| /hal_nxp-3.5.0/mcux/mcux-sdk/CMSIS/DSP/Source/MatrixFunctions/ |
| D | arm_mat_cmplx_mult_f32.c | 147 f32x4_t acc0, acc1, acc2; in arm_mat_cmplx_mult_f32_3x3_mve() local
175 acc2 = vcmulq(vecA, vecB); in arm_mat_cmplx_mult_f32_3x3_mve()
176 acc2 = vcmlaq_rot90(acc2, vecA, vecB); in arm_mat_cmplx_mult_f32_3x3_mve()
190 acc2 = vcmlaq(acc2, vecA, vecB); in arm_mat_cmplx_mult_f32_3x3_mve()
191 acc2 = vcmlaq_rot90(acc2, vecA, vecB); in arm_mat_cmplx_mult_f32_3x3_mve()
198 pOut[2 * CMPLX_DIM * MATRIX_DIM3 + 0] = acc2[0] + acc2[2]; in arm_mat_cmplx_mult_f32_3x3_mve()
199 pOut[2 * CMPLX_DIM * MATRIX_DIM3 + 1] = acc2[1] + acc2[3]; in arm_mat_cmplx_mult_f32_3x3_mve()
218 acc2 = vcmulq(vecA, vecB); in arm_mat_cmplx_mult_f32_3x3_mve()
219 acc2 = vcmlaq_rot90(acc2, vecA, vecB); in arm_mat_cmplx_mult_f32_3x3_mve()
232 acc2 = vcmlaq(acc2, vecA, vecB); in arm_mat_cmplx_mult_f32_3x3_mve()
[all …]
|
| D | arm_mat_mult_q31.c | 138 q63_t acc0, acc1, acc2; in arm_mat_mult_q31_3x3_mve() local
155 acc2 = vrmlaldavhq(vecA, vecB); in arm_mat_mult_q31_3x3_mve()
159 acc2 = asrl(acc2, 23); in arm_mat_mult_q31_3x3_mve()
163 pOut[2 * MATRIX_DIM3] = (q31_t) acc2; in arm_mat_mult_q31_3x3_mve()
176 acc2 = vrmlaldavhq(vecA, vecB); in arm_mat_mult_q31_3x3_mve()
180 acc2 = asrl(acc2, 23); in arm_mat_mult_q31_3x3_mve()
184 pOut[2 * MATRIX_DIM3] = (q31_t) acc2; in arm_mat_mult_q31_3x3_mve()
197 acc2 = vrmlaldavhq(vecA, vecB); in arm_mat_mult_q31_3x3_mve()
201 acc2 = asrl(acc2, 23); in arm_mat_mult_q31_3x3_mve()
205 pOut[2 * MATRIX_DIM3] = (q31_t) acc2; in arm_mat_mult_q31_3x3_mve()
[all …]
|
| D | arm_mat_cmplx_mult_q31.c | 78 q63_t acc0, acc1, acc2, acc3; in arm_mat_cmplx_mult_q31_2x2_mve() local
96 acc2 = vmlsldavq_s32(vecA, vecB); in arm_mat_cmplx_mult_q31_2x2_mve()
101 pOut[1 * CMPLX_DIM * MATRIX_DIM2 + 0] = (q31_t) asrl(acc2, 31); in arm_mat_cmplx_mult_q31_2x2_mve()
114 acc2 = vmlsldavq_s32(vecA, vecB); in arm_mat_cmplx_mult_q31_2x2_mve()
121 pOut[1 * CMPLX_DIM * MATRIX_DIM2 + 0] = (q31_t) asrl(acc2, 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
170 acc2 = vmlsldavq_s32(vecA, vecB); in arm_mat_cmplx_mult_q31_3x3_mve()
178 acc2 = vmlsldavaq_s32(acc2, vecA, vecB1); in arm_mat_cmplx_mult_q31_3x3_mve()
183 pOut[1 * CMPLX_DIM * MATRIX_DIM3 + 0] = (q31_t) asrl(acc2, 31); in arm_mat_cmplx_mult_q31_3x3_mve()
211 acc2 = vmlsldavq_s32(vecA, vecB); in arm_mat_cmplx_mult_q31_3x3_mve()
[all …]
|
| D | arm_mat_cmplx_mult_f16.c | 151 f16x8_t acc0, acc1, acc2; in arm_mat_cmplx_mult_f16_3x3_mve() local
179 acc2 = vcmulq(vecA2, vecB); in arm_mat_cmplx_mult_f16_3x3_mve()
180 acc2 = vcmlaq_rot90(acc2, vecA2, vecB); in arm_mat_cmplx_mult_f16_3x3_mve()
184 mve_cmplx_sum_intra_vec_f16(acc2, &pOut[2 * CMPLX_DIM * MATRIX_DIM]); in arm_mat_cmplx_mult_f16_3x3_mve()
199 acc2 = vcmulq(vecA2, vecB); in arm_mat_cmplx_mult_f16_3x3_mve()
