Home
last modified time | relevance | path

Searched refs:xnn_table_exp2minus_k_over_64 (Results 1 – 25 of 81) sorted by relevance

1234

/external/XNNPACK/src/f32-sigmoid/gen/
Dsse41-lut64-p2-div-x24.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
78 …__m128i vl0 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x24()
79 …t __m128i vl2 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x24()
80 …128i vl01 = _mm_insert_epi32(vl0, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x24()
81 …128i vl23 = _mm_insert_epi32(vl2, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x24()
85 …__m128i vl4 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x24()
86 …t __m128i vl6 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x24()
87 …128i vl45 = _mm_insert_epi32(vl4, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x24()
88 …128i vl67 = _mm_insert_epi32(vl6, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x24()
92 …__m128i vl8 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x24()
[all …]
Dsse41-lut64-p2-div-x16.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
68 …__m128i vl0 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x16()
69 …t __m128i vl2 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x16()
70 …128i vl01 = _mm_insert_epi32(vl0, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x16()
71 …128i vl23 = _mm_insert_epi32(vl2, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x16()
75 …__m128i vl4 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x16()
76 …t __m128i vl6 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x16()
77 …128i vl45 = _mm_insert_epi32(vl4, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x16()
78 …128i vl67 = _mm_insert_epi32(vl6, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x16()
82 …__m128i vl8 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x16()
[all …]
Dsse41-lut64-p2-div-x20.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
73 …__m128i vl0 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x20()
74 …t __m128i vl2 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x20()
75 …128i vl01 = _mm_insert_epi32(vl0, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x20()
76 …128i vl23 = _mm_insert_epi32(vl2, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x20()
80 …__m128i vl4 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x20()
81 …t __m128i vl6 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x20()
82 …128i vl45 = _mm_insert_epi32(vl4, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x20()
83 …128i vl67 = _mm_insert_epi32(vl6, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x20()
87 …__m128i vl8 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x20()
[all …]
Dsse41-lut64-p2-div-x12.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
63 …__m128i vl0 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x12()
64 …t __m128i vl2 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x12()
65 …128i vl01 = _mm_insert_epi32(vl0, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x12()
66 …128i vl23 = _mm_insert_epi32(vl2, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x12()
70 …__m128i vl4 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x12()
71 …t __m128i vl6 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x12()
72 …128i vl45 = _mm_insert_epi32(vl4, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x12()
73 …128i vl67 = _mm_insert_epi32(vl6, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x12()
77 …__m128i vl8 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x12()
[all …]
Dsse41-lut64-p2-div-x8.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
58 …__m128i vl0 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x8()
59 …t __m128i vl2 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x8()
60 …128i vl01 = _mm_insert_epi32(vl0, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x8()
61 …128i vl23 = _mm_insert_epi32(vl2, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x8()
65 …__m128i vl4 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x8()
66 …t __m128i vl6 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x8()
67 …128i vl45 = _mm_insert_epi32(vl4, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x8()
68 …128i vl67 = _mm_insert_epi32(vl6, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x8()
75 …__m128i vl0 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + vidx… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x8()
[all …]
Dsse2-lut64-p2-div-x20.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
73 …__m128i vl0 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x20()
74 …t __m128i vl2 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x20()
75 …t __m128i vl1 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x20()
77 …t __m128i vl3 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x20()
82 …__m128i vl4 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x20()
83 …t __m128i vl6 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x20()
84 …t __m128i vl5 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x20()
86 …t __m128i vl7 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x20()
91 …__m128i vl8 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x20()
[all …]
Dsse2-lut64-p2-div-x24.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
78 …__m128i vl0 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x24()
79 …t __m128i vl2 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x24()
80 …t __m128i vl1 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x24()
82 …t __m128i vl3 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x24()
87 …__m128i vl4 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x24()
88 …t __m128i vl6 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x24()
89 …t __m128i vl5 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x24()
91 …t __m128i vl7 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x24()
96 …__m128i vl8 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x24()
[all …]
Dsse2-lut64-p2-div-x12.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
63 …__m128i vl0 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x12()
64 …t __m128i vl2 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x12()
65 …t __m128i vl1 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x12()
67 …t __m128i vl3 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x12()
72 …__m128i vl4 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x12()
73 …t __m128i vl6 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x12()
74 …t __m128i vl5 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x12()
76 …t __m128i vl7 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x12()
81 …__m128i vl8 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x12()
[all …]
Dsse2-lut64-p2-div-x16.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
