| /external/XNNPACK/src/f32-velu/ |
| D | sse-rr2-p6.c.in | 134 const __m128 ve = _mm_mul_ps(_mm_add_ps(vp, vs), valpha); variable
168 const __m128 ve = _mm_mul_ps(_mm_add_ps(vp, vs), valpha); variable
|
| /external/XNNPACK/src/f32-sigmoid/ |
| D | wasmsimd-p5-div.c.in | 116 const v128_t ve = wasm_f32x4_add(vs, wasm_f32x4_mul(vt, vp)); variable
144 const v128_t ve = wasm_f32x4_add(vs, wasm_f32x4_mul(vt, vp)); variable
|
| D | sse-p5-div.c.in | 126 __m128 ve = _mm_add_ps(_mm_mul_ps(vt, vp), vs); variable
161 __m128 ve = _mm_add_ps(_mm_mul_ps(vt, vp), vs); variable
|
| D | scalar-p5-div.c.in | 118 const float ve = vt * vp + vs; variable
152 const float ve = vt * vp + vs; variable
185 const float ve = vt * vp + vs; variable
|
| D | scalar-lut2048-p1-div.c.in | 98 const uint32_t ve = fp32_to_bits(vn) << 12; variable
129 const uint32_t ve = fp32_to_bits(vn) << 12; variable
159 const uint32_t ve = fp32_to_bits(vn) << 12; variable
|
| D | scalar-lut64-p2-div.c.in | 101 const uint32_t ve = fp32_to_bits(vn) << 17; variable
134 const uint32_t ve = fp32_to_bits(vn) << 17; variable
166 const uint32_t ve = fp32_to_bits(vn) << 17; variable
|
| /external/XNNPACK/src/f32-sigmoid/gen/ |
| D | avx512f-rr1-p5-scalef-div-x16.c | 53 const __m512 ve = _mm512_scalef_ps(vp, vn); in xnn_f32_sigmoid_ukernel__avx512f_rr1_p5_scalef_div_x16() local
84 const __m512 ve = _mm512_scalef_ps(vp, vn); in xnn_f32_sigmoid_ukernel__avx512f_rr1_p5_scalef_div_x16() local
|
| D | avx512f-rr1-lut16-p3-perm-scalef-div-x16.c | 57 const __m512 ve = _mm512_scalef_ps(vp, vn); in xnn_f32_sigmoid_ukernel__avx512f_rr1_lut16_p3_perm_scalef_div_x16() local
89 const __m512 ve = _mm512_scalef_ps(vp, vn); in xnn_f32_sigmoid_ukernel__avx512f_rr1_lut16_p3_perm_scalef_div_x16() local
|
| D | avx512f-rr1-p5-scalef-nr1fma-x16.c | 53 const __m512 ve = _mm512_scalef_ps(vp, vn); in xnn_f32_sigmoid_ukernel__avx512f_rr1_p5_scalef_nr1fma_x16() local
87 const __m512 ve = _mm512_scalef_ps(vp, vn); in xnn_f32_sigmoid_ukernel__avx512f_rr1_p5_scalef_nr1fma_x16() local
|
| D | wasmsimd-p5-div-x4.c | 57 const v128_t ve = wasm_f32x4_add(vs, wasm_f32x4_mul(vt, vp)); in xnn_f32_sigmoid_ukernel__wasmsimd_p5_div_x4() local
85 const v128_t ve = wasm_f32x4_add(vs, wasm_f32x4_mul(vt, vp)); in xnn_f32_sigmoid_ukernel__wasmsimd_p5_div_x4() local
|
| D | avx512f-rr1-lut16-p3-perm-scalef-nr1fma-x16.c | 57 const __m512 ve = _mm512_scalef_ps(vp, vn); in xnn_f32_sigmoid_ukernel__avx512f_rr1_lut16_p3_perm_scalef_nr1fma_x16() local
92 const __m512 ve = _mm512_scalef_ps(vp, vn); in xnn_f32_sigmoid_ukernel__avx512f_rr1_lut16_p3_perm_scalef_nr1fma_x16() local
|
| D | avx512f-rr2-lut32-p2-perm2-scalef-div-x16.c | 63 const __m512 ve = _mm512_scalef_ps(vp, vn); in xnn_f32_sigmoid_ukernel__avx512f_rr2_lut32_p2_perm2_scalef_div_x16() local
95 const __m512 ve = _mm512_scalef_ps(vp, vn); in xnn_f32_sigmoid_ukernel__avx512f_rr2_lut32_p2_perm2_scalef_div_x16() local
|
| D | sse41-p5-div-x4.c | 57 __m128 ve = _mm_add_ps(_mm_mul_ps(vt, vp), vs); in xnn_f32_sigmoid_ukernel__sse41_p5_div_x4() local
88 __m128 ve = _mm_add_ps(_mm_mul_ps(vt, vp), vs); in xnn_f32_sigmoid_ukernel__sse41_p5_div_x4() local
|
