/third_party/cmsis/CMSIS/DSP/Source/MatrixFunctions/ |
D | arm_mat_cmplx_mult_q31.c | 79 q31x4_t vecB, vecA; in arm_mat_cmplx_mult_q31_2x2_mve() local 91 vecA = vldrwq_s32(pInA0); in arm_mat_cmplx_mult_q31_2x2_mve() 92 acc0 = vmlsldavq_s32(vecA, vecB); in arm_mat_cmplx_mult_q31_2x2_mve() 93 acc1 = vmlaldavxq_s32(vecA, vecB); in arm_mat_cmplx_mult_q31_2x2_mve() 95 vecA = vldrwq_s32(pInA1); in arm_mat_cmplx_mult_q31_2x2_mve() 96 acc2 = vmlsldavq_s32(vecA, vecB); in arm_mat_cmplx_mult_q31_2x2_mve() 97 acc3 = vmlaldavxq_s32(vecA, vecB); in arm_mat_cmplx_mult_q31_2x2_mve() 109 vecA = vldrwq_s32(pInA0); in arm_mat_cmplx_mult_q31_2x2_mve() 110 acc0 = vmlsldavq_s32(vecA, vecB); in arm_mat_cmplx_mult_q31_2x2_mve() 111 acc1 = vmlaldavxq_s32(vecA, vecB); in arm_mat_cmplx_mult_q31_2x2_mve() [all …]
|
D | arm_mat_cmplx_mult_f32.c | 83 f32x4_t vecB, vecA; in arm_mat_cmplx_mult_f32_2x2_mve() local 95 vecA = vldrwq_f32(pInA0); in arm_mat_cmplx_mult_f32_2x2_mve() 96 acc0 = vcmulq(vecA, vecB); in arm_mat_cmplx_mult_f32_2x2_mve() 97 acc0 = vcmlaq_rot90(acc0, vecA, vecB); in arm_mat_cmplx_mult_f32_2x2_mve() 99 vecA = vldrwq_f32(pInA1); in arm_mat_cmplx_mult_f32_2x2_mve() 100 acc1 = vcmulq(vecA, vecB); in arm_mat_cmplx_mult_f32_2x2_mve() 101 acc1 = vcmlaq_rot90(acc1, vecA, vecB); in arm_mat_cmplx_mult_f32_2x2_mve() 116 vecA = vldrwq_f32(pInA0); in arm_mat_cmplx_mult_f32_2x2_mve() 117 acc0 = vcmulq(vecA, vecB); in arm_mat_cmplx_mult_f32_2x2_mve() 118 acc0 = vcmlaq_rot90(acc0, vecA, vecB); in arm_mat_cmplx_mult_f32_2x2_mve() [all …]
|
D | arm_mat_mult_q31.c | 139 q31x4_t vecB, vecA; in arm_mat_mult_q31_3x3_mve() local 150 vecA = vldrwq_s32(pInA0); in arm_mat_mult_q31_3x3_mve() 151 acc0 = vrmlaldavhq(vecA, vecB); in arm_mat_mult_q31_3x3_mve() 152 vecA = vldrwq_s32(pInA1); in arm_mat_mult_q31_3x3_mve() 153 acc1 = vrmlaldavhq(vecA, vecB); in arm_mat_mult_q31_3x3_mve() 154 vecA = vldrwq_s32(pInA2); in arm_mat_mult_q31_3x3_mve() 155 acc2 = vrmlaldavhq(vecA, vecB); in arm_mat_mult_q31_3x3_mve() 171 vecA = vldrwq_s32(pInA0); in arm_mat_mult_q31_3x3_mve() 172 acc0 = vrmlaldavhq(vecA, vecB); in arm_mat_mult_q31_3x3_mve() 173 vecA = vldrwq_s32(pInA1); in arm_mat_mult_q31_3x3_mve() [all …]
|