68 …__m128i vl0 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x16()
69 …t __m128i vl2 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x16()
70 …t __m128i vl1 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x16()
72 …t __m128i vl3 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x16()
77 …__m128i vl4 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x16()
78 …t __m128i vl6 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x16()
79 …t __m128i vl5 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x16()
81 …t __m128i vl7 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x16()
86 …__m128i vl8 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x16()
[all …]
Dsse2-lut64-p2-div-x8.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
58 …__m128i vl0 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x8()
59 …t __m128i vl2 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x8()
60 …t __m128i vl1 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x8()
62 …t __m128i vl3 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x8()
67 …__m128i vl4 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x8()
68 …t __m128i vl6 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x8()
69 …t __m128i vl5 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x8()
71 …t __m128i vl7 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x8()
79 …__m128i vl0 = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + vidx… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x8()
[all …]
Dsse41-lut64-p2-div-x4.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
51 …m128i vl_ll = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x4()
52 …__m128i vl_hl = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x4()
53 …i vl_lo = _mm_insert_epi32(vl_ll, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x4()
54 …i vl_hi = _mm_insert_epi32(vl_hl, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x4()
56 …__m128i vl_ll = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x4()
57 …__m128i vl_hl = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x4()
58 …i vl_lo = _mm_insert_epi32(vl_ll, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x4()
59 …i vl_hi = _mm_insert_epi32(vl_hl, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x4()
93 …m128i vl_ll = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse41_lut64_p2_div_x4()
[all …]
Dneonfma-rr1-lut64-p2-div-x24.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
75 float32x2_t vl01 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx01]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x24()
76 float32x2_t vl23 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx23]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x24()
79 float32x2_t vl45 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx45]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x24()
80 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx67]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x24()
83 float32x2_t vl89 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx89]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x24()
84 float32x2_t vlAB = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxAB]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x24()
87 float32x2_t vlCD = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxCD]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x24()
88 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxEF]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x24()
91 float32x2_t vlGH = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxGH]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x24()
[all …]
Dwasmsimd-lut64-p2-div-x24.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
76 …const float vl0 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) vidx0… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x24()
77 …const float vl1 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) (vidx01… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x24()
78 …const float vl2 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) vidx23)… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x24()
79 …const float vl3 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) (vidx23… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x24()
83 …const float vl4 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) vidx4… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x24()
84 …const float vl5 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) (vidx45… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x24()
85 …const float vl6 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) vidx67)… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x24()
86 …const float vl7 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) (vidx67… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x24()
90 …const float vl8 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) vidx8… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x24()
[all …]
Dsse2-lut64-p2-div-x4.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
51 …m128i vl_ll = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x4()
52 …__m128i vl_hl = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x4()
53 …__m128i vl_lh = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x4()
55 …__m128i vl_hh = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x4()
58 …__m128i vl_ll = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x4()
59 …__m128i vl_hl = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x4()
60 …__m128i vl_lh = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x4()
62 …__m128i vl_hh = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x4()
98 …m128i vl_ll = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_f32_sigmoid_ukernel__sse2_lut64_p2_div_x4()
[all …]
Dwasmsimd-lut64-p2-div-x20.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
71 …const float vl0 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) vidx0… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x20()
72 …const float vl1 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) (vidx01… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x20()
73 …const float vl2 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) vidx23)… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x20()
74 …const float vl3 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) (vidx23… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x20()
78 …const float vl4 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) vidx4… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x20()
79 …const float vl5 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) (vidx45… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x20()
80 …const float vl6 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) vidx67)… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x20()
