Searched refs:vacc0_lo (Results 1 – 3 of 3) sorted by relevance
/external/XNNPACK/src/qu8-vadd/ |
D | minmax-neon.c | 49 int32x4_t vacc0_lo = vmulq_s32(vmovl_s16(vget_low_s16(vxa0)), va_multiplier); in xnn_qu8_vadd_minmax_ukernel__neon() local 58 vacc0_lo = vmlaq_s32(vacc0_lo, vmovl_s16(vget_low_s16(vxb0)), vb_multiplier); in xnn_qu8_vadd_minmax_ukernel__neon() 68 vacc0_lo = vsraq_n_s32(vacc0_lo, vbicq_s32(vacc0_lo, vzero_shift_mask), 31); in xnn_qu8_vadd_minmax_ukernel__neon() 77 vacc0_lo = vrshlq_s32(vacc0_lo, vright_shift); in xnn_qu8_vadd_minmax_ukernel__neon() 87 …const int16x8_t vacc0 = vqaddq_s16(vqmovn_high_s32(vqmovn_s32(vacc0_lo), vacc0_hi), vy_zero_point); in xnn_qu8_vadd_minmax_ukernel__neon() 115 int32x4_t vacc0_lo = vmulq_s32(vmovl_s16(vget_low_s16(vxa0)), va_multiplier); in xnn_qu8_vadd_minmax_ukernel__neon() local 123 vacc0_lo = vmlaq_s32(vacc0_lo, vmovl_s16(vget_low_s16(vxb0)), vb_multiplier); in xnn_qu8_vadd_minmax_ukernel__neon() 129 vacc0_lo = vsraq_n_s32(vacc0_lo, vbicq_s32(vacc0_lo, vzero_shift_mask), 31); in xnn_qu8_vadd_minmax_ukernel__neon() 134 vacc0_lo = vrshlq_s32(vacc0_lo, vright_shift); in xnn_qu8_vadd_minmax_ukernel__neon() 140 …const int16x8_t vacc0 = vqaddq_s16(vcombine_s16(vqmovn_s32(vacc0_lo), vqmovn_s32(vacc0_hi)), vy_ze… in xnn_qu8_vadd_minmax_ukernel__neon()
|
/external/XNNPACK/src/f32-prelu/gen/ |
D | avx-2x8.c | 87 __m128 vacc0_lo = _mm256_castps256_ps128(vacc0); in xnn_f32_prelu_ukernel__avx_2x8() local 90 _mm_storeu_ps(o0, vacc0_lo); in xnn_f32_prelu_ukernel__avx_2x8() 93 vacc0_lo = _mm256_extractf128_ps(vacc0, 1); in xnn_f32_prelu_ukernel__avx_2x8() 100 _mm_storel_pi((__m64*) o0, vacc0_lo); in xnn_f32_prelu_ukernel__avx_2x8() 103 vacc0_lo = _mm_movehl_ps(vacc0_lo, vacc0_lo); in xnn_f32_prelu_ukernel__avx_2x8() 110 _mm_store_ss(o0, vacc0_lo); in xnn_f32_prelu_ukernel__avx_2x8()
|
D | avx-2x16.c | 116 __m128 vacc0_lo = _mm256_castps256_ps128(vacc0); in xnn_f32_prelu_ukernel__avx_2x16() local 119 _mm_storeu_ps(o0, vacc0_lo); in xnn_f32_prelu_ukernel__avx_2x16() 122 vacc0_lo = _mm256_extractf128_ps(vacc0, 1); in xnn_f32_prelu_ukernel__avx_2x16() 129 _mm_storel_pi((__m64*) o0, vacc0_lo); in xnn_f32_prelu_ukernel__avx_2x16() 132 vacc0_lo = _mm_movehl_ps(vacc0_lo, vacc0_lo); in xnn_f32_prelu_ukernel__avx_2x16() 139 _mm_store_ss(o0, vacc0_lo); in xnn_f32_prelu_ukernel__avx_2x16()
|