/hal_nxp-3.5.0/mcux/mcux-sdk/CMSIS/DSP/Source/MatrixFunctions/ |
D | arm_mat_mult_fast_q15.c | 87 q31_t inA1, inB1, inA2, inB2; in arm_mat_mult_fast_q15() local 93 q15_t inA1, inB1, inA2, inB2; in arm_mat_mult_fast_q15() local 277 inA2 = read_q15x2_ia ((q15_t **) &pInA2); in arm_mat_mult_fast_q15() 283 sum3 = __SMLAD(inA2, inB1, sum3); in arm_mat_mult_fast_q15() 284 sum4 = __SMLAD(inA2, inB2, sum4); in arm_mat_mult_fast_q15() 292 inA2 = *pInA++; in arm_mat_mult_fast_q15() 294 sum += inA2 * inB2; in arm_mat_mult_fast_q15() 300 inA2 = *pInA++; in arm_mat_mult_fast_q15() 302 sum += inA2 * inB2; in arm_mat_mult_fast_q15() 314 inA2 = *pInA2++; in arm_mat_mult_fast_q15() [all …]
|
D | arm_mat_mult_fast_q31.c | 77 q31_t inA1, inA2, inB1, inB2; in arm_mat_mult_fast_q31() local 142 inA2 = *pInA2++; in arm_mat_mult_fast_q31() 149 sum3 = __SMMLA(inA2, inB1, sum3); in arm_mat_mult_fast_q31() 150 sum4 = __SMMLA(inA2, inB2, sum4); in arm_mat_mult_fast_q31() 154 sum3 = (q31_t) ((((q63_t) sum3 << 32) + ((q63_t) inA2 * inB1)) >> 32); in arm_mat_mult_fast_q31() 155 sum4 = (q31_t) ((((q63_t) sum4 << 32) + ((q63_t) inA2 * inB2)) >> 32); in arm_mat_mult_fast_q31() 303 inA2 = *pInA++; in arm_mat_mult_fast_q31() 310 sum1 = __SMMLA(inA2, inB2, sum1); in arm_mat_mult_fast_q31() 313 sum1 = (q31_t) ((((q63_t) sum1 << 32) + ((q63_t) inA2 * inB2)) >> 32); in arm_mat_mult_fast_q31() 317 inA2 = *pInA++; in arm_mat_mult_fast_q31() [all …]
|
D | arm_mat_scale_q15.c | 147 q31_t inA1, inA2; in arm_mat_scale_q15() local 181 inA2 = read_q15x2_ia ((q15_t **) &pIn); in arm_mat_scale_q15() 187 out3 = (q31_t) ((q15_t) (inA2 >> 16) * scaleFract); in arm_mat_scale_q15() 188 out4 = (q31_t) ((q15_t) (inA2 ) * scaleFract); in arm_mat_scale_q15()
|
/hal_nxp-3.5.0/mcux/mcux-sdk/CMSIS/DSP/Source/ComplexMathFunctions/ |
D | arm_cmplx_mult_real_q15.c | 120 q31_t inA1, inA2; /* Temporary variables to hold input data */ in arm_cmplx_mult_real_q15() local 137 inA2 = read_q15x2_ia ((q15_t **) &pSrcCmplx); in arm_cmplx_mult_real_q15() 145 mul3 = (q31_t) ((q15_t) (inA2) * (q15_t) (inB1 >> 16)); in arm_cmplx_mult_real_q15() 146 mul4 = (q31_t) ((q15_t) (inA2 >> 16) * (q15_t) (inB1 >> 16)); in arm_cmplx_mult_real_q15() 150 mul4 = (q31_t) ((q15_t) (inA2 >> 16) * (q15_t) inB1); in arm_cmplx_mult_real_q15() 151 mul3 = (q31_t) ((q15_t) inA2 * (q15_t) inB1); in arm_cmplx_mult_real_q15() 165 inA2 = read_q15x2_ia ((q15_t **) &pSrcCmplx); in arm_cmplx_mult_real_q15() 171 mul3 = (q31_t) ((q15_t) (inA2) * (q15_t) (inB1 >> 16)); in arm_cmplx_mult_real_q15() 172 mul4 = (q31_t) ((q15_t) (inA2 >> 16) * (q15_t) (inB1 >> 16)); in arm_cmplx_mult_real_q15() 176 mul4 = (q31_t) ((q15_t) (inA2 >> 16) * (q15_t) inB1); in arm_cmplx_mult_real_q15() [all …]
