/cmsis-dsp-latest/Source/MatrixFunctions/ |
D | arm_mat_cmplx_mult_q31.c | 78 q63_t acc0, acc1, acc2, acc3; in arm_mat_cmplx_mult_q31_2x2_mve() local 97 acc3 = vmlaldavxq_s32(vecA, vecB); in arm_mat_cmplx_mult_q31_2x2_mve() 102 pOut[1 * CMPLX_DIM * MATRIX_DIM2 + 1] = (q31_t) asrl(acc3, 31); in arm_mat_cmplx_mult_q31_2x2_mve() 115 acc3 = vmlaldavxq_s32(vecA, vecB); in arm_mat_cmplx_mult_q31_2x2_mve() 122 pOut[1 * CMPLX_DIM * MATRIX_DIM2 + 1] = (q31_t) asrl(acc3, 31); in arm_mat_cmplx_mult_q31_2x2_mve() 141 q63_t acc0, acc1, acc2, acc3; in arm_mat_cmplx_mult_q31_3x3_mve() local 171 acc3 = vmlaldavxq_s32(vecA, vecB); in arm_mat_cmplx_mult_q31_3x3_mve() 179 acc3 = vmlaldavaxq_s32(acc3, vecA, vecB1); in arm_mat_cmplx_mult_q31_3x3_mve() 184 pOut[1 * CMPLX_DIM * MATRIX_DIM3 + 1] = (q31_t) asrl(acc3, 31); in arm_mat_cmplx_mult_q31_3x3_mve() 212 acc3 = vmlaldavxq_s32(vecA, vecB); in arm_mat_cmplx_mult_q31_3x3_mve() [all …]
|
D | arm_mat_cmplx_mult_f32.c | 305 f32x4_t acc0, acc1, acc2, acc3; in arm_mat_cmplx_mult_f32_4x4_mve() local 335 acc3 = vcmulq(vecA, vecB); in arm_mat_cmplx_mult_f32_4x4_mve() 336 acc3 = vcmlaq_rot90(acc3, vecA, vecB); in arm_mat_cmplx_mult_f32_4x4_mve() 353 acc3 = vcmlaq(acc3, vecA, vecB); in arm_mat_cmplx_mult_f32_4x4_mve() 354 acc3 = vcmlaq_rot90(acc3, vecA, vecB); in arm_mat_cmplx_mult_f32_4x4_mve() 362 pOut[3 * CMPLX_DIM * MATRIX_DIM4 + 0] = acc3[0] + acc3[2]; in arm_mat_cmplx_mult_f32_4x4_mve() 363 pOut[3 * CMPLX_DIM * MATRIX_DIM4 + 1] = acc3[1] + acc3[3]; in arm_mat_cmplx_mult_f32_4x4_mve() 386 acc3 = vcmulq(vecA, vecB); in arm_mat_cmplx_mult_f32_4x4_mve() 387 acc3 = vcmlaq_rot90(acc3, vecA, vecB); in arm_mat_cmplx_mult_f32_4x4_mve() 404 acc3 = vcmlaq(acc3, vecA, vecB); in arm_mat_cmplx_mult_f32_4x4_mve() [all …]
|
D | arm_mat_mult_opt_q31.c | 229 q63_t acc0, acc1, acc2, acc3; in arm_mat_mult_opt_q31_4x4_mve() local 245 acc3 = vrmlaldavhq(vecA, vecB); in arm_mat_mult_opt_q31_4x4_mve() 250 acc3 = asrl(acc3, 23); in arm_mat_mult_opt_q31_4x4_mve() 255 pOut[3 * MATRIX_DIM4] = (q31_t) acc3; in arm_mat_mult_opt_q31_4x4_mve() 270 acc3 = vrmlaldavhq(vecA, vecB); in arm_mat_mult_opt_q31_4x4_mve() 275 acc3 = asrl(acc3, 23); in arm_mat_mult_opt_q31_4x4_mve() 280 pOut[3 * MATRIX_DIM4] = (q31_t) acc3; in arm_mat_mult_opt_q31_4x4_mve() 296 acc3 = vrmlaldavhq(vecA, vecB); in arm_mat_mult_opt_q31_4x4_mve() 301 acc3 = asrl(acc3, 23); in arm_mat_mult_opt_q31_4x4_mve() 306 pOut[3 * MATRIX_DIM4] = (q31_t) acc3; in arm_mat_mult_opt_q31_4x4_mve() [all …]
|