200 acc2 = vcmlaq_rot90(acc2, vecA2, vecB); in arm_mat_cmplx_mult_f16_3x3_mve()
204 mve_cmplx_sum_intra_vec_f16(acc2, &pOut[2 * CMPLX_DIM * MATRIX_DIM]); in arm_mat_cmplx_mult_f16_3x3_mve()
219 acc2 = vcmulq(vecA2, vecB); in arm_mat_cmplx_mult_f16_3x3_mve()
220 acc2 = vcmlaq_rot90(acc2, vecA2, vecB); in arm_mat_cmplx_mult_f16_3x3_mve()
224 mve_cmplx_sum_intra_vec_f16(acc2, &pOut[2 * CMPLX_DIM * MATRIX_DIM]); in arm_mat_cmplx_mult_f16_3x3_mve()
[all …]
|
| D | arm_mat_mult_q15.c | 136 q63_t acc0, acc1, acc2; in arm_mat_mult_q15_3x3_mve() local
153 acc2 = vmlaldavq(vecA2, vecB); in arm_mat_mult_q15_3x3_mve()
157 acc2 = asrl(acc2, 15); in arm_mat_mult_q15_3x3_mve()
161 pOut[2 * MATRIX_DIM3] = (q15_t) __SSAT(acc2, 16); in arm_mat_mult_q15_3x3_mve()
171 acc2 = vmlaldavq(vecA2, vecB); in arm_mat_mult_q15_3x3_mve()
175 acc2 = asrl(acc2, 15); in arm_mat_mult_q15_3x3_mve()
179 pOut[2 * MATRIX_DIM3] = (q15_t) __SSAT(acc2, 16); in arm_mat_mult_q15_3x3_mve()
189 acc2 = vmlaldavq(vecA2, vecB); in arm_mat_mult_q15_3x3_mve()
193 acc2 = asrl(acc2, 15); in arm_mat_mult_q15_3x3_mve()
197 pOut[2 * MATRIX_DIM3] = (q15_t) __SSAT(acc2, 16); in arm_mat_mult_q15_3x3_mve()
[all …]
|
| D | arm_mat_mult_q7.c | 121 q31_t acc0, acc1, acc2; in arm_mat_mult_q7_3x3_mve() local
138 acc2 = vmladavq_s8(vecA2, vecB); in arm_mat_mult_q7_3x3_mve()
142 pOut[2 * MATRIX_DIM] = (q7_t) __SSAT(acc2 >> 7, 8); in arm_mat_mult_q7_3x3_mve()
152 acc2 = vmladavq_s8(vecA2, vecB); in arm_mat_mult_q7_3x3_mve()
156 pOut[2 * MATRIX_DIM] = (q7_t) __SSAT(acc2 >> 7, 8); in arm_mat_mult_q7_3x3_mve()
166 acc2 = vmladavq_s8(vecA2, vecB); in arm_mat_mult_q7_3x3_mve()
170 pOut[2 * MATRIX_DIM] = (q7_t) __SSAT(acc2 >> 7, 8); in arm_mat_mult_q7_3x3_mve()
192 q31_t acc0, acc1, acc2, acc3; in arm_mat_mult_q7_4x4_mve() local
209 acc2 = vmladavq_s8(vecA2, vecB); in arm_mat_mult_q7_4x4_mve()
214 pOut[2 * MATRIX_DIM] = (q7_t) __SSAT(acc2 >> 7, 8); in arm_mat_mult_q7_4x4_mve()
[all …]
|
| /hal_nxp-3.5.0/mcux/mcux-sdk/CMSIS/DSP/Source/FilteringFunctions/ |
| D | arm_fir_q31.c | 235 q63_t acc0=0, acc1=0, acc2=0, acc3=0; in arm_fir_q31_1_4_mve() local
260 acc2 = vrmlaldavhq(vecIn0, vecCoeffs); in arm_fir_q31_1_4_mve()
267 acc2 = asrl(acc2, 23); in arm_fir_q31_1_4_mve()
272 *pOutput++ = (q31_t) acc2; in arm_fir_q31_1_4_mve()
301 acc2 = vrmlaldavhq(vecIn0, vecCoeffs); in arm_fir_q31_1_4_mve()
305 acc2 = asrl(acc2, 23); in arm_fir_q31_1_4_mve()
309 *pOutput++ = (q31_t) acc2; in arm_fir_q31_1_4_mve()
465 q63_t acc0, acc1, acc2, acc3; in arm_fir_q31_29_32_mve() local
543 acc2 = vrmlaldavhq(vecIn0, vecCoeffs0); in arm_fir_q31_29_32_mve()
545 acc2 = vrmlaldavhaq(acc2, vecIn0, vecCoeffs1); in arm_fir_q31_29_32_mve()
[all …]
|
| D | arm_fir_interpolate_q15.c | 120 q63_t acc2 = 0LL; in arm_fir_interpolate_q15() local
135 acc2 = vmlaldavaq(acc2, vecState, vecCoef); in arm_fir_interpolate_q15()