| D | avx512f-rr2-lut32-p2-perm2-scalef-nr1fma-x16.c | 63 const __m512 ve = _mm512_scalef_ps(vp, vn); in xnn_f32_sigmoid_ukernel__avx512f_rr2_lut32_p2_perm2_scalef_nr1fma_x16() local
98 const __m512 ve = _mm512_scalef_ps(vp, vn); in xnn_f32_sigmoid_ukernel__avx512f_rr2_lut32_p2_perm2_scalef_nr1fma_x16() local
|
| D | neonfma-rr1-p5-div-x4.c | 53 const float32x4_t ve = vfmaq_f32(vs, vp, vt); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_div_x4() local
79 const float32x4_t ve = vfmaq_f32(vs, vp, vt); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_div_x4() local
|
| D | avx2-rr1-p5-div-x8.c | 58 const __m256 ve = _mm256_fmadd_ps(vt, vp, vs); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x8() local
90 const __m256 ve = _mm256_fmadd_ps(vt, vp, vs); in xnn_f32_sigmoid_ukernel__avx2_rr1_p5_div_x8() local
|
| D | sse2-p5-div-x4.c | 57 __m128 ve = _mm_add_ps(_mm_mul_ps(vt, vp), vs); in xnn_f32_sigmoid_ukernel__sse2_p5_div_x4() local
89 __m128 ve = _mm_add_ps(_mm_mul_ps(vt, vp), vs); in xnn_f32_sigmoid_ukernel__sse2_p5_div_x4() local
|
| D | neonfma-rr1-p5-nr1recps1fma-x4.c | 53 const float32x4_t ve = vfmaq_f32(vs, vp, vt); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_nr1recps1fma_x4() local
83 const float32x4_t ve = vfmaq_f32(vs, vp, vt); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_nr1recps1fma_x4() local
|
| D | neonfma-rr1-p5-nr2recps-x4.c | 53 const float32x4_t ve = vfmaq_f32(vs, vp, vt); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_nr2recps_x4() local
83 const float32x4_t ve = vfmaq_f32(vs, vp, vt); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_nr2recps_x4() local
|
| /external/XNNPACK/src/f32-velu/gen/ |
| D | velu-sse41-rr2-p6-x4.c | 67 const __m128 ve = _mm_mul_ps(_mm_add_ps(vp, vs), valpha); in xnn_f32_velu_ukernel__sse41_rr2_p6_x4() local
96 const __m128 ve = _mm_mul_ps(_mm_add_ps(vp, vs), valpha); in xnn_f32_velu_ukernel__sse41_rr2_p6_x4() local
|
| D | velu-avx2-rr1-p6-x8.c | 63 const __m256 ve = _mm256_fmadd_ps(vp, valpha, vs); in xnn_f32_velu_ukernel__avx2_rr1_p6_x8() local
95 const __m256 ve = _mm256_fmadd_ps(vp, valpha, vs); in xnn_f32_velu_ukernel__avx2_rr1_p6_x8() local
|
| D | velu-neonfma-rr1-p6-x4.c | 64 const float32x4_t ve = vmulq_f32(vaddq_f32(vp, vs), valpha); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x4() local
92 const float32x4_t ve = vmulq_f32(vaddq_f32(vp, vs), valpha); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x4() local
|
| D | velu-wasmsimd-arm-rr2-p6-x4.c | 67 const v128_t ve = wasm_f32x4_mul(wasm_f32x4_add(vp, vs), valpha); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x4() local
97 const v128_t ve = wasm_f32x4_mul(wasm_f32x4_add(vp, vs), valpha); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x4() local
|
| D | velu-sse2-rr2-p6-x4.c | 67 const __m128 ve = _mm_mul_ps(_mm_add_ps(vp, vs), valpha); in xnn_f32_velu_ukernel__sse2_rr2_p6_x4() local
97 const __m128 ve = _mm_mul_ps(_mm_add_ps(vp, vs), valpha); in xnn_f32_velu_ukernel__sse2_rr2_p6_x4() local
|
| /external/XNNPACK/src/f32-vscaleextexp/gen/ |
| D | avx512f-p5-scalef-x16.c | 108 const __m512 ve = _mm512_add_ps(vn, vscalee); in xnn_f32_vscaleextexp_ukernel__avx512f_p5_scalef_x16() local
142 const __m512 ve = _mm512_add_ps(vn, vscalee); in xnn_f32_vscaleextexp_ukernel__avx512f_p5_scalef_x16() local
|