D | arm_mat_mult_q31.c | 225 q63_t acc0, acc1, acc2, acc3; in arm_mat_mult_q31_4x4_mve() local 241 acc3 = vrmlaldavhq(vecA, vecB); in arm_mat_mult_q31_4x4_mve() 246 acc3 = asrl(acc3, 23); in arm_mat_mult_q31_4x4_mve() 251 pOut[3 * MATRIX_DIM4] = (q31_t) acc3; in arm_mat_mult_q31_4x4_mve() 266 acc3 = vrmlaldavhq(vecA, vecB); in arm_mat_mult_q31_4x4_mve() 271 acc3 = asrl(acc3, 23); in arm_mat_mult_q31_4x4_mve() 276 pOut[3 * MATRIX_DIM4] = (q31_t) acc3; in arm_mat_mult_q31_4x4_mve() 292 acc3 = vrmlaldavhq(vecA, vecB); in arm_mat_mult_q31_4x4_mve() 297 acc3 = asrl(acc3, 23); in arm_mat_mult_q31_4x4_mve() 302 pOut[3 * MATRIX_DIM4] = (q31_t) acc3; in arm_mat_mult_q31_4x4_mve() [all …]
|
D | arm_mat_cmplx_mult_f16.c | 250 f16x8_t acc0, acc1, acc2, acc3; in arm_mat_cmplx_mult_f16_4x4_mve() local 277 acc3 = vcmulq(vecA, vecB); in arm_mat_cmplx_mult_f16_4x4_mve() 278 acc3 = vcmlaq_rot90(acc3, vecA, vecB); in arm_mat_cmplx_mult_f16_4x4_mve() 284 mve_cmplx_sum_intra_vec_f16(acc3, &pOut[3 * CMPLX_DIM * MATRIX_DIM]); in arm_mat_cmplx_mult_f16_4x4_mve() 306 acc3 = vcmulq(vecA, vecB); in arm_mat_cmplx_mult_f16_4x4_mve() 307 acc3 = vcmlaq_rot90(acc3, vecA, vecB); in arm_mat_cmplx_mult_f16_4x4_mve() 313 mve_cmplx_sum_intra_vec_f16(acc3, &pOut[3 * CMPLX_DIM * MATRIX_DIM]); in arm_mat_cmplx_mult_f16_4x4_mve() 335 acc3 = vcmulq(vecA, vecB); in arm_mat_cmplx_mult_f16_4x4_mve() 336 acc3 = vcmlaq_rot90(acc3, vecA, vecB); in arm_mat_cmplx_mult_f16_4x4_mve() 342 mve_cmplx_sum_intra_vec_f16(acc3, &pOut[3 * CMPLX_DIM * MATRIX_DIM]); in arm_mat_cmplx_mult_f16_4x4_mve() [all …]
|
D | arm_mat_mult_q15.c | 221 q63_t acc0, acc1, acc2, acc3; in arm_mat_mult_q15_4x4_mve() local 239 acc3 = vmlaldavq(vecA3, vecB); in arm_mat_mult_q15_4x4_mve() 244 acc3 = asrl(acc3, 15); in arm_mat_mult_q15_4x4_mve() 249 pOut[3 * MATRIX_DIM4] = (q15_t) __SSAT(acc3, 16); in arm_mat_mult_q15_4x4_mve() 260 acc3 = vmlaldavq(vecA3, vecB); in arm_mat_mult_q15_4x4_mve() 265 acc3 = asrl(acc3, 15); in arm_mat_mult_q15_4x4_mve() 270 pOut[3 * MATRIX_DIM4] = (q15_t) __SSAT(acc3, 16); in arm_mat_mult_q15_4x4_mve() 282 acc3 = vmlaldavq(vecA3, vecB); in arm_mat_mult_q15_4x4_mve() 287 acc3 = asrl(acc3, 15); in arm_mat_mult_q15_4x4_mve() 292 pOut[3 * MATRIX_DIM4] = (q15_t) __SSAT(acc3, 16); in arm_mat_mult_q15_4x4_mve() [all …]
|
/cmsis-dsp-latest/Source/FilteringFunctions/ |
D | arm_conv_fast_q15.c | 72 q31_t sum, acc0, acc1, acc2, acc3; /* Accumulators */ in arm_conv_fast_q15() local 281 acc3 = 0; in arm_conv_fast_q15() 316 acc3 = __SMLADX(x3, c0, acc3); in arm_conv_fast_q15() 338 acc3 = __SMLADX(x1, c0, acc3); in arm_conv_fast_q15() 368 acc3 = __SMLADX(x3, c0, acc3); in arm_conv_fast_q15() 387 acc3 = __SMLADX(x2, c0, acc3); in arm_conv_fast_q15() 405 acc3 = __SMLADX(x2, c0, acc3); in arm_conv_fast_q15() 423 acc3 = __SMLADX(x3, c0, acc3); in arm_conv_fast_q15() 429 write_q15x2_ia (&pOut, __PKHBT((acc2 >> 15), (acc3 >> 15), 16)); in arm_conv_fast_q15() 432 write_q15x2_ia (&pOut, __PKHBT((acc3 >> 15), (acc2 >> 15), 16)); in arm_conv_fast_q15()
|
