Lines Matching refs:ABC

8 $ABC = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ"
44     const __m128i vx${ABC[0:4]} = _mm_cvtepi8_epi32(_mm_loadu_si32(input_x));
45     const __m128i vy${ABC[0:4]} = _mm_cvtepi8_epi32(_mm_loadu_si32(input_y));
47       const __m128i vx${ABC[N:N+4]} = _mm_cvtepi8_epi32(_mm_loadu_si32(input_x + ${N}));
48       const __m128i vy${ABC[N:N+4]} = _mm_cvtepi8_epi32(_mm_loadu_si32(input_y + ${N}));
54 …   __m128i vacc${ABC[N:N+4]} = _mm_macc_epi32(vx${ABC[N:N+4]}, vx_multiplier, vzero_point_product);
57         vacc${ABC[N:N+4]} = _mm_macc_epi32(vy${ABC[N:N+4]}, vy_multiplier, vacc${ABC[N:N+4]});
60 …__m128i vacc${ABC[N:N+4]} = _mm_add_epi32(vzero_point_product, _mm_mullo_epi32(vx${ABC[N:N+4]}, vx…
63 …vacc${ABC[N:N+4]} = _mm_add_epi32(vacc${ABC[N:N+4]}, _mm_mullo_epi32(vy${ABC[N:N+4]}, vy_multiplie…
66 …__m128i vrem${ABC[N:N+4]} = _mm_add_epi32(_mm_and_si128(vacc${ABC[N:N+4]}, vremainder_mask), _mm_c…
69 …vacc${ABC[N:N+4]} = _mm_sub_epi32(_mm_sra_epi32(vacc${ABC[N:N+4]}, vshift), _mm_cmpgt_epi32(vrem${…
72 …__m128i vout${ABC[N:N+8]} = _mm_adds_epi16(_mm_packs_epi32(vacc${ABC[N:N+4]}, vacc${ABC[N+4:N+8]})…
75       vout${ABC[N:N+8]} = _mm_max_epi16(vout${ABC[N:N+8]}, voutput_min);
78       vout${ABC[N:N+8]} = _mm_min_epi16(vout${ABC[N:N+8]}, voutput_max);
82         const __m128i vout${ABC[N:N+16]} = _mm_packs_epi16(vout${ABC[N:N+8]}, vout${ABC[N+8:N+16]});
84 …const __m128i vout${ABC[N:N+8]}${ABC[N:N+8]} = _mm_packs_epi16(vout${ABC[N:N+8]}, vout${ABC[N:N+8]…
87       _mm_storeu_si128((__m128i*) output, vout${ABC[0:16]});
89       _mm_storel_epi64((__m128i*) output, vout${ABC[0:8]}${ABC[0:8]});
92         _mm_storeu_si128((__m128i*) (output + ${N}), vout${ABC[N:N+16]});
94         _mm_storel_epi64((__m128i*) (output + ${N}), vout${ABC[N:N+8]}${ABC[N:N+8]});
99       const __m128i vx${ABC[0:4]} = _mm_cvtepi8_epi32(_mm_loadu_si32(input_x));
100       const __m128i vy${ABC[0:4]} = _mm_cvtepi8_epi32(_mm_loadu_si32(input_y));
101       const __m128i vx${ABC[4:8]} = _mm_cvtepi8_epi32(_mm_loadu_si32(input_x + 4));
102       const __m128i vy${ABC[4:8]} = _mm_cvtepi8_epi32(_mm_loadu_si32(input_y + 4));
108         __m128i vacc${ABC[0:4]} = _mm_macc_epi32(vx${ABC[0:4]}, vx_multiplier, vzero_point_product);
109         __m128i vacc${ABC[4:8]} = _mm_macc_epi32(vx${ABC[4:8]}, vx_multiplier, vzero_point_product);
111         vacc${ABC[0:4]} = _mm_macc_epi32(vy${ABC[0:4]}, vy_multiplier, vacc${ABC[0:4]});