|
/hal_nxp-3.5.0/mcux/mcux-sdk/CMSIS/NN/Source/ConvolutionFunctions/ |
D | arm_depthwise_separable_conv_HWC_q7.c | 164 q31_t inA1, inA2, inB1, inB2, opA, opB; in arm_depthwise_separable_conv_HWC_q7() local 176 inA2 = __PKHTB(opB, inA1, 16); in arm_depthwise_separable_conv_HWC_q7() 184 opA = __SXTB16(inA2); in arm_depthwise_separable_conv_HWC_q7() 187 opA = __SXTB16(__ROR(inA2, 8)); in arm_depthwise_separable_conv_HWC_q7() 196 q31_t inA1, inA2, inB1, inB2, opA, opB; in arm_depthwise_separable_conv_HWC_q7() local 208 inA2 = __PKHBT(opB, inA1, 16); in arm_depthwise_separable_conv_HWC_q7() 216 opA = __SXTB16(inA2); in arm_depthwise_separable_conv_HWC_q7() 219 opA = __SXTB16(__ROR(inA2, 8)); in arm_depthwise_separable_conv_HWC_q7()
|
D | arm_depthwise_separable_conv_HWC_q7_nonsquare.c | 174 q31_t inA1, inA2, inB1, inB2, opA, opB; in arm_depthwise_separable_conv_HWC_q7_nonsquare() local 186 inA2 = __PKHTB(opB, inA1, 16); in arm_depthwise_separable_conv_HWC_q7_nonsquare() 194 opA = __SXTB16(inA2); in arm_depthwise_separable_conv_HWC_q7_nonsquare() 197 opA = __SXTB16(__ROR(inA2, 8)); in arm_depthwise_separable_conv_HWC_q7_nonsquare() 206 q31_t inA1, inA2, inB1, inB2, opA, opB; in arm_depthwise_separable_conv_HWC_q7_nonsquare() local 218 inA2 = __PKHBT(opB, inA1, 16); in arm_depthwise_separable_conv_HWC_q7_nonsquare() 226 opA = __SXTB16(inA2); in arm_depthwise_separable_conv_HWC_q7_nonsquare() 229 opA = __SXTB16(__ROR(inA2, 8)); in arm_depthwise_separable_conv_HWC_q7_nonsquare()
|
D | arm_convolve_HWC_q15_fast.c | 168 q31_t inA2 = arm_nn_read_q15x2_ia(&pA2); in arm_convolve_HWC_q15_fast() local 173 sum3 = __SMLAD(inA2, inB1, sum3); in arm_convolve_HWC_q15_fast() 174 sum4 = __SMLAD(inA2, inB2, sum4); in arm_convolve_HWC_q15_fast() 183 q15_t inA2 = *pA2++; in arm_convolve_HWC_q15_fast() local 188 sum3 += inA2 * inB1; in arm_convolve_HWC_q15_fast() 189 sum4 += inA2 * inB2; in arm_convolve_HWC_q15_fast()
|
D | arm_convolve_HWC_q15_fast_nonsquare.c | 180 q31_t inA2 = arm_nn_read_q15x2_ia(&pA2); in arm_convolve_HWC_q15_fast_nonsquare() local 185 sum3 = __SMLAD(inA2, inB1, sum3); in arm_convolve_HWC_q15_fast_nonsquare() 186 sum4 = __SMLAD(inA2, inB2, sum4); in arm_convolve_HWC_q15_fast_nonsquare() 195 q15_t inA2 = *pA2++; in arm_convolve_HWC_q15_fast_nonsquare() local 200 sum3 += inA2 * inB1; in arm_convolve_HWC_q15_fast_nonsquare() 201 sum4 += inA2 * inB2; in arm_convolve_HWC_q15_fast_nonsquare()