D | arm_correlate_fast_q15.c | 72 q31_t sum, acc0, acc1, acc2, acc3; /* Accumulators */ in arm_correlate_fast_q15() local 267 acc3 = 0; in arm_correlate_fast_q15() 302 acc3 = __SMLAD(x3, c0, acc3); in arm_correlate_fast_q15() 324 acc3 = __SMLAD(x1, c0, acc3); in arm_correlate_fast_q15() 354 acc3 = __SMLADX(x3, c0, acc3); in arm_correlate_fast_q15() 373 acc3 = __SMLAD(x2, c0, acc3); in arm_correlate_fast_q15() 391 acc3 = __SMLAD(x2, c0, acc3); in arm_correlate_fast_q15() 409 acc3 = __SMLADX(x3, c0, acc3); in arm_correlate_fast_q15() 423 *pOut = (q15_t) (acc3 >> 15); in arm_correlate_fast_q15()
|
D | arm_fir_decimate_fast_q31.c | 80 q63_t acc1, acc2, acc3; in arm_fir_decimate_fast_q31() local 108 acc3 = 0; in arm_fir_decimate_fast_q31() 140 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x3 * c0)) >> 32); in arm_fir_decimate_fast_q31() 155 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x3 * c0)) >> 32); in arm_fir_decimate_fast_q31() 170 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x3 * c0)) >> 32); in arm_fir_decimate_fast_q31() 185 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x3 * c0)) >> 32); in arm_fir_decimate_fast_q31() 209 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x3 * c0)) >> 32); in arm_fir_decimate_fast_q31() 223 *pDst++ = (q31_t) (acc3 << 1); in arm_fir_decimate_fast_q31()
|
D | arm_fir_fast_q15.c | 75 q31_t acc1, acc2, acc3; /* Accumulators */ in arm_fir_fast_q15() local 107 acc3 = 0; in arm_fir_fast_q15() 157 acc3 = __SMLADX(x1, c0, acc3); in arm_fir_fast_q15() 182 acc3 = __SMLADX(x1, c0, acc3); in arm_fir_fast_q15() 220 acc3 = __SMLADX(x1, c0, acc3); in arm_fir_fast_q15() 227 write_q15x2_ia (&pDst, __PKHBT(__SSAT((acc2 >> 15), 16), __SSAT((acc3 >> 15), 16), 16)); in arm_fir_fast_q15() 230 write_q15x2_ia (&pDst, __PKHBT(__SSAT((acc3 >> 15), 16), __SSAT((acc2 >> 15), 16), 16)); in arm_fir_fast_q15()
|
D | arm_conv_partial_fast_q15.c | 68 q31_t sum, acc0, acc1, acc2, acc3; /* Accumulator */ in arm_conv_partial_fast_q15() local 304 acc3 = 0; in arm_conv_partial_fast_q15() 341 acc3 = __SMLADX(x3, c0, acc3); in arm_conv_partial_fast_q15() 363 acc3 = __SMLADX(x1, c0, acc3); in arm_conv_partial_fast_q15() 392 acc3 = __SMLADX(x3, c0, acc3); in arm_conv_partial_fast_q15() 411 acc3 = __SMLADX(x2, c0, acc3); in arm_conv_partial_fast_q15() 429 acc3 = __SMLADX(x2, c0, acc3); in arm_conv_partial_fast_q15() 446 acc3 = __SMLADX(x3, c0, acc3); in arm_conv_partial_fast_q15() 452 write_q15x2_ia (&pOut, __PKHBT(acc2 >> 15, acc3 >> 15, 16)); in arm_conv_partial_fast_q15() 455 write_q15x2_ia (&pOut, __PKHBT(acc3 >> 15, acc2 >> 15, 16)); in arm_conv_partial_fast_q15()
|
