| /external/XNNPACK/src/f32-velu/gen/ |
| D | velu-wasm-rr2-p6-x6.c | 36 const float vc6 = 0x1.6b7338p-10f; in xnn_f32_velu_ukernel__wasm_rr2_p6_x6() local
94 float vp0 = vc6 * vt0 + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x6()
95 float vp1 = vc6 * vt1 + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x6()
96 float vp2 = vc6 * vt2 + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x6()
97 float vp3 = vc6 * vt3 + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x6()
98 float vp4 = vc6 * vt4 + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x6()
99 float vp5 = vc6 * vt5 + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x6()
191 float vp = vc6 * vt + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x6()
|
| D | velu-wasm-rr2-p6-x5.c | 36 const float vc6 = 0x1.6b7338p-10f; in xnn_f32_velu_ukernel__wasm_rr2_p6_x5() local
87 float vp0 = vc6 * vt0 + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x5()
88 float vp1 = vc6 * vt1 + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x5()
89 float vp2 = vc6 * vt2 + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x5()
90 float vp3 = vc6 * vt3 + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x5()
91 float vp4 = vc6 * vt4 + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x5()
172 float vp = vc6 * vt + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x5()
|
| D | velu-scalar-rr2-p6-x6.c | 36 const float vc6 = 0x1.6b7338p-10f; in xnn_f32_velu_ukernel__scalar_rr2_p6_x6() local
118 float vp0 = vc6 * vt0 + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x6()
119 float vp1 = vc6 * vt1 + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x6()
120 float vp2 = vc6 * vt2 + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x6()
121 float vp3 = vc6 * vt3 + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x6()
122 float vp4 = vc6 * vt4 + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x6()
123 float vp5 = vc6 * vt5 + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x6()
231 float vp = vc6 * vt + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x6()
|
| D | velu-wasm-rr2-p6-x4.c | 36 const float vc6 = 0x1.6b7338p-10f; in xnn_f32_velu_ukernel__wasm_rr2_p6_x4() local
80 float vp0 = vc6 * vt0 + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x4()
81 float vp1 = vc6 * vt1 + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x4()
82 float vp2 = vc6 * vt2 + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x4()
83 float vp3 = vc6 * vt3 + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x4()
153 float vp = vc6 * vt + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x4()
|
| D | velu-scalar-rr2-p6-x5.c | 36 const float vc6 = 0x1.6b7338p-10f; in xnn_f32_velu_ukernel__scalar_rr2_p6_x5() local
107 float vp0 = vc6 * vt0 + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x5()
108 float vp1 = vc6 * vt1 + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x5()
109 float vp2 = vc6 * vt2 + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x5()
110 float vp3 = vc6 * vt3 + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x5()
111 float vp4 = vc6 * vt4 + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x5()
206 float vp = vc6 * vt + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x5()
|
| D | velu-scalar-rr2-p6-x4.c | 36 const float vc6 = 0x1.6b7338p-10f; in xnn_f32_velu_ukernel__scalar_rr2_p6_x4() local
96 float vp0 = vc6 * vt0 + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x4()
97 float vp1 = vc6 * vt1 + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x4()
98 float vp2 = vc6 * vt2 + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x4()
99 float vp3 = vc6 * vt3 + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x4()
181 float vp = vc6 * vt + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x4()
|
| D | velu-scalar-rr2-p6-x3.c | 36 const float vc6 = 0x1.6b7338p-10f; in xnn_f32_velu_ukernel__scalar_rr2_p6_x3() local
