Searched refs:_lo (Results 1 – 15 of 15) sorted by relevance
98 #define COUNTER(_sel, _lo, _hi) { \ argument100 .counter_reg_lo = REG(_lo), \104 #define COUNTER2(_sel, _lo, _hi, _en, _clr) { \ argument106 .counter_reg_lo = REG(_lo), \
119 __m128 vacc${M}_lo = _mm256_castps256_ps128(vacc${M});122 _mm_storeu_ps(o${M}, vacc${M}_lo);125 vacc${M}_lo = _mm256_extractf128_ps(vacc${M}, 1);132 _mm_storel_pi((__m64*) o${M}, vacc${M}_lo);135 vacc${M}_lo = _mm_movehl_ps(vacc${M}_lo, vacc${M}_lo);142 _mm_store_ss(o${M}, vacc${M}_lo);
60 … const __m128i vidx${N}_lo = _mm_slli_epi32(_mm_castps_si128(_mm256_castps256_ps128(vidx${N})), 2);63 const uint64_t vidx${N}_ll = (uint64_t) _mm_cvtsi128_si64(vidx${N}_lo);64 const uint64_t vidx${N}_lh = (uint64_t) _mm_extract_epi64(vidx${N}_lo, 1);76 …st void*) ((uintptr_t) xnn_table_exp2minus_k_over_16 + (uint32_t) _mm_cvtsi128_si32(vidx${N}_lo)));77 …void*) ((uintptr_t) xnn_table_exp2minus_k_over_16 + (uint32_t) _mm_extract_epi32(vidx${N}_lo, 2)));80 …*) ((uintptr_t) xnn_table_exp2minus_k_over_16 + (uint32_t) _mm_extract_epi32(vidx${N}_lo, 1))), 1);81 …*) ((uintptr_t) xnn_table_exp2minus_k_over_16 + (uint32_t) _mm_extract_epi32(vidx${N}_lo, 3))), 1);85 const __m128i vl${N}_lo = _mm_unpacklo_epi64(vl${N}_ll, vl${N}_lh);89 … const __m128i ven${N}_lo = _mm_slli_epi32(_mm_castps_si128(_mm256_castps256_ps128(vn${N})), 19);92 const __m128 vs${N}_lo = _mm_castsi128_ps(_mm_add_epi32(vl${N}_lo, ven${N}_lo));[all …]
57 …const __m128 vs${N}_lo = _mm_castsi128_ps(_mm_slli_epi32(_mm_castps_si128(_mm256_castps256_ps128(v…63 __m256 vs${N} = _mm256_insertf128_ps(_mm256_castps128_ps256(vs${N}_lo), vs${N}_hi, 1);
61 …const __m128 ven${N}_lo = _mm_castsi128_ps(_mm_slli_epi32(_mm_castps_si128(_mm256_castps256_ps128(…69 ven${N} = _mm256_insertf128_ps(_mm256_castps128_ps256(ven${N}_lo), ven${N}_hi, 1);
519 dst##_lo[element] = src##_lo[element]; \525 b##_lo[b_index] = \533 b##_lo[b_index] = \541 c##_lo[c_index] = vaddw_s16(a##_lo[a_index], vget_low_s16(b[b_index])); \547 temp##_lo[temp_index] = vmovl_s16(vget_low_s16(a[a_index])); \549 c##_lo[c_index] = vsubq_s32(temp##_lo[temp_index], b##_lo[b_index]); \555 b##_lo[b_index] = vaddq_s32(a##_lo[left_index], a##_lo[right_index]); \561 b##_lo[b_index] = vsubq_s32(a##_lo[left_index], a##_lo[right_index]); \586 &b##_lo[add_index], &b##_hi[add_index], \587 &b##_lo[sub_index], &b##_hi[sub_index]); \[all …]