D | arm_conv_fast_q31.c | 75 q31_t sum, acc0, acc1, acc2, acc3; /* Accumulators */ in arm_conv_fast_q31() local 244 acc3 = 0; in arm_conv_fast_q31() 271 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x3 * c0)) >> 32); in arm_conv_fast_q31() 287 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x0 * c0)) >> 32); in arm_conv_fast_q31() 303 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x1 * c0)) >> 32); in arm_conv_fast_q31() 319 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x2 * c0)) >> 32); in arm_conv_fast_q31() 343 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x3 * c0)) >> 32); in arm_conv_fast_q31() 358 *pOut++ = (q31_t) (acc3 << 1); in arm_conv_fast_q31()
|
D | arm_correlate_fast_q31.c | 77 q31_t sum, acc0, acc1, acc2, acc3; /* Accumulators */ in arm_correlate_fast_q31() local 270 acc3 = 0; in arm_correlate_fast_q31() 297 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x3 * c0)) >> 32); in arm_correlate_fast_q31() 313 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x0 * c0)) >> 32); in arm_correlate_fast_q31() 329 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x1 * c0)) >> 32); in arm_correlate_fast_q31() 345 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x2 * c0)) >> 32); in arm_correlate_fast_q31() 369 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x3 * c0)) >> 32); in arm_correlate_fast_q31() 391 *pOut = (q31_t) (acc3 << 1); in arm_correlate_fast_q31()
|
D | arm_fir_q31.c | 234 q63_t acc0=0, acc1=0, acc2=0, acc3=0; in arm_fir_q31_1_4_mve() local 262 acc3 = vrmlaldavhq(vecIn0, vecCoeffs); in arm_fir_q31_1_4_mve() 267 acc3 = asrl(acc3, 23); in arm_fir_q31_1_4_mve() 272 *pOutput++ = (q31_t) acc3; in arm_fir_q31_1_4_mve() 464 q63_t acc0, acc1, acc2, acc3; in arm_fir_q31_29_32_mve() local 560 acc3 = vrmlaldavhq(vecIn0, vecCoeffs0); in arm_fir_q31_29_32_mve() 562 acc3 = vrmlaldavhaq(acc3, vecIn0, vecCoeffs1); in arm_fir_q31_29_32_mve() 564 acc3 = vrmlaldavhaq(acc3, vecIn0, vecCoeffs2); in arm_fir_q31_29_32_mve() 566 acc3 = vrmlaldavhaq(acc3, vecIn0, vecCoeffs3); in arm_fir_q31_29_32_mve() 568 acc3 = vrmlaldavhaq(acc3, vecIn0, vecCoeffs4); in arm_fir_q31_29_32_mve() [all …]
|
D | arm_conv_fast_opt_q15.c | 85 q31_t acc1, acc2, acc3; /* Accumulators */ in arm_conv_fast_opt_q15() local 205 acc3 = 0; in arm_conv_fast_opt_q15() 250 acc3 = __SMLADX(x3, y1, acc3); in arm_conv_fast_opt_q15() 261 acc3 = __SMLADX(x3, y2, acc3); in arm_conv_fast_opt_q15() 279 acc3 += (*pScr1++ * *pIn2++); in arm_conv_fast_opt_q15() 292 write_q15x2_ia (&pOut, __PKHBT(__SSAT((acc2 >> 15), 16), __SSAT((acc3 >> 15), 16), 16)); in arm_conv_fast_opt_q15() 295 write_q15x2_ia (&pOut, __PKHBT(__SSAT((acc3 >> 15), 16), __SSAT((acc2 >> 15), 16), 16)); in arm_conv_fast_opt_q15()
|