85 float vp0 = vc6 * vt0 + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x3()
86 float vp1 = vc6 * vt1 + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x3()
87 float vp2 = vc6 * vt2 + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x3()
156 float vp = vc6 * vt + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x3()
|
| D | velu-wasm-rr2-p6-x3.c | 36 const float vc6 = 0x1.6b7338p-10f; in xnn_f32_velu_ukernel__wasm_rr2_p6_x3() local
73 float vp0 = vc6 * vt0 + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x3()
74 float vp1 = vc6 * vt1 + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x3()
75 float vp2 = vc6 * vt2 + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x3()
134 float vp = vc6 * vt + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x3()
|
| D | velu-avx2-rr1-p6-x72.c | 36 const __m256 vc6 = _mm256_set1_ps(0x1.6b7338p-10f); in xnn_f32_velu_ukernel__avx2_rr1_p6_x72() local
103 __m256 vp0 = _mm256_fmadd_ps(vc6, vt0, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x72()
104 __m256 vp1 = _mm256_fmadd_ps(vc6, vt1, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x72()
105 __m256 vp2 = _mm256_fmadd_ps(vc6, vt2, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x72()
106 __m256 vp3 = _mm256_fmadd_ps(vc6, vt3, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x72()
107 __m256 vp4 = _mm256_fmadd_ps(vc6, vt4, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x72()
108 __m256 vp5 = _mm256_fmadd_ps(vc6, vt5, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x72()
109 __m256 vp6 = _mm256_fmadd_ps(vc6, vt6, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x72()
110 __m256 vp7 = _mm256_fmadd_ps(vc6, vt7, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x72()
111 __m256 vp8 = _mm256_fmadd_ps(vc6, vt8, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x72()
[all …]
|
| D | velu-avx2-rr1-p6-x80.c | 36 const __m256 vc6 = _mm256_set1_ps(0x1.6b7338p-10f); in xnn_f32_velu_ukernel__avx2_rr1_p6_x80() local
109 __m256 vp0 = _mm256_fmadd_ps(vc6, vt0, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x80()
110 __m256 vp1 = _mm256_fmadd_ps(vc6, vt1, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x80()
111 __m256 vp2 = _mm256_fmadd_ps(vc6, vt2, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x80()
112 __m256 vp3 = _mm256_fmadd_ps(vc6, vt3, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x80()
113 __m256 vp4 = _mm256_fmadd_ps(vc6, vt4, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x80()
114 __m256 vp5 = _mm256_fmadd_ps(vc6, vt5, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x80()
115 __m256 vp6 = _mm256_fmadd_ps(vc6, vt6, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x80()
116 __m256 vp7 = _mm256_fmadd_ps(vc6, vt7, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x80()
117 __m256 vp8 = _mm256_fmadd_ps(vc6, vt8, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x80()
[all …]
|
| D | velu-avx2-rr1-p6-x64.c | 36 const __m256 vc6 = _mm256_set1_ps(0x1.6b7338p-10f); in xnn_f32_velu_ukernel__avx2_rr1_p6_x64() local
97 __m256 vp0 = _mm256_fmadd_ps(vc6, vt0, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x64()
98 __m256 vp1 = _mm256_fmadd_ps(vc6, vt1, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x64()
99 __m256 vp2 = _mm256_fmadd_ps(vc6, vt2, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x64()
100 __m256 vp3 = _mm256_fmadd_ps(vc6, vt3, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x64()
101 __m256 vp4 = _mm256_fmadd_ps(vc6, vt4, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x64()
102 __m256 vp5 = _mm256_fmadd_ps(vc6, vt5, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x64()
103 __m256 vp6 = _mm256_fmadd_ps(vc6, vt6, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x64()
104 __m256 vp7 = _mm256_fmadd_ps(vc6, vt7, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x64()
215 __m256 vp = _mm256_fmadd_ps(vc6, vt, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x64()
[all …]
|