112         vacc${ABC[4:8]} = _mm_macc_epi32(vy${ABC[4:8]}, vy_multiplier, vacc${ABC[4:8]});
114 …__m128i vacc${ABC[0:4]} = _mm_add_epi32(vzero_point_product, _mm_mullo_epi32(vx${ABC[0:4]}, vx_mul…
115 …__m128i vacc${ABC[4:8]} = _mm_add_epi32(vzero_point_product, _mm_mullo_epi32(vx${ABC[4:8]}, vx_mul…
117 …   vacc${ABC[0:4]} = _mm_add_epi32(vacc${ABC[0:4]}, _mm_mullo_epi32(vy${ABC[0:4]}, vy_multiplier));
118 …   vacc${ABC[4:8]} = _mm_add_epi32(vacc${ABC[4:8]}, _mm_mullo_epi32(vy${ABC[4:8]}, vy_multiplier));
120 …t __m128i vrem${ABC[0:4]} = _mm_add_epi32(_mm_and_si128(vacc${ABC[0:4]}, vremainder_mask), _mm_cmp…
121 …t __m128i vrem${ABC[4:8]} = _mm_add_epi32(_mm_and_si128(vacc${ABC[4:8]}, vremainder_mask), _mm_cmp…
123 …vacc${ABC[0:4]} = _mm_sub_epi32(_mm_sra_epi32(vacc${ABC[0:4]}, vshift), _mm_cmpgt_epi32(vrem${ABC[…
124 …vacc${ABC[4:8]} = _mm_sub_epi32(_mm_sra_epi32(vacc${ABC[4:8]}, vshift), _mm_cmpgt_epi32(vrem${ABC[…
126 …__m128i vout${ABC[0:8]} = _mm_adds_epi16(_mm_packs_epi32(vacc${ABC[0:4]}, vacc${ABC[4:8]}), voutpu…
127       vout${ABC[0:8]} = _mm_max_epi16(vout${ABC[0:8]}, voutput_min);
128       vout${ABC[0:8]} = _mm_min_epi16(vout${ABC[0:8]}, voutput_max);
130       __m128i vout${ABC[0:8]}${ABC[0:8]} = _mm_packs_epi16(vout${ABC[0:8]}, vout${ABC[0:8]});
134           _mm_storel_epi64((__m128i*) output, vout${ABC[0:8]}${ABC[0:8]});
139             *((uint32_t*) output) = (uint32_t) _mm_cvtsi128_si32(vout${ABC[0:8]}${ABC[0:8]});
140             vout${ABC[0:8]}${ABC[0:8]} = _mm_srli_epi64(vout${ABC[0:8]}${ABC[0:8]}, 32);
144             *((uint16_t*) output) = (uint16_t) _mm_extract_epi16(vout${ABC[0:8]}${ABC[0:8]}, 0);
145             vout${ABC[0:8]}${ABC[0:8]} = _mm_srli_epi32(vout${ABC[0:8]}${ABC[0:8]}, 16);
150               *output = (int8_t) _mm_extract_epi8(vout${ABC[0:8]}${ABC[0:8]}, 0);
152               *output = (int32_t) _mm_cvtsi128_si32(vout${ABC[0:8]}${ABC[0:8]});
158           *((uint32_t*) output) = (uint32_t) _mm_cvtsi128_si32(vout${ABC[0:8]}${ABC[0:8]});
159           vout${ABC[0:8]}${ABC[0:8]} = _mm_srli_epi64(vout${ABC[0:8]}${ABC[0:8]}, 32);
163           *((uint16_t*) output) = (uint16_t) _mm_extract_epi16(vout${ABC[0:8]}${ABC[0:8]}, 0);
164           vout${ABC[0:8]}${ABC[0:8]} = _mm_srli_epi32(vout${ABC[0:8]}${ABC[0:8]}, 16);
169             *output = (int8_t) _mm_extract_epi8(vout${ABC[0:8]}${ABC[0:8]}, 0);
171             *output = (int32_t) _mm_cvtsi128_si32(vout${ABC[0:8]}${ABC[0:8]});