| /external/vixl/src/aarch64/ |
| D | assembler-aarch64.cc | 315 const VRegister& vn, in NEONTable()
327 const VRegister& vn, in tbl()
335 const VRegister& vn, in tbl()
347 const VRegister& vn, in tbl()
360 const VRegister& vn, in tbl()
374 const VRegister& vn, in tbx()
382 const VRegister& vn, in tbx()
394 const VRegister& vn, in tbx()
407 const VRegister& vn, in tbx()
2381 const VRegister& vn, in NEON3DifferentL()
[all …]
|
| /external/XNNPACK/src/f32-velu/ |
| D | sse-rr2-p6.c.in | 118 __m128 vn = _mm_add_ps(_mm_mul_ps(vz, vlog2e), vmagic_bias); variable
152 __m128 vn = _mm_add_ps(_mm_mul_ps(vz, vlog2e), vmagic_bias); variable
|
| /external/XNNPACK/src/f32-sigmoid/ |
| D | wasmsimd-p5-div.c.in | 103 v128_t vn = wasm_f32x4_add(vmagic_bias, wasm_f32x4_mul(vz, vminus_log2e)); variable
131 v128_t vn = wasm_f32x4_add(vmagic_bias, wasm_f32x4_mul(vz, vminus_log2e)); variable
|
| D | sse-p5-div.c.in | 113 __m128 vn = _mm_add_ps(_mm_mul_ps(vz, vlog2e), vmagic_bias); variable
148 __m128 vn = _mm_add_ps(_mm_mul_ps(vz, vlog2e), vmagic_bias); variable
|
| D | scalar-p5-div.c.in | 105 float vn = vz * vminus_log2e + vmagic_bias; variable
139 float vn = vz * vminus_log2e + vmagic_bias; variable
172 float vn = vz * vminus_log2e + vmagic_bias; variable
|
| D | scalar-lut2048-p1-div.c.in | 97 float vn = vz * vminus_log2e + vmagic_bias; variable
128 float vn = vz * vminus_log2e + vmagic_bias; variable
158 float vn = vz * vminus_log2e + vmagic_bias; variable
|
| D | scalar-lut64-p2-div.c.in | 100 float vn = vz * vminus_log2e + vmagic_bias; variable
133 float vn = vz * vminus_log2e + vmagic_bias; variable
165 float vn = vz * vminus_log2e + vmagic_bias; variable
|
| /external/XNNPACK/src/f32-sigmoid/gen/ |
| D | avx512f-rr1-p5-scalef-div-x16.c | 43 const __m512 vn = _mm512_roundscale_ps(_mm512_mul_ps(vz, vlog2e), 0); in xnn_f32_sigmoid_ukernel__avx512f_rr1_p5_scalef_div_x16() local
74 const __m512 vn = _mm512_roundscale_ps(_mm512_mul_ps(vz, vlog2e), 0); in xnn_f32_sigmoid_ukernel__avx512f_rr1_p5_scalef_div_x16() local
|
| D | avx512f-rr1-lut16-p3-perm-scalef-div-x16.c | 46 __m512 vn = _mm512_fmadd_ps(vz, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx512f_rr1_lut16_p3_perm_scalef_div_x16() local
78 __m512 vn = _mm512_fmadd_ps(vz, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx512f_rr1_lut16_p3_perm_scalef_div_x16() local
|
| D | avx512f-rr1-p5-scalef-nr1fma-x16.c | 43 const __m512 vn = _mm512_roundscale_ps(_mm512_mul_ps(vz, vlog2e), 0); in xnn_f32_sigmoid_ukernel__avx512f_rr1_p5_scalef_nr1fma_x16() local
77 const __m512 vn = _mm512_roundscale_ps(_mm512_mul_ps(vz, vlog2e), 0); in xnn_f32_sigmoid_ukernel__avx512f_rr1_p5_scalef_nr1fma_x16() local
|
| D | wasmsimd-p5-div-x4.c | 44 v128_t vn = wasm_f32x4_add(vmagic_bias, wasm_f32x4_mul(vz, vminus_log2e)); in xnn_f32_sigmoid_ukernel__wasmsimd_p5_div_x4() local
72 v128_t vn = wasm_f32x4_add(vmagic_bias, wasm_f32x4_mul(vz, vminus_log2e)); in xnn_f32_sigmoid_ukernel__wasmsimd_p5_div_x4() local
|
| D | avx512f-rr1-lut16-p3-perm-scalef-nr1fma-x16.c | 46 __m512 vn = _mm512_fmadd_ps(vz, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx512f_rr1_lut16_p3_perm_scalef_nr1fma_x16() local
81 __m512 vn = _mm512_fmadd_ps(vz, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx512f_rr1_lut16_p3_perm_scalef_nr1fma_x16() local
|
| D | avx512f-rr2-lut32-p2-perm2-scalef-div-x16.c | 52 __m512 vn = _mm512_fmadd_ps(vz, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx512f_rr2_lut32_p2_perm2_scalef_div_x16() local
84 __m512 vn = _mm512_fmadd_ps(vz, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx512f_rr2_lut32_p2_perm2_scalef_div_x16() local
