Home
last modified time | relevance | path

Searched refs:_mm256_cvtepi8_epi32 (Results 1 – 22 of 22) sorted by relevance

/external/XNNPACK/src/qs8-dwconv/gen/
Dup32x9-minmax-avx2-mul32.c89 const __m256i vi0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i0)); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32()
90 …const __m256i vk0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w … in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32()
91 const __m256i vi0x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i0 + 8))); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32()
92 …const __m256i vk0x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w … in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32()
93 … const __m256i vi0xGHIJKLMN = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i0 + 16))); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32()
94 …const __m256i vk0xGHIJKLMN = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w … in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32()
95 … const __m256i vi0xOPQRSTUV = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i0 + 24))); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32()
96 …const __m256i vk0xOPQRSTUV = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w … in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32()
104 const __m256i vi1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i1)); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32()
105 …const __m256i vk1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w … in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32()
[all …]
Dup24x9-minmax-avx2-mul32.c88 const __m256i vi0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i0)); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32()
89 …const __m256i vk0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w … in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32()
90 const __m256i vi0x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i0 + 8))); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32()
91 …const __m256i vk0x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w … in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32()
92 … const __m256i vi0xGHIJKLMN = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i0 + 16))); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32()
93 …const __m256i vk0xGHIJKLMN = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w … in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32()
100 const __m256i vi1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i1)); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32()
101 …const __m256i vk1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w … in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32()
102 const __m256i vi1x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i1 + 8))); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32()
103 …const __m256i vk1x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w … in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32()
[all …]
Dup16x9-minmax-avx2-mul32.c87 const __m256i vi0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i0)); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32()
88 …const __m256i vk0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w … in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32()
89 const __m256i vi0x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i0 + 8))); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32()
90 …const __m256i vk0x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w … in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32()
96 const __m256i vi1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i1)); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32()
97 …const __m256i vk1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w … in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32()
98 const __m256i vi1x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i1 + 8))); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32()
99 …const __m256i vk1x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w … in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32()
105 const __m256i vi2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i2)); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32()
106 …const __m256i vk2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w … in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32()
[all …]
Dup8x9-minmax-avx2-mul32.c86 const __m256i vi0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i0)); in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32()
87 …const __m256i vk0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w … in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32()
92 const __m256i vi1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i1)); in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32()
93 …const __m256i vk1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w … in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32()
98 const __m256i vi2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i2)); in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32()
99 …const __m256i vk2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w … in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32()
104 const __m256i vi3x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i3)); in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32()
105 …const __m256i vk3x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w … in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32()
110 const __m256i vi4x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i4)); in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32()
111 …const __m256i vk4x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w … in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32()
[all …]
/external/XNNPACK/src/qs8-vadd/gen/
Dminmax-avx2-mul32-ld64-x32.c36 const __m256i vx01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_x)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x32()
37 const __m256i vy01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_y)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x32()
38 … const __m256i vx89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_x + 8))); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x32()
39 … const __m256i vy89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_y + 8))); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x32()
40 … const __m256i vxGHIJKLMN = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_x + 16))); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x32()
41 … const __m256i vyGHIJKLMN = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_y + 16))); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x32()
42 … const __m256i vxOPQRSTUV = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_x + 24))); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x32()
43 … const __m256i vyOPQRSTUV = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_y + 24))); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x32()
82 const __m256i vx01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_x)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x32()
83 const __m256i vy01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_y)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x32()