81 …const float vl7 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) (vidx67… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x20()
85 …const float vl8 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) vidx8… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x20()
[all …]
Dneonfma-rr1-lut64-p2-div-x20.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
70 float32x2_t vl01 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx01]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x20()
71 float32x2_t vl23 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx23]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x20()
74 float32x2_t vl45 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx45]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x20()
75 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx67]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x20()
78 float32x2_t vl89 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx89]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x20()
79 float32x2_t vlAB = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxAB]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x20()
82 float32x2_t vlCD = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxCD]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x20()
83 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxEF]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x20()
86 float32x2_t vlGH = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxGH]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x20()
[all …]
Dwasmsimd-lut64-p2-div-x16.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
66 …const float vl0 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) vidx0… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x16()
67 …const float vl1 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) (vidx01… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x16()
68 …const float vl2 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) vidx23)… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x16()
69 …const float vl3 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) (vidx23… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x16()
73 …const float vl4 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) vidx4… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x16()
74 …const float vl5 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) (vidx45… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x16()
75 …const float vl6 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) vidx67)… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x16()
76 …const float vl7 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) (vidx67… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x16()
80 …const float vl8 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) vidx8… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x16()
[all …]
Dneonfma-rr1-lut64-p2-div-x16.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
65 float32x2_t vl01 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx01]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x16()
66 float32x2_t vl23 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx23]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x16()
69 float32x2_t vl45 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx45]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x16()
70 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx67]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x16()
73 float32x2_t vl89 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx89]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x16()
74 float32x2_t vlAB = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxAB]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x16()
77 float32x2_t vlCD = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxCD]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x16()
78 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxEF]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x16()
80 vl01 = vld1_lane_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) (vidx01 >> 32)], vl01, 1); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x16()
[all …]
Dwasmsimd-lut64-p2-div-x12.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
61 …const float vl0 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) vidx0… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x12()
62 …const float vl1 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) (vidx01… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x12()
63 …const float vl2 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) vidx23)… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x12()
64 …const float vl3 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) (vidx23… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x12()
68 …const float vl4 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) vidx4… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x12()
69 …const float vl5 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) (vidx45… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x12()
70 …const float vl6 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) vidx67)… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x12()
71 …const float vl7 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) (vidx67… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x12()
75 …const float vl8 = *((const float*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint32_t) vidx8… in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x12()
[all …]
Dneonfma-rr1-lut64-p2-div-x12.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
60 float32x2_t vl01 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx01]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x12()
61 float32x2_t vl23 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx23]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x12()
64 float32x2_t vl45 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx45]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x12()
65 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx67]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x12()
68 float32x2_t vl89 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx89]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x12()
69 float32x2_t vlAB = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxAB]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x12()
71 vl01 = vld1_lane_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) (vidx01 >> 32)], vl01, 1); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x12()
72 vl23 = vld1_lane_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) (vidx23 >> 32)], vl23, 1); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x12()