| D | velu-avx512f-rr1-p6-x128.c | 36 const __m512 vc6 = _mm512_set1_ps(0x1.6b7338p-10f); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x128() local
97 __m512 vp0 = _mm512_fmadd_ps(vc6, vt0, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x128()
98 __m512 vp1 = _mm512_fmadd_ps(vc6, vt1, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x128()
99 __m512 vp2 = _mm512_fmadd_ps(vc6, vt2, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x128()
100 __m512 vp3 = _mm512_fmadd_ps(vc6, vt3, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x128()
101 __m512 vp4 = _mm512_fmadd_ps(vc6, vt4, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x128()
102 __m512 vp5 = _mm512_fmadd_ps(vc6, vt5, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x128()
103 __m512 vp6 = _mm512_fmadd_ps(vc6, vt6, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x128()
104 __m512 vp7 = _mm512_fmadd_ps(vc6, vt7, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x128()
218 __m512 vp = _mm512_fmadd_ps(vc6, vt, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x128()
[all …]
|
| D | velu-avx512f-rr1-p6-x96.c | 36 const __m512 vc6 = _mm512_set1_ps(0x1.6b7338p-10f); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x96() local
85 __m512 vp0 = _mm512_fmadd_ps(vc6, vt0, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x96()
86 __m512 vp1 = _mm512_fmadd_ps(vc6, vt1, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x96()
87 __m512 vp2 = _mm512_fmadd_ps(vc6, vt2, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x96()
88 __m512 vp3 = _mm512_fmadd_ps(vc6, vt3, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x96()
89 __m512 vp4 = _mm512_fmadd_ps(vc6, vt4, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x96()
90 __m512 vp5 = _mm512_fmadd_ps(vc6, vt5, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x96()
182 __m512 vp = _mm512_fmadd_ps(vc6, vt, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x96()
216 __m512 vp = _mm512_fmadd_ps(vc6, vt, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x96()
|
| D | velu-avx2-rr1-p6-x48.c | 36 const __m256 vc6 = _mm256_set1_ps(0x1.6b7338p-10f); in xnn_f32_velu_ukernel__avx2_rr1_p6_x48() local
85 __m256 vp0 = _mm256_fmadd_ps(vc6, vt0, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x48()
86 __m256 vp1 = _mm256_fmadd_ps(vc6, vt1, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x48()
87 __m256 vp2 = _mm256_fmadd_ps(vc6, vt2, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x48()
88 __m256 vp3 = _mm256_fmadd_ps(vc6, vt3, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x48()
89 __m256 vp4 = _mm256_fmadd_ps(vc6, vt4, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x48()
90 __m256 vp5 = _mm256_fmadd_ps(vc6, vt5, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x48()
179 __m256 vp = _mm256_fmadd_ps(vc6, vt, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x48()
211 __m256 vp = _mm256_fmadd_ps(vc6, vt, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x48()
|
| D | velu-avx512f-rr1-p6-x112.c | 36 const __m512 vc6 = _mm512_set1_ps(0x1.6b7338p-10f); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x112() local
91 __m512 vp0 = _mm512_fmadd_ps(vc6, vt0, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x112()
92 __m512 vp1 = _mm512_fmadd_ps(vc6, vt1, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x112()
93 __m512 vp2 = _mm512_fmadd_ps(vc6, vt2, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x112()
94 __m512 vp3 = _mm512_fmadd_ps(vc6, vt3, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x112()
95 __m512 vp4 = _mm512_fmadd_ps(vc6, vt4, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x112()
96 __m512 vp5 = _mm512_fmadd_ps(vc6, vt5, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x112()
97 __m512 vp6 = _mm512_fmadd_ps(vc6, vt6, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x112()
200 __m512 vp = _mm512_fmadd_ps(vc6, vt, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x112()
