Home
last modified time | relevance | path

Searched refs:vz0 (Results 1 – 25 of 34) sorted by relevance

12

/external/XNNPACK/src/f32-sigmoid/gen/
Dscalar-lut2048-p1-div-x2.c56 const float vz0 = fabsf(vx0); in xnn_f32_sigmoid_ukernel__scalar_lut2048_p1_div_x2() local
66 float vn0 = vz0 * vminus_log2e_x2048 + vmagic_bias; in xnn_f32_sigmoid_ukernel__scalar_lut2048_p1_div_x2()
95 float vt0 = vn0 * vln2_o2048_hi + vz0; in xnn_f32_sigmoid_ukernel__scalar_lut2048_p1_div_x2()
119 if XNN_UNPREDICTABLE(vz0 > vdenorm_cutoff) { in xnn_f32_sigmoid_ukernel__scalar_lut2048_p1_div_x2()
Dscalar-p5-div-x2.c55 const float vz0 = fabsf(vx0); in xnn_f32_sigmoid_ukernel__scalar_p5_div_x2() local
64 float vn0 = vz0 * vminus_log2e + vmagic_bias; in xnn_f32_sigmoid_ukernel__scalar_p5_div_x2()
78 float vt0 = vn0 * vln2_hi + vz0; in xnn_f32_sigmoid_ukernel__scalar_p5_div_x2()
114 if XNN_UNPREDICTABLE(vz0 > vdenorm_cutoff) { in xnn_f32_sigmoid_ukernel__scalar_p5_div_x2()
Dscalar-lut64-p2-div-x2.c56 const float vz0 = fabsf(vx0); in xnn_f32_sigmoid_ukernel__scalar_lut64_p2_div_x2() local
66 float vn0 = vz0 * vminus_log2e_x64 + vmagic_bias; in xnn_f32_sigmoid_ukernel__scalar_lut64_p2_div_x2()
95 float vt0 = vn0 * vln2_o64_hi + vz0; in xnn_f32_sigmoid_ukernel__scalar_lut64_p2_div_x2()
123 if XNN_UNPREDICTABLE(vz0 > vdenorm_cutoff) { in xnn_f32_sigmoid_ukernel__scalar_lut64_p2_div_x2()
Dscalar-lut2048-p1-div-x4.c58 const float vz0 = fabsf(vx0); in xnn_f32_sigmoid_ukernel__scalar_lut2048_p1_div_x4() local
70 float vn0 = vz0 * vminus_log2e_x2048 + vmagic_bias; in xnn_f32_sigmoid_ukernel__scalar_lut2048_p1_div_x4()
109 float vt0 = vn0 * vln2_o2048_hi + vz0; in xnn_f32_sigmoid_ukernel__scalar_lut2048_p1_div_x4()
143 if XNN_UNPREDICTABLE(vz0 > vdenorm_cutoff) { in xnn_f32_sigmoid_ukernel__scalar_lut2048_p1_div_x4()
Dscalar-p5-div-x4.c57 const float vz0 = fabsf(vx0); in xnn_f32_sigmoid_ukernel__scalar_p5_div_x4() local
68 float vn0 = vz0 * vminus_log2e + vmagic_bias; in xnn_f32_sigmoid_ukernel__scalar_p5_div_x4()
88 float vt0 = vn0 * vln2_hi + vz0; in xnn_f32_sigmoid_ukernel__scalar_p5_div_x4()
142 if XNN_UNPREDICTABLE(vz0 > vdenorm_cutoff) { in xnn_f32_sigmoid_ukernel__scalar_p5_div_x4()
Dscalar-lut64-p2-div-x4.c58 const float vz0 = fabsf(vx0); in xnn_f32_sigmoid_ukernel__scalar_lut64_p2_div_x4() local
70 float vn0 = vz0 * vminus_log2e_x64 + vmagic_bias; in xnn_f32_sigmoid_ukernel__scalar_lut64_p2_div_x4()
109 float vt0 = vn0 * vln2_o64_hi + vz0; in xnn_f32_sigmoid_ukernel__scalar_lut64_p2_div_x4()
149 if XNN_UNPREDICTABLE(vz0 > vdenorm_cutoff) { in xnn_f32_sigmoid_ukernel__scalar_lut64_p2_div_x4()
Davx2-rr1-p5-div-x16.c55 const __m256 vz0 = _mm256_or_ps(vx0, vsign_mask); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x16() local
64 __m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x16()
77 __m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x16()
113 vf0 = _mm256_andnot_ps(_mm256_cmp_ps(vz0, vdenorm_cutoff, _CMP_LT_OS), vf0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x16()
Davx2-rr1-p5-nr1fma-x16.c55 const __m256 vz0 = _mm256_or_ps(vx0, vsign_mask); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x16() local
64 __m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x16()
77 __m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x16()
123 vf0 = _mm256_andnot_ps(_mm256_cmp_ps(vz0, vdenorm_cutoff, _CMP_LT_OS), vf0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x16()
Davx2-rr1-p5-div-x24.c56 const __m256 vz0 = _mm256_or_ps(vx0, vsign_mask); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x24() local
66 __m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x24()
82 __m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x24()
127 vf0 = _mm256_andnot_ps(_mm256_cmp_ps(vz0, vdenorm_cutoff, _CMP_LT_OS), vf0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x24()
Davx2-rr1-p5-nr2fma-x16.c55 const __m256 vz0 = _mm256_or_ps(vx0, vsign_mask); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr2fma_x16() local
64 __m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr2fma_x16()
77 __m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr2fma_x16()
125 vf0 = _mm256_andnot_ps(_mm256_cmp_ps(vz0, vdenorm_cutoff, _CMP_LT_OS), vf0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr2fma_x16()
Davx2-rr1-p5-nr1fma-x24.c56 const __m256 vz0 = _mm256_or_ps(vx0, vsign_mask); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x24() local
66 __m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x24()
82 __m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x24()
139 vf0 = _mm256_andnot_ps(_mm256_cmp_ps(vz0, vdenorm_cutoff, _CMP_LT_OS), vf0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x24()
Davx2-rr1-p5-div-x32.c57 const __m256 vz0 = _mm256_or_ps(vx0, vsign_mask); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x32() local
68 __m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x32()
87 __m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x32()
141 vf0 = _mm256_andnot_ps(_mm256_cmp_ps(vz0, vdenorm_cutoff, _CMP_LT_OS), vf0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x32()
Davx2-rr1-p5-nr2fma-x24.c56 const __m256 vz0 = _mm256_or_ps(vx0, vsign_mask); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr2fma_x24() local