74 vl45 = vld1_lane_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) (vidx45 >> 32)], vl45, 1); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x12()
[all …]
Dneonfma-rr1-lut64-p2-nr2fma-x24.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
75 float32x2_t vl01 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx01]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2fma_x24()
76 float32x2_t vl23 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx23]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2fma_x24()
79 float32x2_t vl45 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx45]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2fma_x24()
80 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx67]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2fma_x24()
83 float32x2_t vl89 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx89]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2fma_x24()
84 float32x2_t vlAB = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxAB]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2fma_x24()
87 float32x2_t vlCD = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxCD]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2fma_x24()
88 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxEF]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2fma_x24()
91 float32x2_t vlGH = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxGH]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2fma_x24()
[all …]
Dneonfma-rr1-lut64-p2-nr2recps-x24.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
75 float32x2_t vl01 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx01]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2recps_x24()
76 float32x2_t vl23 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx23]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2recps_x24()
79 float32x2_t vl45 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx45]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2recps_x24()
80 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx67]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2recps_x24()
83 float32x2_t vl89 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx89]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2recps_x24()
84 float32x2_t vlAB = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxAB]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2recps_x24()
87 float32x2_t vlCD = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxCD]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2recps_x24()
88 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxEF]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2recps_x24()
91 float32x2_t vlGH = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxGH]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2recps_x24()
[all …]
Dneonfma-rr1-lut64-p2-nr1recps1fma-x24.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
75 float32x2_t vl01 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx01]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr1recps1fma_x24()
76 float32x2_t vl23 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx23]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr1recps1fma_x24()
79 float32x2_t vl45 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx45]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr1recps1fma_x24()
80 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx67]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr1recps1fma_x24()
83 float32x2_t vl89 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx89]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr1recps1fma_x24()
84 float32x2_t vlAB = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxAB]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr1recps1fma_x24()
87 float32x2_t vlCD = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxCD]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr1recps1fma_x24()
88 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxEF]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr1recps1fma_x24()
91 float32x2_t vlGH = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxGH]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr1recps1fma_x24()
[all …]
Dneon-rr2-lut64-p2-nr2recps-x24.c18 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
76 float32x2_t vl01 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx01]); in xnn_f32_sigmoid_ukernel__neon_rr2_lut64_p2_nr2recps_x24()
77 float32x2_t vl23 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx23]); in xnn_f32_sigmoid_ukernel__neon_rr2_lut64_p2_nr2recps_x24()
80 float32x2_t vl45 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx45]); in xnn_f32_sigmoid_ukernel__neon_rr2_lut64_p2_nr2recps_x24()
81 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx67]); in xnn_f32_sigmoid_ukernel__neon_rr2_lut64_p2_nr2recps_x24()
84 float32x2_t vl89 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx89]); in xnn_f32_sigmoid_ukernel__neon_rr2_lut64_p2_nr2recps_x24()
85 float32x2_t vlAB = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxAB]); in xnn_f32_sigmoid_ukernel__neon_rr2_lut64_p2_nr2recps_x24()
88 float32x2_t vlCD = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxCD]); in xnn_f32_sigmoid_ukernel__neon_rr2_lut64_p2_nr2recps_x24()
89 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxEF]); in xnn_f32_sigmoid_ukernel__neon_rr2_lut64_p2_nr2recps_x24()
92 float32x2_t vlGH = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxGH]); in xnn_f32_sigmoid_ukernel__neon_rr2_lut64_p2_nr2recps_x24()
[all …]
/external/XNNPACK/src/math/
Dsigmoid-avx-rr2-lut64-p2-div.c16 extern XNN_INTERNAL const float xnn_table_exp2minus_k_over_64[64];
84 …__m128i vl_ll = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_math_f32_sigmoid__avx_rr2_lut64_p2_div()
85 …__m128i vl_lh = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_math_f32_sigmoid__avx_rr2_lut64_p2_div()
86 …__m128i vl_hl = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_math_f32_sigmoid__avx_rr2_lut64_p2_div()
87 …__m128i vl_hh = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_math_f32_sigmoid__avx_rr2_lut64_p2_div()
88 …vl_ll = _mm_insert_epi32(vl_ll, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint3… in xnn_math_f32_sigmoid__avx_rr2_lut64_p2_div()
89 …vl_lh = _mm_insert_epi32(vl_lh, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint3… in xnn_math_f32_sigmoid__avx_rr2_lut64_p2_div()
90 …vl_hl = _mm_insert_epi32(vl_hl, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint3… in xnn_math_f32_sigmoid__avx_rr2_lut64_p2_div()
91 …vl_hh = _mm_insert_epi32(vl_hh, *((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uint3… in xnn_math_f32_sigmoid__avx_rr2_lut64_p2_div()
93 …__m128i vl_ll = _mm_cvtsi32_si128(*((const int*) ((uintptr_t) xnn_table_exp2minus_k_over_64 + (uin… in xnn_math_f32_sigmoid__avx_rr2_lut64_p2_div()
[all …]

1234