|
| D | sse41-p5-div-x4.c | 44 __m128 vn = _mm_add_ps(_mm_mul_ps(vz, vlog2e), vmagic_bias); in xnn_f32_sigmoid_ukernel__sse41_p5_div_x4() local
75 __m128 vn = _mm_add_ps(_mm_mul_ps(vz, vlog2e), vmagic_bias); in xnn_f32_sigmoid_ukernel__sse41_p5_div_x4() local
|
| D | avx512f-rr2-lut32-p2-perm2-scalef-nr1fma-x16.c | 52 __m512 vn = _mm512_fmadd_ps(vz, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx512f_rr2_lut32_p2_perm2_scalef_nr1fma_x16() local
87 __m512 vn = _mm512_fmadd_ps(vz, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx512f_rr2_lut32_p2_perm2_scalef_nr1fma_x16() local
|
| D | neonfma-rr1-p5-div-x4.c | 42 float32x4_t vn = vfmaq_f32(vmagic_bias, vz, vminus_log2e); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_div_x4() local
68 float32x4_t vn = vfmaq_f32(vmagic_bias, vz, vminus_log2e); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_div_x4() local
|
| D | avx2-rr1-p5-div-x8.c | 46 __m256 vn = _mm256_fmadd_ps(vz, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x8() local
78 __m256 vn = _mm256_fmadd_ps(vz, vlog2e, vmagic_bias); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x8() local
|
| /external/XNNPACK/src/f32-velu/gen/ |
| D | velu-avx512f-rr1-p6-x16.c | 49 __m512 vn = _mm512_fmadd_ps(vz, vlog2e, vmagic_bias); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x16() local
83 __m512 vn = _mm512_fmadd_ps(vz, vlog2e, vmagic_bias); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x16() local
|
| D | velu-sse41-rr2-p6-x4.c | 51 __m128 vn = _mm_add_ps(_mm_mul_ps(vz, vlog2e), vmagic_bias); in xnn_f32_velu_ukernel__sse41_rr2_p6_x4() local
80 __m128 vn = _mm_add_ps(_mm_mul_ps(vz, vlog2e), vmagic_bias); in xnn_f32_velu_ukernel__sse41_rr2_p6_x4() local
|
| D | velu-avx512f-rr1-lut16-p3-perm-x16.c | 49 __m512 vn = _mm512_fmadd_ps(vz, vlog2e, vmagic_bias); in xnn_f32_velu_ukernel__avx512f_rr1_lut16_p3_perm_x16() local
82 __m512 vn = _mm512_fmadd_ps(vz, vlog2e, vmagic_bias); in xnn_f32_velu_ukernel__avx512f_rr1_lut16_p3_perm_x16() local
|
| D | velu-avx2-rr1-p6-x8.c | 48 __m256 vn = _mm256_fmadd_ps(vz, vlog2e, vmagic_bias); in xnn_f32_velu_ukernel__avx2_rr1_p6_x8() local
80 __m256 vn = _mm256_fmadd_ps(vz, vlog2e, vmagic_bias); in xnn_f32_velu_ukernel__avx2_rr1_p6_x8() local
|
| /external/XNNPACK/src/f32-raddstoreexpminusmax/gen/ |
| D | neonfma-p5-x4.c | 56 float32x4_t vn = vfmaq_f32(vmagic_bias, vx, vlog2e); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_p5_x4() local
113 float32x4_t vn = vfmaq_f32(vmagic_bias, vx, vlog2e); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_p5_x4() local
|
| D | neon-p5-x4.c | 57 float32x4_t vn = vmlaq_f32(vmagic_bias, vx, vlog2e); in xnn_f32_raddstoreexpminusmax_ukernel__neon_p5_x4() local
114 float32x4_t vn = vmlaq_f32(vmagic_bias, vx, vlog2e); in xnn_f32_raddstoreexpminusmax_ukernel__neon_p5_x4() local
|
| /external/XNNPACK/src/f32-vscaleextexp/gen/ |
| D | avx512f-p5-scalef-x16.c | 92 const __m512 vn = _mm512_roundscale_ps(_mm512_mul_ps(vx, vlog2e), 0); in xnn_f32_vscaleextexp_ukernel__avx512f_p5_scalef_x16() local
126 const __m512 vn = _mm512_roundscale_ps(_mm512_mul_ps(vx, vlog2e), 0); in xnn_f32_vscaleextexp_ukernel__avx512f_p5_scalef_x16() local
|
| /external/XNNPACK/src/f32-vscaleexpminusmax/gen/ |
| D | avx512f-p5-scalef-x16.c | 91 __m512 vn = _mm512_roundscale_ps(_mm512_mul_ps(vx, vlog2e), 0); in xnn_f32_vscaleexpminusmax_ukernel__avx512f_p5_scalef_x16() local
129 __m512 vn = _mm512_roundscale_ps(_mm512_mul_ps(vx, vlog2e), 0); in xnn_f32_vscaleexpminusmax_ukernel__avx512f_p5_scalef_x16() local
|