D | arm_mat_cmplx_mult_q15.c | 136 q15x8_t vecA, vecB, vecB2; in arm_mat_cmplx_mult_q15() local 170 vecA = vld1q(pSrcAVec); in arm_mat_cmplx_mult_q15() 174 acc0 = vmlsldavaq(acc0, vecA, vecB); in arm_mat_cmplx_mult_q15() 175 acc1 = vmlaldavaxq(acc1, vecA, vecB); in arm_mat_cmplx_mult_q15() 182 acc2 = vmlsldavaq(acc2, vecA, vecB2); in arm_mat_cmplx_mult_q15() 183 acc3 = vmlaldavaxq(acc3, vecA, vecB2); in arm_mat_cmplx_mult_q15() 197 vecA = vldrhq_z_s16(pSrcAVec, p0); in arm_mat_cmplx_mult_q15() 199 acc0 = vmlsldavaq(acc0, vecA, vecB); in arm_mat_cmplx_mult_q15() 200 acc1 = vmlaldavaxq(acc1, vecA, vecB); in arm_mat_cmplx_mult_q15() 208 acc2 = vmlsldavaq(acc2, vecA, vecB2); in arm_mat_cmplx_mult_q15() [all …]
|
D | arm_mat_mult_q15.c | 430 q15x8_t vecB, vecA; in arm_mat_mult_q15() local 435 vecA = vld1q(pSrcA0Vec); pSrcA0Vec += 8; in arm_mat_mult_q15() 436 acc0 = vmlaldavaq(acc0, vecA, vecB); in arm_mat_mult_q15() 437 vecA = vld1q(pSrcA1Vec); pSrcA1Vec += 8; in arm_mat_mult_q15() 438 acc1 = vmlaldavaq(acc1, vecA, vecB); in arm_mat_mult_q15() 439 vecA = vld1q(pSrcA2Vec); pSrcA2Vec += 8; in arm_mat_mult_q15() 440 acc2 = vmlaldavaq(acc2, vecA, vecB); in arm_mat_mult_q15() 441 vecA = vld1q(pSrcA3Vec); pSrcA3Vec += 8; in arm_mat_mult_q15() 442 acc3 = vmlaldavaq(acc3, vecA, vecB); in arm_mat_mult_q15() 453 q15x8_t vecB, vecA; in arm_mat_mult_q15() local [all …]
|
D | arm_mat_inverse_f32.c | 237 f32x4_t vecA, vecB; in arm_mat_inverse_f32() local 248 vecA = vldrwq_f32(pTmpA); in arm_mat_inverse_f32() 250 vstrwq_f32(pTmpB, vecA); in arm_mat_inverse_f32() 269 vecA = vldrwq_f32(pTmpA); in arm_mat_inverse_f32() 271 vstrwq_p_f32(pTmpB, vecA, p0); in arm_mat_inverse_f32() 283 vecA = vldrwq_f32(pTmpA); in arm_mat_inverse_f32() 285 vstrwq_f32(pTmpB, vecA); in arm_mat_inverse_f32() 302 vecA = vldrwq_f32(pTmpA); in arm_mat_inverse_f32() 304 vstrwq_p_f32(pTmpB, vecA, p0); in arm_mat_inverse_f32() 357 f32x4_t vecA; in arm_mat_inverse_f32() local [all …]
|
/third_party/cmsis/CMSIS/DSP/Source/SVMFunctions/ |
D | arm_svm_linear_predict_f32.c | 104 f32x4_t vecA; in arm_svm_linear_predict_f32() local 108 vecA = vld1q(pSrcA0Vec); in arm_svm_linear_predict_f32() 110 acc0 = vfmaq(acc0, vecIn, vecA); in arm_svm_linear_predict_f32() 111 vecA = vld1q(pSrcA1Vec); in arm_svm_linear_predict_f32() 113 acc1 = vfmaq(acc1, vecIn, vecA); in arm_svm_linear_predict_f32() 114 vecA = vld1q(pSrcA2Vec); in arm_svm_linear_predict_f32() 116 acc2 = vfmaq(acc2, vecIn, vecA); in arm_svm_linear_predict_f32() 117 vecA = vld1q(pSrcA3Vec); in arm_svm_linear_predict_f32() 119 acc3 = vfmaq(acc3, vecIn, vecA); in arm_svm_linear_predict_f32() 130 f32x4_t vecA; in arm_svm_linear_predict_f32() local [all …]
