/third_party/cmsis/CMSIS/DSP/Source/MatrixFunctions/ |
D | arm_mat_cmplx_mult_q15.c | 80 uint16x8_t vecOffs, vecColBOffs; in arm_mat_cmplx_mult_q15() local 164 vecOffs = vecColBOffs; in arm_mat_cmplx_mult_q15() 172 vecB = vldrhq_gather_shifted_offset(pInB, vecOffs); in arm_mat_cmplx_mult_q15() 176 vecB2 = vldrhq_gather_shifted_offset(pInB2, vecOffs); in arm_mat_cmplx_mult_q15() 180 vecOffs = vaddq_n_u16(vecOffs, (uint16_t) (numColsB * 4 * CMPLX_DIM)); in arm_mat_cmplx_mult_q15() 195 vecB = vldrhq_gather_shifted_offset(pInB, vecOffs); in arm_mat_cmplx_mult_q15() 201 vecB2 = vldrhq_gather_shifted_offset(pInB2, vecOffs); in arm_mat_cmplx_mult_q15() 206 vecOffs = vaddq_n_u16(vecOffs, (uint16_t) (numColsB * 4 * CMPLX_DIM)); in arm_mat_cmplx_mult_q15() 256 vecOffs = vecColBOffs; in arm_mat_cmplx_mult_q15() 265 vecB = vldrhq_gather_shifted_offset(pInB, vecOffs); in arm_mat_cmplx_mult_q15() [all …]
|
D | arm_mat_trans_q7.c | 52 uint16x8_t vecOffs; in arm_mat_trans_q7() local 71 vecOffs = vidupq_u16((uint32_t)0, 1); in arm_mat_trans_q7() 72 vecOffs = vecOffs * pSrc->numCols; in arm_mat_trans_q7() 84 vecIn = vldrbq_gather_offset_u16(pDataC, vecOffs); in arm_mat_trans_q7() 102 vecIn = vldrbq_gather_offset_u16(pDataC, vecOffs); in arm_mat_trans_q7()
|
D | arm_mat_mult_q15.c | 332 uint16x8_t vecOffs, vecColBOffs; in arm_mat_mult_q15() local 424 vecOffs = vecColBOffs; in arm_mat_mult_q15() 431 vecB = vldrhq_gather_shifted_offset((int16_t const *)pInB, vecOffs); in arm_mat_mult_q15() 432 vecOffs = vecOffs + (uint16_t) (numColsB * 8); in arm_mat_mult_q15() 454 vecB = vldrhq_gather_shifted_offset((int16_t const *)pInB, vecOffs); in arm_mat_mult_q15() 455 vecOffs = vecOffs + (uint16_t) (numColsB * 8); in arm_mat_mult_q15() 530 vecOffs = vecColBOffs; in arm_mat_mult_q15() 537 vecB = vldrhq_gather_shifted_offset((int16_t const *)pInB, vecOffs); in arm_mat_mult_q15() 538 vecOffs = vecOffs + (uint16_t) (numColsB * 8); in arm_mat_mult_q15() 556 vecB = vldrhq_gather_shifted_offset((int16_t const *)pInB, vecOffs); in arm_mat_mult_q15() [all …]
|
D | arm_mat_cmplx_mult_f16.c | 392 uint16x8_t vecOffs, vecColBOffs; in arm_mat_cmplx_mult_f16() local 492 vecOffs = vecColBOffs; in arm_mat_cmplx_mult_f16() 502 vecB = vldrhq_gather_shifted_offset_f16(pInB, vecOffs); in arm_mat_cmplx_mult_f16() 506 vecOffs = vaddq_n_u16(vecOffs , (uint16_t) (numColsB * 4 * CMPLX_DIM)); in arm_mat_cmplx_mult_f16() 539 vecB = vldrhq_gather_shifted_offset_z_f16(pInB, vecOffs, p0); in arm_mat_cmplx_mult_f16() 543 vecOffs = vaddq_n_u16(vecOffs, (uint16_t) (numColsB * 4 * CMPLX_DIM)); in arm_mat_cmplx_mult_f16() 628 vecOffs = vecColBOffs; in arm_mat_cmplx_mult_f16() 638 vecB = vldrhq_gather_shifted_offset(pInB, vecOffs); in arm_mat_cmplx_mult_f16() 642 vecOffs = vaddq_n_u16(vecOffs, (uint16_t) (4*numColsB * CMPLX_DIM)); in arm_mat_cmplx_mult_f16() 663 vecB = vldrhq_gather_shifted_offset_z(pInB, vecOffs, p0); in arm_mat_cmplx_mult_f16()
|
D | arm_mat_mult_q31.c | 349 uint32x4_t vecOffs, vecColBOffs; in arm_mat_mult_q31() local 439 vecOffs = vecColBOffs; in arm_mat_mult_q31() 447 vecB = vldrwq_gather_shifted_offset(pInB, vecOffs); in arm_mat_mult_q31() 449 vecOffs = vecOffs + (uint32_t) (numColsB * 4); in arm_mat_mult_q31() 472 vecB = vldrwq_gather_shifted_offset_z(pInB, vecOffs, p0); in arm_mat_mult_q31() 553 vecOffs = vecColBOffs; in arm_mat_mult_q31() 561 vecB = vldrwq_gather_shifted_offset(pInB, vecOffs); in arm_mat_mult_q31() 563 vecOffs = vecOffs + (uint32_t) (numColsB * 4); in arm_mat_mult_q31() 581 vecB = vldrwq_gather_shifted_offset_z(pInB, vecOffs, p0); in arm_mat_mult_q31()