158 acc2 = vmlaldavaq(acc2, vecState, vecCoef); in arm_fir_interpolate_q15()
166 acc2 = asrl(acc2, 15); in arm_fir_interpolate_q15()
171 *pDst++ = (q15_t) __SSAT(acc2, 16); in arm_fir_interpolate_q15()
184 q63_t acc2 = 0LL; in arm_fir_interpolate_q15() local
198 acc2 = vmlaldavaq(acc2, vecState, vecCoef); in arm_fir_interpolate_q15()
218 acc2 = vmlaldavaq(acc2, vecState, vecCoef); in arm_fir_interpolate_q15()
223 acc2 = asrl(acc2, 15); in arm_fir_interpolate_q15()
227 *pDst++ = (q15_t) __SSAT(acc2, 16);; in arm_fir_interpolate_q15()
[all …]
|
| D | arm_fir_interpolate_q31.c | 117 q63_t acc2 = 0LL; in arm_fir_interpolate_q31() local
132 acc2 = vrmlaldavhaq(acc2, vecState, vecCoef); in arm_fir_interpolate_q31()
155 acc2 = vrmlaldavhaq(acc2, vecState, vecCoef); in arm_fir_interpolate_q31()
163 acc2 = asrl(acc2, 31 - 8); in arm_fir_interpolate_q31()
168 *pDst++ = (q31_t) acc2; in arm_fir_interpolate_q31()
181 q63_t acc2 = 0LL; in arm_fir_interpolate_q31() local
195 acc2 = vrmlaldavhaq(acc2, vecState, vecCoef); in arm_fir_interpolate_q31()
215 acc2 = vrmlaldavhaq(acc2, vecState, vecCoef); in arm_fir_interpolate_q31()
220 acc2 = asrl(acc2, 31 - 8); in arm_fir_interpolate_q31()
224 *pDst++ = (q31_t) acc2; in arm_fir_interpolate_q31()
[all …]
|
| D | arm_conv_fast_q15.c | 73 q31_t sum, acc0, acc1, acc2, acc3; /* Accumulators */ in arm_conv_fast_q15() local
281 acc2 = 0; in arm_conv_fast_q15()
314 acc2 = __SMLADX(x2, c0, acc2); in arm_conv_fast_q15()
336 acc2 = __SMLADX(x0, c0, acc2); in arm_conv_fast_q15()
368 acc2 = __SMLADX(x1, c0, acc2); in arm_conv_fast_q15()
387 acc2 = __SMLADX(x3, c0, acc2); in arm_conv_fast_q15()
405 acc2 = __SMLADX(x3, c0, acc2); in arm_conv_fast_q15()
423 acc2 = __SMLADX(x2, c0, acc2); in arm_conv_fast_q15()
430 write_q15x2_ia (&pOut, __PKHBT((acc2 >> 15), (acc3 >> 15), 16)); in arm_conv_fast_q15()
433 write_q15x2_ia (&pOut, __PKHBT((acc3 >> 15), (acc2 >> 15), 16)); in arm_conv_fast_q15()
|
| D | arm_correlate_fast_q15.c | 73 q31_t sum, acc0, acc1, acc2, acc3; /* Accumulators */ in arm_correlate_fast_q15() local
267 acc2 = 0; in arm_correlate_fast_q15()
300 acc2 = __SMLAD(x2, c0, acc2); in arm_correlate_fast_q15()
322 acc2 = __SMLAD(x0, c0, acc2); in arm_correlate_fast_q15()
354 acc2 = __SMLADX(x1, c0, acc2); in arm_correlate_fast_q15()
373 acc2 = __SMLAD(x3, c0, acc2); in arm_correlate_fast_q15()
391 acc2 = __SMLAD(x3, c0, acc2); in arm_correlate_fast_q15()
409 acc2 = __SMLADX(x2, c0, acc2); in arm_correlate_fast_q15()
421 *pOut = (q15_t) (acc2 >> 15); in arm_correlate_fast_q15()
|
| D | arm_fir_decimate_fast_q31.c | 81 q63_t acc1, acc2, acc3; in arm_fir_decimate_fast_q31() local
108 acc2 = 0; in arm_fir_decimate_fast_q31()
140 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x2 * c0)) >> 32); in arm_fir_decimate_fast_q31()