D | arm_conv_opt_q15.c | 81 q63_t acc1, acc2, acc3; /* Accumulators */ in arm_conv_opt_q15() local 201 acc3 = 0; in arm_conv_opt_q15() 246 acc3 = __SMLALDX(x3, y1, acc3); in arm_conv_opt_q15() 257 acc3 = __SMLALDX(x3, y2, acc3); in arm_conv_opt_q15() 275 acc3 += (*pScr1++ * *pIn2++); in arm_conv_opt_q15() 288 write_q15x2_ia (&pOut, __PKHBT(__SSAT((acc2 >> 15), 16), __SSAT((acc3 >> 15), 16), 16)); in arm_conv_opt_q15() 291 write_q15x2_ia (&pOut, __PKHBT(__SSAT((acc3 >> 15), 16), __SSAT((acc2 >> 15), 16), 16)); in arm_conv_opt_q15()
|
D | arm_conv_partial_fast_q31.c | 78 q31_t acc0, acc1, acc2, acc3; /* Accumulators */ in arm_conv_partial_fast_q31() local 280 acc3 = 0; in arm_conv_partial_fast_q31() 307 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x3 * c0)) >> 32); in arm_conv_partial_fast_q31() 322 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x0 * c0)) >> 32); in arm_conv_partial_fast_q31() 337 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x1 * c0)) >> 32); in arm_conv_partial_fast_q31() 352 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x2 * c0)) >> 32); in arm_conv_partial_fast_q31() 375 acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x3 * c0)) >> 32); in arm_conv_partial_fast_q31() 390 *pOut++ = (q31_t) (acc3 << 1); in arm_conv_partial_fast_q31()
|
D | arm_conv_partial_q15.c | 74 q63_t sum, acc0, acc1, acc2, acc3; /* Accumulator */ in arm_conv_partial_q15() local 310 acc3 = 0; in arm_conv_partial_q15() 347 acc3 = __SMLALDX(x3, c0, acc3); in arm_conv_partial_q15() 369 acc3 = __SMLALDX(x1, c0, acc3); in arm_conv_partial_q15() 398 acc3 = __SMLALDX(x3, c0, acc3); in arm_conv_partial_q15() 417 acc3 = __SMLALDX(x2, c0, acc3); in arm_conv_partial_q15() 435 acc3 = __SMLALDX(x2, c0, acc3); in arm_conv_partial_q15() 452 acc3 = __SMLALDX(x3, c0, acc3); in arm_conv_partial_q15() 460 int32_t sat3 = __SSAT((acc3 >> 15), 16); in arm_conv_partial_q15()
|
D | arm_conv_q15.c | 150 int64_t acc3 = 0LL; in arm_conv_q15() local 159 MVE_INTR_CONV_QUAD_INC_X_FIXED_SIZE_Q15(acc0, acc1, acc2, acc3, pX, pY, count); in arm_conv_q15() 163 *pDst++ = (q15_t) acc3; in arm_conv_q15() 242 q63_t sum, acc0, acc1, acc2, acc3; /* Accumulators */ in arm_conv_q15() local 454 acc3 = 0; in arm_conv_q15() 490 acc3 = __SMLALDX(x3, c0, acc3); in arm_conv_q15() 513 acc3 = __SMLALDX(x1, c0, acc3); in arm_conv_q15() 542 acc3 = __SMLALDX(x3, c0, acc3); in arm_conv_q15() 561 acc3 = __SMLALDX(x2, c0, acc3); in arm_conv_q15() 579 acc3 = __SMLALDX(x2, c0, acc3); in arm_conv_q15() [all …]
|
D | arm_correlate_q15.c | 172 int64_t acc3 = 0LL; in arm_correlate_q15() local 181 MVE_INTR_CORR_QUAD_INC_X_FIXED_SIZE_Q15(acc0, acc1, acc2, acc3, pX, pY, srcBLen); in arm_correlate_q15() 189 *pDst = (q15_t) acc3; in arm_correlate_q15() 280 q63_t sum, acc0, acc1, acc2, acc3; /* Accumulators */ in arm_correlate_q15() local 473 acc3 = 0; in arm_correlate_q15() 509 acc3 = __SMLALD(x3, c0, acc3); in arm_correlate_q15() 531 acc3 = __SMLALD(x1, c0, acc3); in arm_correlate_q15() 557 acc3 = __SMLALDX(x3, c0, acc3); in arm_correlate_q15() 576 acc3 = __SMLALD(x2, c0, acc3); in arm_correlate_q15() 594 acc3 = __SMLALD(x2, c0, acc3); in arm_correlate_q15() [all …]