|
D | arm_svm_sigmoid_predict_f32.c | 107 f32x4_t vecA; in arm_svm_sigmoid_predict_f32() local 111 vecA = vld1q(pSrcA0Vec); in arm_svm_sigmoid_predict_f32() 113 acc0 = vfmaq(acc0, vecIn, vecA); in arm_svm_sigmoid_predict_f32() 114 vecA = vld1q(pSrcA1Vec); in arm_svm_sigmoid_predict_f32() 116 acc1 = vfmaq(acc1, vecIn, vecA); in arm_svm_sigmoid_predict_f32() 117 vecA = vld1q(pSrcA2Vec); in arm_svm_sigmoid_predict_f32() 119 acc2 = vfmaq(acc2, vecIn, vecA); in arm_svm_sigmoid_predict_f32() 120 vecA = vld1q(pSrcA3Vec); in arm_svm_sigmoid_predict_f32() 122 acc3 = vfmaq(acc3, vecIn, vecA); in arm_svm_sigmoid_predict_f32() 133 f32x4_t vecA; in arm_svm_sigmoid_predict_f32() local [all …]
|
D | arm_svm_polynomial_predict_f32.c | 110 f32x4_t vecA; in arm_svm_polynomial_predict_f32() local 114 vecA = vld1q(pSrcA0Vec); in arm_svm_polynomial_predict_f32() 116 acc0 = vfmaq(acc0, vecIn, vecA); in arm_svm_polynomial_predict_f32() 117 vecA = vld1q(pSrcA1Vec); in arm_svm_polynomial_predict_f32() 119 acc1 = vfmaq(acc1, vecIn, vecA); in arm_svm_polynomial_predict_f32() 120 vecA = vld1q(pSrcA2Vec); in arm_svm_polynomial_predict_f32() 122 acc2 = vfmaq(acc2, vecIn, vecA); in arm_svm_polynomial_predict_f32() 123 vecA = vld1q(pSrcA3Vec); in arm_svm_polynomial_predict_f32() 125 acc3 = vfmaq(acc3, vecIn, vecA); in arm_svm_polynomial_predict_f32() 136 f32x4_t vecA; in arm_svm_polynomial_predict_f32() local [all …]
|
D | arm_svm_rbf_predict_f32.c | 107 f32x4_t vecA; in arm_svm_rbf_predict_f32() local 112 vecA = vld1q(pSrcA0Vec); in arm_svm_rbf_predict_f32() 114 vecDif = vsubq(vecIn, vecA); in arm_svm_rbf_predict_f32() 116 vecA = vld1q(pSrcA1Vec); in arm_svm_rbf_predict_f32() 118 vecDif = vsubq(vecIn, vecA); in arm_svm_rbf_predict_f32() 120 vecA = vld1q(pSrcA2Vec); in arm_svm_rbf_predict_f32() 122 vecDif = vsubq(vecIn, vecA); in arm_svm_rbf_predict_f32() 124 vecA = vld1q(pSrcA3Vec); in arm_svm_rbf_predict_f32() 126 vecDif = vsubq(vecIn, vecA); in arm_svm_rbf_predict_f32() 138 f32x4_t vecA; in arm_svm_rbf_predict_f32() local [all …]
|
/third_party/cmsis/CMSIS/DSP/Source/BasicMathFunctions/ |