|
D | arm_nn_mat_mult_kernel_q7_q15_reordered.c | 103 q7_t inA2 = *pA2++; in arm_nn_mat_mult_kernel_q7_q15_reordered() local 108 sum3 += inA2 * inB1; in arm_nn_mat_mult_kernel_q7_q15_reordered() 109 sum4 += inA2 * inB2; in arm_nn_mat_mult_kernel_q7_q15_reordered()
|
D | arm_convolve_1x1_HWC_q7_fast_nonsquare.c | 157 q31_t inA1, inA2; in arm_convolve_1x1_HWC_q7_fast_nonsquare() local 160 pA = read_and_pad_reordered(pA, &inA1, &inA2); in arm_convolve_1x1_HWC_q7_fast_nonsquare() 166 sum = __SMLAD(inA2, inB2, sum); in arm_convolve_1x1_HWC_q7_fast_nonsquare()
|
D | arm_convolve_HWC_q7_basic.c | 159 q31_t inA1, inA2; in arm_convolve_HWC_q7_basic() local 162 pA = read_and_pad(pA, &inA1, &inA2); in arm_convolve_HWC_q7_basic() 168 sum = __SMLAD(inA2, inB2, sum); in arm_convolve_HWC_q7_basic()
|
D | arm_convolve_HWC_q7_basic_nonsquare.c | 160 q31_t inA1, inA2; in arm_convolve_HWC_q7_basic_nonsquare() local 163 pA = read_and_pad(pA, &inA1, &inA2); in arm_convolve_HWC_q7_basic_nonsquare() 169 sum = __SMLAD(inA2, inB2, sum); in arm_convolve_HWC_q7_basic_nonsquare()
|
D | arm_nn_mat_mult_kernel_q7_q15.c | 103 q7_t inA2 = *pA2++; in arm_nn_mat_mult_kernel_q7_q15() local 108 sum3 += inA2 * inB1; in arm_nn_mat_mult_kernel_q7_q15() 109 sum4 += inA2 * inB2; in arm_nn_mat_mult_kernel_q7_q15()
|
D | arm_convolve_HWC_q7_RGB.c | 206 q31_t inA1, inA2; in arm_convolve_HWC_q7_RGB() local 209 pA = read_and_pad(pA, &inA1, &inA2); in arm_convolve_HWC_q7_RGB() 214 sum = __SMLAD(inA2, inB2, sum); in arm_convolve_HWC_q7_RGB()
|
D | arm_convolve_HWC_q7_fast.c | 306 q31_t inA1, inA2; in arm_convolve_HWC_q7_fast() local 309 pA = read_and_pad_reordered(pA, &inA1, &inA2); in arm_convolve_HWC_q7_fast() 314 sum = __SMLAD(inA2, inB2, sum); in arm_convolve_HWC_q7_fast()
|
D | arm_convolve_HWC_q7_fast_nonsquare.c | 302 q31_t inA1, inA2; in arm_convolve_HWC_q7_fast_nonsquare() local 305 pA = read_and_pad_reordered(pA, &inA1, &inA2); in arm_convolve_HWC_q7_fast_nonsquare() 310 sum = __SMLAD(inA2, inB2, sum); in arm_convolve_HWC_q7_fast_nonsquare()
|
/hal_nxp-3.5.0/mcux/mcux-sdk/CMSIS/NN/Source/NNSupportFunctions/ |
D | arm_nn_mult_q15.c | 63 q31_t inA1, inA2, inB1, inB2; /* temporary input variables */ in arm_nn_mult_q15() local 79 inA2 = arm_nn_read_q15x2_ia((const q15_t **)&pSrcA); in arm_nn_mult_q15() 86 mul3 = (q31_t)((q15_t)(inA2 >> 16) * (q15_t)(inB2 >> 16)); in arm_nn_mult_q15() 87 mul4 = (q31_t)((q15_t)inA2 * (q15_t)inB2); in arm_nn_mult_q15()
|
/hal_nxp-3.5.0/mcux/mcux-sdk/CMSIS/DSP/Source/BasicMathFunctions/ |