|
D | arm_fir_fast_q31.c | 76 q31_t acc1, acc2, acc3; /* Accumulators */ in arm_fir_fast_q31() local 108 acc3 = 0; in arm_fir_fast_q31() 145 multAcc_32x32_keep32_R(acc3, x3, c0); in arm_fir_fast_q31() 157 multAcc_32x32_keep32_R(acc3, x0, c0); in arm_fir_fast_q31() 169 multAcc_32x32_keep32_R(acc3, x1, c0); in arm_fir_fast_q31() 181 multAcc_32x32_keep32_R(acc3, x2, c0); in arm_fir_fast_q31() 206 multAcc_32x32_keep32_R(acc3, x3, c0); in arm_fir_fast_q31() 222 *pDst++ = (q31_t) (acc3 << 1); in arm_fir_fast_q31()
|
D | arm_conv_partial_opt_q15.c | 86 q63_t acc1, acc2, acc3; /* Accumulator */ in arm_conv_partial_opt_q15() local 217 acc3 = 0; in arm_conv_partial_opt_q15() 262 acc3 = __SMLALDX(x3, y1, acc3); in arm_conv_partial_opt_q15() 273 acc3 = __SMLALDX(x3, y2, acc3); in arm_conv_partial_opt_q15() 291 acc3 += (*pScr1++ * *pIn2++); in arm_conv_partial_opt_q15() 304 write_q15x2_ia (&pOut, __PKHBT(__SSAT((acc2 >> 15), 16), __SSAT((acc3 >> 15), 16), 16)); in arm_conv_partial_opt_q15() 307 write_q15x2_ia (&pOut, __PKHBT(__SSAT((acc3 >> 15), 16), __SSAT((acc2 >> 15), 16), 16)); in arm_conv_partial_opt_q15()
|
D | arm_conv_partial_fast_opt_q15.c | 85 q31_t acc1, acc2, acc3; /* Accumulator */ in arm_conv_partial_fast_opt_q15() local 216 acc3 = 0; in arm_conv_partial_fast_opt_q15() 261 acc3 = __SMLADX(x3, y1, acc3); in arm_conv_partial_fast_opt_q15() 273 acc3 = __SMLADX(x3, y2, acc3); in arm_conv_partial_fast_opt_q15() 291 acc3 += (*pScr1++ * *pIn2++); in arm_conv_partial_fast_opt_q15() 304 write_q15x2_ia (&pOut, __PKHBT(__SSAT((acc2 >> 15), 16), __SSAT((acc3 >> 15), 16), 16)); in arm_conv_partial_fast_opt_q15() 307 write_q15x2_ia (&pOut, __PKHBT(__SSAT((acc3 >> 15), 16), __SSAT((acc2 >> 15), 16), 16)); in arm_conv_partial_fast_opt_q15()
|
D | arm_correlate_opt_q15.c | 79 q63_t acc1, acc2, acc3; /* Accumulators */ in arm_correlate_opt_q15() local 178 acc3 = 0; in arm_correlate_opt_q15() 222 acc3 = __SMLALDX(x3, y1, acc3); in arm_correlate_opt_q15() 233 acc3 = __SMLALDX(x3, y2, acc3); in arm_correlate_opt_q15() 251 acc3 += (*pScr1++ * *pIn2++); in arm_correlate_opt_q15() 269 *pOut = (__SSAT(acc3 >> 15U, 16)); in arm_correlate_opt_q15()
|
D | arm_correlate_fast_opt_q15.c | 82 q31_t acc1, acc2, acc3; /* Accumulators */ in arm_correlate_fast_opt_q15() local 183 acc3 = 0; in arm_correlate_fast_opt_q15() 227 acc3 = __SMLADX(x3, y1, acc3); in arm_correlate_fast_opt_q15() 238 acc3 = __SMLADX(x3, y2, acc3); in arm_correlate_fast_opt_q15() 256 acc3 += (*pScr1++ * *pIn2++); in arm_correlate_fast_opt_q15() 273 *pOut = (__SSAT(acc3 >> 15U, 16)); in arm_correlate_fast_opt_q15()
|