcux-sdk/CMSIS/DSP/Source/FilteringFunctions/ |
D | arm_biquad_cascade_df2T_f16.c | 65 f16x8_t b0Coeffs, a0Coeffs; /* Coefficients vector */ in arm_biquad_cascade_df2T_f16() 66 f16x8_t b1Coeffs, a1Coeffs; /* Modified coef. vector */ in arm_biquad_cascade_df2T_f16() 67 f16x8_t state; /* State vector */ in arm_biquad_cascade_df2T_f16() 86 state = *(f16x8_t *) pState; in arm_biquad_cascade_df2T_f16() 92 b1Coeffs = (f16x8_t)vshlcq_s16((int16x8_t)b0Coeffs, &tmp, 16); in arm_biquad_cascade_df2T_f16() 93 a1Coeffs = (f16x8_t)vshlcq_s16((int16x8_t)a0Coeffs, &tmp, 16); in arm_biquad_cascade_df2T_f16()
|
D | arm_biquad_cascade_stereo_df2T_f16.c | 68 f16x8_t aCoeffs, bCoeffs; in arm_biquad_cascade_stereo_df2T_f16() 73 f16x8_t stateVec0, stateVec1; in arm_biquad_cascade_stereo_df2T_f16() 74 f16x8_t inVec; in arm_biquad_cascade_stereo_df2T_f16() 118 *(f16x8_t *) scratch = *(f16x8_t *) pState; in arm_biquad_cascade_stereo_df2T_f16() 165 *(f16x8_t *) scratch = stateVec1; in arm_biquad_cascade_stereo_df2T_f16()
|
D | arm_levinson_durbin_f16.c | 79 f16x8_t vecA,vecRevPhi,vecPhi,vecSumA, vecSumB; in arm_levinson_durbin_f16() 131 f16x8_t vecRevA,tmp; in arm_levinson_durbin_f16() 167 vecRevA = (f16x8_t) vsetq_lane_u64(tmpb,(uint64x2_t)vecRevA,0); in arm_levinson_durbin_f16() 168 vecRevA = (f16x8_t) vsetq_lane_u64(tmpa,(uint64x2_t)vecRevA,1); in arm_levinson_durbin_f16()
|
/hal_nxp-3.5.0/mcux/mcux-sdk/CMSIS/DSP/Source/FastMathFunctions/ |
D | arm_vinverse_f16.c | 46 f16x8_t src; in arm_vinverse_f16() 47 f16x8_t dst; in arm_vinverse_f16()
|
D | arm_vlog_f16.c | 45 f16x8_t src; in arm_vlog_f16() 46 f16x8_t dst; in arm_vlog_f16()
|