D | arm_add_q31.c | 64 q31x4_t vecA; in arm_add_q31() local 75 vecA = vld1q(pSrcA); in arm_add_q31() 77 vst1q(pDst, vqaddq(vecA, vecB)); in arm_add_q31() 96 vecA = vld1q(pSrcA); in arm_add_q31() 98 vstrwq_p(pDst, vqaddq(vecA, vecB), p0); in arm_add_q31()
|
D | arm_sub_q31.c | 64 q31x4_t vecA; in arm_sub_q31() local 75 vecA = vld1q(pSrcA); in arm_sub_q31() 77 vst1q(pDst, vqsubq(vecA, vecB)); in arm_sub_q31() 96 vecA = vld1q(pSrcA); in arm_sub_q31() 98 vstrwq_p(pDst, vqsubq(vecA, vecB), p0); in arm_sub_q31()
|
D | arm_dot_prod_q31.c | 68 q31x4_t vecA; in arm_dot_prod_q31() local 80 vecA = vld1q(pSrcA); in arm_dot_prod_q31() 82 sum = vrmlaldavhaq(sum, vecA, vecB); in arm_dot_prod_q31() 100 vecA = vld1q(pSrcA); in arm_dot_prod_q31() 102 sum = vrmlaldavhaq_p(sum, vecA, vecB, p0); in arm_dot_prod_q31()
|
D | arm_add_q7.c | 64 q7x16_t vecA; in arm_add_q7() local 75 vecA = vld1q(pSrcA); in arm_add_q7() 77 vst1q(pDst, vqaddq(vecA, vecB)); in arm_add_q7() 96 vecA = vld1q(pSrcA); in arm_add_q7() 98 vstrbq_p(pDst, vqaddq(vecA, vecB), p0); in arm_add_q7()
|
D | arm_sub_q7.c | 63 q7x16_t vecA; in arm_sub_q7() local 74 vecA = vld1q(pSrcA); in arm_sub_q7() 76 vst1q(pDst, vqsubq(vecA, vecB)); in arm_sub_q7() 95 vecA = vld1q(pSrcA); in arm_sub_q7() 97 vstrbq_p(pDst, vqsubq(vecA, vecB), p0); in arm_sub_q7()
|
D | arm_mult_q31.c | 63 q31x4_t vecA, vecB; in arm_mult_q31() local 73 vecA = vld1q(pSrcA); in arm_mult_q31() 75 vst1q(pDst, vqdmulhq(vecA, vecB)); in arm_mult_q31() 94 vecA = vld1q(pSrcA); in arm_mult_q31() 96 vstrwq_p(pDst, vqdmulhq(vecA, vecB), p0); in arm_mult_q31()
|
D | arm_dot_prod_q15.c | 66 q15x8_t vecA; in arm_dot_prod_q15() local 78 vecA = vld1q(pSrcA); in arm_dot_prod_q15() 80 sum = vmlaldavaq(sum, vecA, vecB); in arm_dot_prod_q15() 98 vecA = vld1q(pSrcA); in arm_dot_prod_q15() 100 sum = vmlaldavaq_p(sum, vecA, vecB, p0); in arm_dot_prod_q15()
|
D | arm_sub_q15.c | 64 q15x8_t vecA; in arm_sub_q15() local 75 vecA = vld1q(pSrcA); in arm_sub_q15() 77 vst1q(pDst, vqsubq(vecA, vecB)); in arm_sub_q15() 96 vecA = vld1q(pSrcA); in arm_sub_q15() 98 vstrhq_p(pDst, vqsubq(vecA, vecB), p0); in arm_sub_q15()
|
D | arm_mult_q7.c | 63 q7x16_t vecA, vecB; in arm_mult_q7() local 73 vecA = vld1q(pSrcA); in arm_mult_q7() 75 vst1q(pDst, vqdmulhq(vecA, vecB)); in arm_mult_q7() 94 vecA = vld1q(pSrcA); in arm_mult_q7() 96 vstrbq_p(pDst, vqdmulhq(vecA, vecB), p0); in arm_mult_q7()