66 __m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr2fma_x24()
82 __m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr2fma_x24()
142 vf0 = _mm256_andnot_ps(_mm256_cmp_ps(vz0, vdenorm_cutoff, _CMP_LT_OS), vf0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr2fma_x24()
Davx2-rr1-p5-div-x40.c58 const __m256 vz0 = _mm256_or_ps(vx0, vsign_mask); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x40() local
70 __m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x40()
92 __m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x40()
155 vf0 = _mm256_andnot_ps(_mm256_cmp_ps(vz0, vdenorm_cutoff, _CMP_LT_OS), vf0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x40()
Davx2-rr1-p5-nr1fma-x32.c57 const __m256 vz0 = _mm256_or_ps(vx0, vsign_mask); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x32() local
68 __m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x32()
87 __m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x32()
155 vf0 = _mm256_andnot_ps(_mm256_cmp_ps(vz0, vdenorm_cutoff, _CMP_LT_OS), vf0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x32()
Davx2-rr1-p5-nr2fma-x32.c57 const __m256 vz0 = _mm256_or_ps(vx0, vsign_mask); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr2fma_x32() local
68 __m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr2fma_x32()
87 __m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr2fma_x32()
159 vf0 = _mm256_andnot_ps(_mm256_cmp_ps(vz0, vdenorm_cutoff, _CMP_LT_OS), vf0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr2fma_x32()
Davx2-rr1-p5-nr2fma-x40.c58 const __m256 vz0 = _mm256_or_ps(vx0, vsign_mask); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr2fma_x40() local
70 __m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr2fma_x40()
92 __m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr2fma_x40()
176 vf0 = _mm256_andnot_ps(_mm256_cmp_ps(vz0, vdenorm_cutoff, _CMP_LT_OS), vf0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr2fma_x40()
Davx2-rr1-p5-nr1fma-x40.c58 const __m256 vz0 = _mm256_or_ps(vx0, vsign_mask); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x40() local
70 __m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x40()
92 __m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x40()
171 vf0 = _mm256_andnot_ps(_mm256_cmp_ps(vz0, vdenorm_cutoff, _CMP_LT_OS), vf0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x40()
Davx2-rr1-p5-div-x48.c59 const __m256 vz0 = _mm256_or_ps(vx0, vsign_mask); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x48() local
72 __m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x48()
97 __m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x48()
169 vf0 = _mm256_andnot_ps(_mm256_cmp_ps(vz0, vdenorm_cutoff, _CMP_LT_OS), vf0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x48()
Davx2-rr1-p5-nr1fma-x48.c59 const __m256 vz0 = _mm256_or_ps(vx0, vsign_mask); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x48() local
72 __m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x48()
97 __m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x48()
187 vf0 = _mm256_andnot_ps(_mm256_cmp_ps(vz0, vdenorm_cutoff, _CMP_LT_OS), vf0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x48()
Davx2-rr1-p5-div-x56.c60 const __m256 vz0 = _mm256_or_ps(vx0, vsign_mask); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x56() local
74 __m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x56()
102 __m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x56()
183 vf0 = _mm256_andnot_ps(_mm256_cmp_ps(vz0, vdenorm_cutoff, _CMP_LT_OS), vf0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x56()
Davx2-rr1-p5-nr2fma-x48.c59 const __m256 vz0 = _mm256_or_ps(vx0, vsign_mask); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr2fma_x48() local
72 __m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr2fma_x48()
97 __m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr2fma_x48()
193 vf0 = _mm256_andnot_ps(_mm256_cmp_ps(vz0, vdenorm_cutoff, _CMP_LT_OS), vf0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr2fma_x48()
Davx2-rr1-p5-div-x64.c61 const __m256 vz0 = _mm256_or_ps(vx0, vsign_mask); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x64() local
76 __m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x64()
107 __m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x64()
197 vf0 = _mm256_andnot_ps(_mm256_cmp_ps(vz0, vdenorm_cutoff, _CMP_LT_OS), vf0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x64()
Davx2-rr1-p5-nr1fma-x56.c60 const __m256 vz0 = _mm256_or_ps(vx0, vsign_mask); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x56() local
74 __m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x56()
102 __m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x56()
203 vf0 = _mm256_andnot_ps(_mm256_cmp_ps(vz0, vdenorm_cutoff, _CMP_LT_OS), vf0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_nr1fma_x56()
Davx2-rr1-p5-div-x80.c63 const __m256 vz0 = _mm256_or_ps(vx0, vsign_mask); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x80() local
80 __m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x80()
117 __m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x80()
225 vf0 = _mm256_andnot_ps(_mm256_cmp_ps(vz0, vdenorm_cutoff, _CMP_LT_OS), vf0); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x80()

12