16 ; GCN: v_cmp_eq_u32{{(_e[0-9]+)?}} vcc{{(_lo)?}}, 0, v[[mbcnt]]34 ; GCN: v_cmp_eq_u32{{(_e[0-9]+)?}} vcc{{(_lo)?}}, 0, v[[mbcnt]]74 ; GCN: v_cmp_eq_u32{{(_e[0-9]+)?}} vcc{{(_lo)?}}, 0, v[[mbcnt]]93 ; GCN: v_cmp_eq_u32{{(_e[0-9]+)?}} vcc{{(_lo)?}}, 0, v[[mbcnt]]124 ; GCN: v_cmp_eq_u32{{(_e[0-9]+)?}} vcc{{(_lo)?}}, 0, v[[mbcnt]]142 ; GCN: v_cmp_eq_u32{{(_e[0-9]+)?}} vcc{{(_lo)?}}, 0, v[[mbcnt]]182 ; GCN: v_cmp_eq_u32{{(_e[0-9]+)?}} vcc{{(_lo)?}}, 0, v[[mbcnt]]201 ; GCN: v_cmp_eq_u32{{(_e[0-9]+)?}} vcc{{(_lo)?}}, 0, v[[mbcnt]]
18 ; GCN: v_cmp_eq_u32{{(_e[0-9]+)?}} vcc{{(_lo)?}}, 0, v[[mbcnt]]36 ; GCN: v_cmp_eq_u32{{(_e[0-9]+)?}} vcc{{(_lo)?}}, 0, v[[mbcnt]]89 ; GCN: v_cmp_eq_u32{{(_e[0-9]+)?}} vcc{{(_lo)?}}, 0, v[[mbcnt]]107 ; GCN: v_cmp_eq_u32{{(_e[0-9]+)?}} vcc{{(_lo)?}}, 0, v[[mbcnt]]
18 ; GCN: v_cmp_eq_u32{{(_e[0-9]+)?}} vcc{{(_lo)?}}, 0, v[[mbcnt]]36 ; GCN: v_cmp_eq_u32{{(_e[0-9]+)?}} vcc{{(_lo)?}}, 0, v[[mbcnt]]102 ; GCN: v_cmp_eq_u32{{(_e[0-9]+)?}} vcc{{(_lo)?}}, 0, v[[mbcnt]]120 ; GCN: v_cmp_eq_u32{{(_e[0-9]+)?}} vcc{{(_lo)?}}, 0, v[[mbcnt]]
19 ; GCN: v_cmp_eq_u32{{(_e[0-9]+)?}} vcc{{(_lo)?}}, 0, v[[mbcnt]]37 ; GCN: v_cmp_eq_u32{{(_e[0-9]+)?}} vcc{{(_lo)?}}, 0, v[[mbcnt]]121 ; GCN: v_cmp_eq_u32{{(_e[0-9]+)?}} vcc{{(_lo)?}}, 0, v[[mbcnt]]139 ; GCN: v_cmp_eq_u32{{(_e[0-9]+)?}} vcc{{(_lo)?}}, 0, v[[mbcnt]]
213 float32x2_t vo${M}_lo = vget_low_f32(vo${M});216 vst1_f32(o${M}, vo${M}_lo); o${M} += 2;219 vo${M}_lo = vget_high_f32(vo${M});223 vst1_lane_f32(o${M}, vo${M}_lo, 0); o${M} += 1;
228 float32x2_t vo${M}_lo = vget_low_f32(vo${M});231 vst1_f32(o${M}, vo${M}_lo); o${M} += 2;234 vo${M}_lo = vget_high_f32(vo${M});238 vst1_lane_f32(o${M}, vo${M}_lo, 0); o${M} += 1;
346 float32x2_t vo${M}_lo = vget_low_f32(vo${M});349 vst1_f32(o${M}, vo${M}_lo); o${M} += 2;352 vo${M}_lo = vget_high_f32(vo${M});356 vst1_lane_f32(o${M}, vo${M}_lo, 0); o${M} += 1;
468 float32x2_t vo${M}_lo = vget_low_f32(vo${M});471 vst1_f32(o${M}, vo${M}_lo); o${M} += 2;474 vo${M}_lo = vget_high_f32(vo${M});478 vst1_lane_f32(o${M}, vo${M}_lo, 0); o${M} += 1;
850 #define CLAMP( _lo, _x, _hi ) ( (_x) < (_lo) ? (_lo) : ((_x) > (_hi) ? (_hi) : (_x))) argument