|
D | arm_add_q15.c | 64 q15x8_t vecA; in arm_add_q15() local 75 vecA = vld1q(pSrcA); in arm_add_q15() 77 vst1q(pDst, vqaddq(vecA, vecB)); in arm_add_q15() 96 vecA = vld1q(pSrcA); in arm_add_q15() 98 vstrhq_p(pDst, vqaddq(vecA, vecB), p0); in arm_add_q15()
|
D | arm_dot_prod_q7.c | 67 q7x16_t vecA; in arm_dot_prod_q7() local 79 vecA = vld1q(pSrcA); in arm_dot_prod_q7() 81 sum = vmladavaq(sum, vecA, vecB); in arm_dot_prod_q7() 99 vecA = vld1q(pSrcA); in arm_dot_prod_q7() 101 sum = vmladavaq_p(sum, vecA, vecB, p0); in arm_dot_prod_q7()
|
D | arm_dot_prod_f32.c | 73 f32x4_t vecA, vecB; in arm_dot_prod_f32() local 88 vecA = vld1q(pSrcA); in arm_dot_prod_f32() 94 vecSum = vfmaq(vecSum, vecA, vecB); in arm_dot_prod_f32() 108 vecA = vld1q(pSrcA); in arm_dot_prod_f32() 110 vecSum = vfmaq_m(vecSum, vecA, vecB, p0); in arm_dot_prod_f32()
|
/third_party/cmsis/CMSIS/DSP/Source/ComplexMathFunctions/ |
D | arm_cmplx_mag_squared_f32.c | 131 float32x4x2_t vecA; in arm_cmplx_mag_squared_f32() local 148 vecA = vld2q_f32(pSrc); in arm_cmplx_mag_squared_f32() 151 vRealA = vmulq_f32(vecA.val[0], vecA.val[0]); in arm_cmplx_mag_squared_f32() 152 vImagA = vmulq_f32(vecA.val[1], vecA.val[1]); in arm_cmplx_mag_squared_f32()
|
D | arm_cmplx_mag_f32.c | 166 float32x4x2_t vecA; in arm_cmplx_mag_f32() local 183 vecA = vld2q_f32(pSrc); in arm_cmplx_mag_f32() 189 vRealA = vmulq_f32(vecA.val[0], vecA.val[0]); in arm_cmplx_mag_f32() 190 vImagA = vmulq_f32(vecA.val[1], vecA.val[1]); in arm_cmplx_mag_f32()
|
/third_party/cmsis/CMSIS/DSP/Source/TransformFunctions/ |
D | arm_cfft_q31.c | 76 q31x4_t vecA, vecB, vecC, vecD; in _arm_radix4_butterfly_q31_mve() local 122 vecA = vldrwq_s32(inA); in _arm_radix4_butterfly_q31_mve() 129 vecSum0 = vhaddq(vecA, vecC); in _arm_radix4_butterfly_q31_mve() 130 vecDiff0 = vhsubq(vecA, vecC); in _arm_radix4_butterfly_q31_mve() 178 vecA = vldrwq_s32(inA); in _arm_radix4_butterfly_q31_mve() 207 vecA = vldrwq_gather_base_wb_s32(&vecScGathAddr, 64); in _arm_radix4_butterfly_q31_mve() 213 vecSum0 = vhaddq(vecA, vecC); in _arm_radix4_butterfly_q31_mve() 214 vecDiff0 = vhsubq(vecA, vecC); in _arm_radix4_butterfly_q31_mve() 224 vecA = vldrwq_gather_base_wb_s32(&vecScGathAddr, 64); in _arm_radix4_butterfly_q31_mve() 326 q31x4_t vecA, vecB, vecC, vecD; in _arm_radix4_butterfly_inverse_q31_mve() local [all …]
|