D | arm_scale_q15.c | 122 q31_t inA1, inA2; in arm_scale_q15() local 140 inA2 = read_q15x2_ia ((q15_t **) &pSrc); in arm_scale_q15() 146 out3 = (q31_t) ((q15_t) (inA2 >> 16) * scaleFract); in arm_scale_q15() 147 out4 = (q31_t) ((q15_t) (inA2 ) * scaleFract); in arm_scale_q15()
|
D | arm_mult_q15.c | 112 q31_t inA1, inA2, inB1, inB2; /* Temporary input variables */ in arm_mult_q15() local 130 inA2 = read_q15x2_ia ((q15_t **) &pSrcA); in arm_mult_q15() 137 mul3 = (q31_t) ((q15_t) (inA2 >> 16) * (q15_t) (inB2 >> 16)); in arm_mult_q15() 138 mul4 = (q31_t) ((q15_t) (inA2 ) * (q15_t) (inB2 )); in arm_mult_q15()
|
D | arm_sub_q15.c | 115 q31_t inA1, inA2; in arm_sub_q15() local 129 inA2 = read_q15x2_ia ((q15_t **) &pSrcA); in arm_sub_q15() 136 write_q15x2_ia (&pDst, __QSUB16(inA2, inB2)); in arm_sub_q15()
|
D | arm_add_q15.c | 114 q31_t inA1, inA2; in arm_add_q15() local 128 inA2 = read_q15x2_ia ((q15_t **) &pSrcA); in arm_add_q15() 135 write_q15x2_ia (&pDst, __QADD16(inA2, inB2)); in arm_add_q15()
|
D | arm_dot_prod_q7.c | 120 q31_t inA1, inA2, inB1, inB2; /* Temporary variables */ in arm_dot_prod_q7() local 139 inA2 = __SXTB16(input1); in arm_dot_prod_q7() 147 sum = __SMLAD(inA2, inB2, sum); in arm_dot_prod_q7()
|
/hal_nxp-3.5.0/mcux/mcux-sdk/CMSIS/NN/Source/FullyConnectedFunctions/ |
D | arm_fully_connected_q15_opt.c | 268 q15_t inA2 = *pA++; in arm_fully_connected_q15_opt() local 272 sum += inA1 * inB1 + inA2 * inB2; in arm_fully_connected_q15_opt() 276 sum2 += inA1 * inB1 + inA2 * inB2; in arm_fully_connected_q15_opt() 280 sum3 += inA1 * inB1 + inA2 * inB2; in arm_fully_connected_q15_opt() 284 sum4 += inA1 * inB1 + inA2 * inB2; in arm_fully_connected_q15_opt()
|
D | arm_fully_connected_mat_q7_vec_q15_opt.c | 346 q15_t inA2 = *pA++; in arm_fully_connected_mat_q7_vec_q15_opt() local 353 sum += inA1 * inB1 + inA2 * inB2; in arm_fully_connected_mat_q7_vec_q15_opt() 354 sum2 += inA1 * inB3 + inA2 * inB4; in arm_fully_connected_mat_q7_vec_q15_opt() 361 sum3 += inA1 * inB1 + inA2 * inB2; in arm_fully_connected_mat_q7_vec_q15_opt() 362 sum4 += inA1 * inB3 + inA2 * inB4; in arm_fully_connected_mat_q7_vec_q15_opt()
|
D | arm_fully_connected_q7_opt.c | 406 q7_t inA2 = *pA++; in arm_fully_connected_q7_opt() 414 sum += inA1 * inB1 + inA2 * inB2; in arm_fully_connected_q7_opt() 415 sum2 += inA1 * inB3 + inA2 * inB4; in arm_fully_connected_q7_opt() 422 sum3 += inA1 * inB1 + inA2 * inB2; in arm_fully_connected_q7_opt() 423 sum4 += inA1 * inB3 + inA2 * inB4; in arm_fully_connected_q7_opt()
|