Dminmax-avx2-mul32-ld64-x24.c36 const __m256i vx01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_x)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x24()
37 const __m256i vy01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_y)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x24()
38 … const __m256i vx89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_x + 8))); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x24()
39 … const __m256i vy89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_y + 8))); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x24()
40 … const __m256i vxGHIJKLMN = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_x + 16))); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x24()
41 … const __m256i vyGHIJKLMN = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_y + 16))); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x24()
76 const __m256i vx01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_x)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x24()
77 const __m256i vy01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_y)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x24()
Dminmax-avx2-mul32-ld64-x16.c36 const __m256i vx01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_x)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x16()
37 const __m256i vy01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_y)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x16()
38 … const __m256i vx89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_x + 8))); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x16()
39 … const __m256i vy89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_y + 8))); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x16()
66 const __m256i vx01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_x)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x16()
67 const __m256i vy01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_y)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x16()
Dminmax-avx2-mul32-ld64-x8.c36 const __m256i vx01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_x)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x8()
37 const __m256i vy01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_y)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x8()
60 const __m256i vx01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_x)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x8()
61 const __m256i vy01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_y)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x8()
/external/XNNPACK/src/qs8-vaddc/gen/
Dminmax-avx2-mul32-ld64-x32.c37 const __m256i vx01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_x)); in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x32()
38 … const __m256i vx89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_x + 8))); in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x32()
39 … const __m256i vxGHIJKLMN = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_x + 16))); in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x32()
40 … const __m256i vxOPQRSTUV = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_x + 24))); in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x32()
73 const __m256i vx01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_x)); in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x32()
Dminmax-avx2-mul32-ld64-x24.c37 const __m256i vx01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_x)); in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x24()
38 … const __m256i vx89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_x + 8))); in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x24()
39 … const __m256i vxGHIJKLMN = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_x + 16))); in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x24()
69 const __m256i vx01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_x)); in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x24()
Dminmax-avx2-mul32-ld64-x16.c37 const __m256i vx01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_x)); in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x16()
38 … const __m256i vx89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_x + 8))); in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x16()
61 const __m256i vx01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_x)); in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x16()
Dminmax-avx2-mul32-ld64-x8.c37 const __m256i vx01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_x)); in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x8()
57 const __m256i vx01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_x)); in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x8()
/external/XNNPACK/src/qs8-vadd/
Davx2-mul32-ld64.c.in40 const __m256i vx${ABC[0:8]} = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_x));
41 const __m256i vy${ABC[0:8]} = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_y));
43 …const __m256i vx${ABC[N:N+8]} = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_x + $…
44 …const __m256i vy${ABC[N:N+8]} = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_y + $…
95 const __m256i vx${ABC[0:8]} = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_x));
96 const __m256i vy${ABC[0:8]} = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_y));
/external/XNNPACK/src/qs8-dwconv/
Dunipass-avx2-mul32.c.in52 … const __m256i vi${K}x${ABC[0:8]} = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i${K}));
54 …const __m256i vi${K}x${ABC[C:C+8]} = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i${K} …
55 …const __m256i vk${K}x${ABC[C:C+8]} = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintp…
142 … const __m256i vi${K}x${ABC[0:8]} = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i${K}));
145 … const __m256i vk${K}x${ABC[0:8]} = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) k));
147 …const __m256i vk${K}x${ABC[0:8]} = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + ${K …
149 …const __m256i vk${K}x${ABC[0:8]} = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr…
/external/XNNPACK/src/qs8-vaddc/
Davx2-mul32-ld64.c.in41 const __m256i vx${ABC[0:8]} = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_x));
43 …const __m256i vx${ABC[N:N+8]} = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_x + $…
90 const __m256i vx${ABC[0:8]} = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_x));
/external/libaom/libaom/av1/common/x86/
Dselfguided_avx2.c237 mask[idx] = _mm256_cvtepi8_epi32(_mm_srl_epi64(ones32, shift)); in calc_ab()
374 mask[idx] = _mm256_cvtepi8_epi32(_mm_srl_epi64(ones32, shift)); in calc_ab_fast()
/external/clang/test/CodeGen/
Davx2-builtins.c303 return _mm256_cvtepi8_epi32(a); in test_mm256_cvtepi8_epi32()
/external/llvm-project/clang/test/CodeGen/X86/
Davx2-builtins.c309 return _mm256_cvtepi8_epi32(a); in test_mm256_cvtepi8_epi32()
/external/mesa3d/src/gallium/drivers/swr/rasterizer/core/
Dformat_types.h235 _mm256_cvtepi8_epi32(_mm_castps_si128(_mm256_castps256_ps128(in))));
/external/llvm-project/clang/lib/Headers/
Davx2intrin.h340 _mm256_cvtepi8_epi32(__m128i __V) in _mm256_cvtepi8_epi32() function
Davx512vlintrin.h3998 (__v8si)_mm256_cvtepi8_epi32(__A), in _mm256_mask_cvtepi8_epi32()
4006 (__v8si)_mm256_cvtepi8_epi32(__A), in _mm256_maskz_cvtepi8_epi32()
/external/clang/lib/Headers/
Davx2intrin.h369 _mm256_cvtepi8_epi32(__m128i __V) in _mm256_cvtepi8_epi32() function