155 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x2 * c0)) >> 32); in arm_fir_decimate_fast_q31()
170 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x2 * c0)) >> 32); in arm_fir_decimate_fast_q31()
185 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x2 * c0)) >> 32); in arm_fir_decimate_fast_q31()
209 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x2 * c0)) >> 32); in arm_fir_decimate_fast_q31()
223 *pDst++ = (q31_t) (acc2 << 1); in arm_fir_decimate_fast_q31()
|
| D | arm_fir_q15.c | 216 q63_t acc0, acc1, acc2, acc3; in arm_fir_q15() local
253 acc2 = 0LL; in arm_fir_q15()
278 acc2 = vmlaldavaq(acc2, vecIn0, vecCoeffs); in arm_fir_q15()
293 *pOutput++ = (q15_t) MVE_ASRL_SAT16(acc2, 15); in arm_fir_q15()
313 acc2 = 0LL; in arm_fir_q15()
337 acc2 = vmlaldavaq(acc2, vecIn0, vecCoeffs); in arm_fir_q15()
349 acc2 = asrl(acc2, 15); in arm_fir_q15()
353 *pOutput++ = (q15_t) MVE_ASRL_SAT16(acc2, 15); in arm_fir_q15()
475 q63_t acc1, acc2, acc3; /* Accumulators */ in arm_fir_q15() local
506 acc2 = 0; in arm_fir_q15()
[all …]
|
| D | arm_conv_partial_fast_q15.c | 68 q31_t sum, acc0, acc1, acc2, acc3; /* Accumulator */ in arm_conv_partial_fast_q15() local
303 acc2 = 0; in arm_conv_partial_fast_q15()
338 acc2 = __SMLADX(x2, c0, acc2); in arm_conv_partial_fast_q15()
360 acc2 = __SMLADX(x0, c0, acc2); in arm_conv_partial_fast_q15()
391 acc2 = __SMLADX(x1, c0, acc2); in arm_conv_partial_fast_q15()
410 acc2 = __SMLADX(x3, c0, acc2); in arm_conv_partial_fast_q15()
428 acc2 = __SMLADX(x3, c0, acc2); in arm_conv_partial_fast_q15()
445 acc2 = __SMLADX(x2, c0, acc2); 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_fir_q7.c | 207 q31_t acc0, acc1, acc2, acc3; in arm_fir_q7() local
263 acc2 = 0; in arm_fir_q7()
290 acc2 = vmladavaq(acc2, vecIn0, vecCoeffs); in arm_fir_q7()
307 *pOutput++ = (q7_t) __SSAT((acc2 >> 7U), 8); in arm_fir_q7()
327 acc2 = 0; in arm_fir_q7()
347 acc2 = vmladavaq(acc2, vecIn0, vecCoeffs); in arm_fir_q7()
356 *pOutput++ = (q7_t) __SSAT((acc2 >> 7U), 8); in arm_fir_q7()
463 q31_t acc1, acc2, acc3; /* Accumulators */ in arm_fir_q7() local
494 acc2 = 0; in arm_fir_q7()
529 acc2 += ((q15_t) x2 * c0); in arm_fir_q7()
[all …]
|
| D | arm_fir_fast_q15.c | 76 q31_t acc1, acc2, acc3; /* Accumulators */ in arm_fir_fast_q15() local
107 acc2 = 0; in arm_fir_fast_q15()
135 acc2 = __SMLAD(x2, c0, acc2); in arm_fir_fast_q15()
170 acc2 = __SMLAD(x0, c0, acc2); in arm_fir_fast_q15()
198 acc2 = __SMLAD(x2, c0, acc2); in arm_fir_fast_q15()
228 write_q15x2_ia (&pDst, __PKHBT(__SSAT((acc2 >> 15), 16), __SSAT((acc3 >> 15), 16), 16)); in arm_fir_fast_q15()
231 write_q15x2_ia (&pDst, __PKHBT(__SSAT((acc3 >> 15), 16), __SSAT((acc2 >> 15), 16), 16)); in arm_fir_fast_q15()
|
| D | arm_conv_fast_q31.c | 76 q31_t sum, acc0, acc1, acc2, acc3; /* Accumulators */ in arm_conv_fast_q31() local