234 __m512 vp = _mm512_fmadd_ps(vc6, vt, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x112()
|
| D | velu-avx2-rr1-p6-x56.c | 36 const __m256 vc6 = _mm256_set1_ps(0x1.6b7338p-10f); in xnn_f32_velu_ukernel__avx2_rr1_p6_x56() local
91 __m256 vp0 = _mm256_fmadd_ps(vc6, vt0, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x56()
92 __m256 vp1 = _mm256_fmadd_ps(vc6, vt1, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x56()
93 __m256 vp2 = _mm256_fmadd_ps(vc6, vt2, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x56()
94 __m256 vp3 = _mm256_fmadd_ps(vc6, vt3, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x56()
95 __m256 vp4 = _mm256_fmadd_ps(vc6, vt4, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x56()
96 __m256 vp5 = _mm256_fmadd_ps(vc6, vt5, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x56()
97 __m256 vp6 = _mm256_fmadd_ps(vc6, vt6, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x56()
197 __m256 vp = _mm256_fmadd_ps(vc6, vt, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x56()
229 __m256 vp = _mm256_fmadd_ps(vc6, vt, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x56()
|
| D | velu-wasm-rr2-p6-x2.c | 36 const float vc6 = 0x1.6b7338p-10f; in xnn_f32_velu_ukernel__wasm_rr2_p6_x2() local
66 float vp0 = vc6 * vt0 + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x2()
67 float vp1 = vc6 * vt1 + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x2()
114 float vp = vc6 * vt + vc5; in xnn_f32_velu_ukernel__wasm_rr2_p6_x2()
|
| D | velu-scalar-rr2-p6-x2.c | 36 const float vc6 = 0x1.6b7338p-10f; in xnn_f32_velu_ukernel__scalar_rr2_p6_x2() local
74 float vp0 = vc6 * vt0 + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x2()
75 float vp1 = vc6 * vt1 + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x2()
130 float vp = vc6 * vt + vc5; in xnn_f32_velu_ukernel__scalar_rr2_p6_x2()
|
| D | velu-avx2-rr1-p6-x32.c | 36 const __m256 vc6 = _mm256_set1_ps(0x1.6b7338p-10f); in xnn_f32_velu_ukernel__avx2_rr1_p6_x32() local
73 __m256 vp0 = _mm256_fmadd_ps(vc6, vt0, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x32()
74 __m256 vp1 = _mm256_fmadd_ps(vc6, vt1, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x32()
75 __m256 vp2 = _mm256_fmadd_ps(vc6, vt2, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x32()
76 __m256 vp3 = _mm256_fmadd_ps(vc6, vt3, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x32()
143 __m256 vp = _mm256_fmadd_ps(vc6, vt, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x32()
175 __m256 vp = _mm256_fmadd_ps(vc6, vt, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x32()
|
| D | velu-avx512f-rr1-p6-x64.c | 36 const __m512 vc6 = _mm512_set1_ps(0x1.6b7338p-10f); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x64() local
73 __m512 vp0 = _mm512_fmadd_ps(vc6, vt0, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x64()
74 __m512 vp1 = _mm512_fmadd_ps(vc6, vt1, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x64()
75 __m512 vp2 = _mm512_fmadd_ps(vc6, vt2, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x64()
76 __m512 vp3 = _mm512_fmadd_ps(vc6, vt3, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x64()
146 __m512 vp = _mm512_fmadd_ps(vc6, vt, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x64()
180 __m512 vp = _mm512_fmadd_ps(vc6, vt, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x64()
|
| D | velu-neonfma-rr1-p6-x20.c | 37 const float32x4_t vc6 = vmovq_n_f32(0x1.6b7338p-10f); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x20() local
80 float32x4_t vp0123 = vfmaq_f32(vc5, vc6, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x20()
81 float32x4_t vp4567 = vfmaq_f32(vc5, vc6, vt4567); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x20()
82 float32x4_t vp89AB = vfmaq_f32(vc5, vc6, vt89AB); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x20()
83 float32x4_t vpCDEF = vfmaq_f32(vc5, vc6, vtCDEF); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x20()