|
D | arm_mat_cmplx_mult_q31.c | 531 uint32x4_t vecOffs, vecColBOffs; in arm_mat_cmplx_mult_q31() local 629 vecOffs = vecColBOffs; in arm_mat_cmplx_mult_q31() 639 vecB = vldrwq_gather_shifted_offset(pInB, vecOffs); in arm_mat_cmplx_mult_q31() 643 vecOffs = vecOffs + (uint32_t) (numColsB * 2 * CMPLX_DIM); in arm_mat_cmplx_mult_q31() 672 vecB = vldrwq_gather_shifted_offset_z(pInB, vecOffs, p0); in arm_mat_cmplx_mult_q31() 677 vecOffs = vecOffs + (uint32_t) (numColsB * 2 * CMPLX_DIM); in arm_mat_cmplx_mult_q31() 754 vecOffs = vecColBOffs; in arm_mat_cmplx_mult_q31() 764 vecB = vldrwq_gather_shifted_offset(pInB, vecOffs); in arm_mat_cmplx_mult_q31() 768 vecOffs = vecOffs + (uint32_t) (numColsB * 2 * CMPLX_DIM); in arm_mat_cmplx_mult_q31() 789 vecB = vldrwq_gather_shifted_offset_z(pInB, vecOffs, p0); in arm_mat_cmplx_mult_q31() [all …]
|
D | arm_mat_ldlt_f32.c | 179 int32x4_t vecOffs; in arm_mat_ldlt_f32() local 181 vecOffs = vidupq_u32((uint32_t)0, 1); in arm_mat_ldlt_f32() 182 vecOffs = vmulq_n_s32(vecOffs,n); in arm_mat_ldlt_f32() 207 vecX = vldrwq_gather_shifted_offset_z_f32(&pA[x*n+k], vecOffs, p0); in arm_mat_ldlt_f32() 250 vecX = vldrwq_gather_shifted_offset_z_f32(&pA[x*n+k], vecOffs, p0); in arm_mat_ldlt_f32()
|
D | arm_mat_cmplx_mult_f32.c | 537 uint32x4_t vecOffs, vecColBOffs; in arm_mat_cmplx_mult_f32() local 632 vecOffs = vecColBOffs; in arm_mat_cmplx_mult_f32() 642 vecB = vldrwq_gather_shifted_offset(pInB, vecOffs); in arm_mat_cmplx_mult_f32() 646 vecOffs = vecOffs + (uint32_t) (numColsB * 2 * CMPLX_DIM); in arm_mat_cmplx_mult_f32() 675 vecB = vldrwq_gather_shifted_offset_z(pInB, vecOffs, p0); in arm_mat_cmplx_mult_f32() 679 vecOffs = vecOffs + (uint32_t) (numColsB * 2 * CMPLX_DIM); in arm_mat_cmplx_mult_f32() 763 vecOffs = vecColBOffs; in arm_mat_cmplx_mult_f32() 773 vecB = vldrwq_gather_shifted_offset(pInB, vecOffs); in arm_mat_cmplx_mult_f32() 777 vecOffs = vecOffs + (uint32_t) (numColsB * 2 * CMPLX_DIM); in arm_mat_cmplx_mult_f32() 798 vecB = vldrwq_gather_shifted_offset_z(pInB, vecOffs, p0); in arm_mat_cmplx_mult_f32()
|
/third_party/cmsis/CMSIS/DSP/Include/ |
D | arm_helium_utils.h | 195 static const uint32x4_t vecOffs = { 0, 2, 1, 3 }; in arm_mat_trans_32bit_2x2_mve() local 203 vstrwq_scatter_shifted_offset_u32(pDataDest, vecOffs, vecIn); in arm_mat_trans_32bit_2x2_mve() 265 uint32x4_t vecOffs; in arm_mat_trans_32bit_generic_mve() local 272 vecOffs = vidupq_u32((uint32_t)0, 1); in arm_mat_trans_32bit_generic_mve() 273 vecOffs = vecOffs * srcCols; in arm_mat_trans_32bit_generic_mve() 284 vecIn = vldrwq_gather_shifted_offset_u32(pDataC, vecOffs); in arm_mat_trans_32bit_generic_mve() 301 vecIn = vldrwq_gather_shifted_offset_u32(pDataC, vecOffs); in arm_mat_trans_32bit_generic_mve() 464 uint16x8_t vecOffs; in arm_mat_trans_16bit_generic() local 471 vecOffs = vidupq_u16((uint32_t)0, 1); in arm_mat_trans_16bit_generic() 472 vecOffs = vecOffs * srcCols; in arm_mat_trans_16bit_generic() [all …]
|