244 acc2 = 0; in arm_conv_fast_q31()
270 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x2 * c0)) >> 32); in arm_conv_fast_q31()
286 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x3 * c0)) >> 32); in arm_conv_fast_q31()
302 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x0 * c0)) >> 32); in arm_conv_fast_q31()
318 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x1 * c0)) >> 32); in arm_conv_fast_q31()
342 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x2 * c0)) >> 32); in arm_conv_fast_q31()
358 *pOut++ = (q31_t) (acc2 << 1); in arm_conv_fast_q31()
|
| D | arm_conv_partial_q15.c | 74 q63_t sum, acc0, acc1, acc2, acc3; /* Accumulator */ in arm_conv_partial_q15() local
309 acc2 = 0; in arm_conv_partial_q15()
344 acc2 = __SMLALDX(x2, c0, acc2); in arm_conv_partial_q15()
366 acc2 = __SMLALDX(x0, c0, acc2); in arm_conv_partial_q15()
397 acc2 = __SMLALDX(x1, c0, acc2); in arm_conv_partial_q15()
416 acc2 = __SMLALDX(x3, c0, acc2); in arm_conv_partial_q15()
434 acc2 = __SMLALDX(x3, c0, acc2); in arm_conv_partial_q15()
451 acc2 = __SMLALDX(x2, c0, acc2); in arm_conv_partial_q15()
458 write_q15x2_ia (&pOut, __PKHBT(__SSAT((acc2 >> 15), 16), __SSAT((acc3 >> 15), 16), 16)); in arm_conv_partial_q15()
461 write_q15x2_ia (&pOut, __PKHBT(__SSAT((acc3 >> 15), 16), __SSAT((acc2 >> 15), 16), 16)); in arm_conv_partial_q15()
|
| D | arm_conv_q15.c | 145 int64_t acc2 = 0LL; in arm_conv_q15() local
155 MVE_INTR_CONV_QUAD_INC_X_FIXED_SIZE_Q15(acc0, acc1, acc2, acc3, pX, pY, count); in arm_conv_q15()
158 *pDst++ = (q15_t) acc2; in arm_conv_q15()
236 q63_t sum, acc0, acc1, acc2, acc3; /* Accumulators */ in arm_conv_q15() local
442 acc2 = 0; in arm_conv_q15()
476 acc2 = __SMLALDX(x2, c0, acc2); in arm_conv_q15()
499 acc2 = __SMLALDX(x0, c0, acc2); in arm_conv_q15()
530 acc2 = __SMLALDX(x1, c0, acc2); in arm_conv_q15()
549 acc2 = __SMLALDX(x3, c0, acc2); in arm_conv_q15()
567 acc2 = __SMLALDX(x3, c0, acc2); in arm_conv_q15()
[all …]
|
| D | arm_correlate_fast_q31.c | 78 q31_t sum, acc0, acc1, acc2, acc3; /* Accumulators */ in arm_correlate_fast_q31() local
270 acc2 = 0; in arm_correlate_fast_q31()
296 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x2 * c0)) >> 32); in arm_correlate_fast_q31()
312 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x3 * c0)) >> 32); in arm_correlate_fast_q31()
328 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x0 * c0)) >> 32); in arm_correlate_fast_q31()
344 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x1 * c0)) >> 32); in arm_correlate_fast_q31()
368 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x2 * c0)) >> 32); in arm_correlate_fast_q31()
389 *pOut = (q31_t) (acc2 << 1); in arm_correlate_fast_q31()
|
| D | arm_conv_opt_q15.c | 82 q63_t acc1, acc2, acc3; /* Accumulators */ in arm_conv_opt_q15() local