84 float32x4_t vpGHIJ = vfmaq_f32(vc5, vc6, vtGHIJ); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x20()
167 float32x4_t vp = vfmaq_f32(vc5, vc6, vt); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x20()
195 float32x4_t vp = vfmaq_f32(vc5, vc6, vt); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x20()
|
| D | velu-avx512f-rr1-p6-x80.c | 36 const __m512 vc6 = _mm512_set1_ps(0x1.6b7338p-10f); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x80() local
79 __m512 vp0 = _mm512_fmadd_ps(vc6, vt0, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x80()
80 __m512 vp1 = _mm512_fmadd_ps(vc6, vt1, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x80()
81 __m512 vp2 = _mm512_fmadd_ps(vc6, vt2, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x80()
82 __m512 vp3 = _mm512_fmadd_ps(vc6, vt3, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x80()
83 __m512 vp4 = _mm512_fmadd_ps(vc6, vt4, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x80()
164 __m512 vp = _mm512_fmadd_ps(vc6, vt, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x80()
198 __m512 vp = _mm512_fmadd_ps(vc6, vt, vc5); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x80()
|
| D | velu-avx2-rr1-p6-x40.c | 36 const __m256 vc6 = _mm256_set1_ps(0x1.6b7338p-10f); in xnn_f32_velu_ukernel__avx2_rr1_p6_x40() local
79 __m256 vp0 = _mm256_fmadd_ps(vc6, vt0, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x40()
80 __m256 vp1 = _mm256_fmadd_ps(vc6, vt1, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x40()
81 __m256 vp2 = _mm256_fmadd_ps(vc6, vt2, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x40()
82 __m256 vp3 = _mm256_fmadd_ps(vc6, vt3, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x40()
83 __m256 vp4 = _mm256_fmadd_ps(vc6, vt4, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x40()
161 __m256 vp = _mm256_fmadd_ps(vc6, vt, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x40()
193 __m256 vp = _mm256_fmadd_ps(vc6, vt, vc5); in xnn_f32_velu_ukernel__avx2_rr1_p6_x40()
|
| D | velu-neon-rr2-p6-x24.c | 38 const float32x4_t vc6 = vmovq_n_f32(0x1.6b7338p-10f); in xnn_f32_velu_ukernel__neon_rr2_p6_x24() local
94 float32x4_t vp0123 = vmlaq_f32(vc5, vc6, vt0123); in xnn_f32_velu_ukernel__neon_rr2_p6_x24()
95 float32x4_t vp4567 = vmlaq_f32(vc5, vc6, vt4567); in xnn_f32_velu_ukernel__neon_rr2_p6_x24()
96 float32x4_t vp89AB = vmlaq_f32(vc5, vc6, vt89AB); in xnn_f32_velu_ukernel__neon_rr2_p6_x24()
97 float32x4_t vpCDEF = vmlaq_f32(vc5, vc6, vtCDEF); in xnn_f32_velu_ukernel__neon_rr2_p6_x24()
98 float32x4_t vpGHIJ = vmlaq_f32(vc5, vc6, vtGHIJ); in xnn_f32_velu_ukernel__neon_rr2_p6_x24()
99 float32x4_t vpKLMN = vmlaq_f32(vc5, vc6, vtKLMN); in xnn_f32_velu_ukernel__neon_rr2_p6_x24()
195 float32x4_t vp = vmlaq_f32(vc5, vc6, vt); in xnn_f32_velu_ukernel__neon_rr2_p6_x24()
224 float32x4_t vp = vmlaq_f32(vc5, vc6, vt); in xnn_f32_velu_ukernel__neon_rr2_p6_x24()
|
| D | velu-neonfma-rr1-p6-x24.c | 37 const float32x4_t vc6 = vmovq_n_f32(0x1.6b7338p-10f); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x24() local
86 float32x4_t vp0123 = vfmaq_f32(vc5, vc6, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x24()
87 float32x4_t vp4567 = vfmaq_f32(vc5, vc6, vt4567); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x24()
88 float32x4_t vp89AB = vfmaq_f32(vc5, vc6, vt89AB); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x24()
89 float32x4_t vpCDEF = vfmaq_f32(vc5, vc6, vtCDEF); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x24()
90 float32x4_t vpGHIJ = vfmaq_f32(vc5, vc6, vtGHIJ); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x24()
91 float32x4_t vpKLMN = vfmaq_f32(vc5, vc6, vtKLMN); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x24()
186 float32x4_t vp = vfmaq_f32(vc5, vc6, vt); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x24()
214 float32x4_t vp = vfmaq_f32(vc5, vc6, vt); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x24()
|