201 acc2 = 0; in arm_conv_opt_q15()
221 acc2 = __SMLALD(x2, y1, acc2); in arm_conv_opt_q15()
238 acc2 = __SMLALD(x1, y2, acc2); in arm_conv_opt_q15()
275 acc2 += (*pScr1++ * *pIn2); in arm_conv_opt_q15()
289 write_q15x2_ia (&pOut, __PKHBT(__SSAT((acc2 >> 15), 16), __SSAT((acc3 >> 15), 16), 16)); in arm_conv_opt_q15()
292 write_q15x2_ia (&pOut, __PKHBT(__SSAT((acc3 >> 15), 16), __SSAT((acc2 >> 15), 16), 16)); in arm_conv_opt_q15()
|
| D | arm_conv_fast_opt_q15.c | 86 q31_t acc1, acc2, acc3; /* Accumulators */ in arm_conv_fast_opt_q15() local
205 acc2 = 0; in arm_conv_fast_opt_q15()
225 acc2 = __SMLAD(x2, y1, acc2); in arm_conv_fast_opt_q15()
242 acc2 = __SMLAD(x1, y2, acc2); in arm_conv_fast_opt_q15()
279 acc2 += (*pScr1++ * *pIn2); in arm_conv_fast_opt_q15()
293 write_q15x2_ia (&pOut, __PKHBT(__SSAT((acc2 >> 15), 16), __SSAT((acc3 >> 15), 16), 16)); in arm_conv_fast_opt_q15()
296 write_q15x2_ia (&pOut, __PKHBT(__SSAT((acc3 >> 15), 16), __SSAT((acc2 >> 15), 16), 16)); in arm_conv_fast_opt_q15()
|
| D | arm_conv_partial_fast_q31.c | 78 q31_t acc0, acc1, acc2, acc3; /* Accumulators */ in arm_conv_partial_fast_q31() local
279 acc2 = 0; in arm_conv_partial_fast_q31()
305 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x2 * c0)) >> 32); in arm_conv_partial_fast_q31()
320 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x3 * c0)) >> 32); in arm_conv_partial_fast_q31()
335 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x0 * c0)) >> 32); in arm_conv_partial_fast_q31()
350 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x1 * c0)) >> 32); in arm_conv_partial_fast_q31()
373 acc2 = (q31_t) ((((q63_t) acc2 << 32) + ((q63_t) x2 * c0)) >> 32); in arm_conv_partial_fast_q31()
389 *pOut++ = (q31_t) (acc2 << 1); in arm_conv_partial_fast_q31()
|
| D | arm_correlate_q15.c | 172 int64_t acc2 = 0LL; in arm_correlate_q15() local
182 MVE_INTR_CORR_QUAD_INC_X_FIXED_SIZE_Q15(acc0, acc1, acc2, acc3, pX, pY, srcBLen); in arm_correlate_q15()
188 *pDst = (q15_t) acc2; in arm_correlate_q15()
281 q63_t sum, acc0, acc1, acc2, acc3; /* Accumulators */ in arm_correlate_q15() local
473 acc2 = 0; in arm_correlate_q15()
507 acc2 = __SMLALD(x2, c0, acc2); in arm_correlate_q15()
529 acc2 = __SMLALD(x0, c0, acc2); in arm_correlate_q15()
557 acc2 = __SMLALDX(x1, c0, acc2); in arm_correlate_q15()
576 acc2 = __SMLALD(x3, c0, acc2); in arm_correlate_q15()
594 acc2 = __SMLALD(x3, c0, acc2); in arm_correlate_q15()
[all …]
|
| D | arm_conv_partial_fast_opt_q15.c | 85 q31_t acc1, acc2, acc3; /* Accumulator */ in arm_conv_partial_fast_opt_q15() local
215 acc2 = 0; in arm_conv_partial_fast_opt_q15()
235 acc2 = __SMLAD(x2, y1, acc2); in arm_conv_partial_fast_opt_q15()
252 acc2 = __SMLAD(x1, y2, acc2); in arm_conv_partial_fast_opt_q15()
290 acc2 += (*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()
|
