/external/XNNPACK/src/f32-velu/gen/ |
D | velu-neon-rr2-p6-x24.c | 99 float32x4_t vpKLMN = vmlaq_f32(vc5, vc6, vtKLMN); in xnn_f32_velu_ukernel__neon_rr2_p6_x24() local 106 vpKLMN = vmlaq_f32(vc4, vpKLMN, vtKLMN); in xnn_f32_velu_ukernel__neon_rr2_p6_x24() 113 vpKLMN = vmlaq_f32(vc3, vpKLMN, vtKLMN); in xnn_f32_velu_ukernel__neon_rr2_p6_x24() 120 vpKLMN = vmlaq_f32(vc2, vpKLMN, vtKLMN); in xnn_f32_velu_ukernel__neon_rr2_p6_x24() 127 vpKLMN = vmulq_f32(vpKLMN, vtKLMN); in xnn_f32_velu_ukernel__neon_rr2_p6_x24() 147 vpKLMN = vmlaq_f32(vtKLMN, vpKLMN, vtKLMN); in xnn_f32_velu_ukernel__neon_rr2_p6_x24() 154 const float32x4_t veKLMN = vmulq_f32(vaddq_f32(vpKLMN, vsKLMN), valpha); in xnn_f32_velu_ukernel__neon_rr2_p6_x24()
|
D | velu-neonfma-rr1-p6-x24.c | 91 float32x4_t vpKLMN = vfmaq_f32(vc5, vc6, vtKLMN); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x24() local 98 vpKLMN = vfmaq_f32(vc4, vpKLMN, vtKLMN); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x24() 105 vpKLMN = vfmaq_f32(vc3, vpKLMN, vtKLMN); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x24() 112 vpKLMN = vfmaq_f32(vc2, vpKLMN, vtKLMN); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x24() 119 vpKLMN = vmulq_f32(vpKLMN, vtKLMN); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x24() 139 vpKLMN = vfmaq_f32(vtKLMN, vpKLMN, vtKLMN); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x24() 146 const float32x4_t veKLMN = vmulq_f32(vaddq_f32(vpKLMN, vsKLMN), valpha); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x24()
|
D | velu-sse41-rr2-p6-x24.c | 101 __m128 vpKLMN = _mm_add_ps(_mm_mul_ps(vc6, vtKLMN), vc5); in xnn_f32_velu_ukernel__sse41_rr2_p6_x24() local 108 vpKLMN = _mm_add_ps(_mm_mul_ps(vpKLMN, vtKLMN), vc4); in xnn_f32_velu_ukernel__sse41_rr2_p6_x24() 115 vpKLMN = _mm_add_ps(_mm_mul_ps(vpKLMN, vtKLMN), vc3); in xnn_f32_velu_ukernel__sse41_rr2_p6_x24() 122 vpKLMN = _mm_add_ps(_mm_mul_ps(vpKLMN, vtKLMN), vc2); in xnn_f32_velu_ukernel__sse41_rr2_p6_x24() 129 vpKLMN = _mm_mul_ps(vpKLMN, vtKLMN); in xnn_f32_velu_ukernel__sse41_rr2_p6_x24() 149 vpKLMN = _mm_add_ps(_mm_mul_ps(vpKLMN, vtKLMN), vtKLMN); in xnn_f32_velu_ukernel__sse41_rr2_p6_x24() 156 const __m128 veKLMN = _mm_mul_ps(_mm_add_ps(vpKLMN, vsKLMN), valpha); in xnn_f32_velu_ukernel__sse41_rr2_p6_x24()
|
D | velu-wasmsimd-arm-rr2-p6-x24.c | 101 v128_t vpKLMN = wasm_f32x4_add(wasm_f32x4_mul(vc6, vtKLMN), vc5); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x24() local 108 vpKLMN = wasm_f32x4_add(wasm_f32x4_mul(vpKLMN, vtKLMN), vc4); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x24() 115 vpKLMN = wasm_f32x4_add(wasm_f32x4_mul(vpKLMN, vtKLMN), vc3); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x24() 122 vpKLMN = wasm_f32x4_add(wasm_f32x4_mul(vpKLMN, vtKLMN), vc2); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x24() 129 vpKLMN = wasm_f32x4_mul(vpKLMN, vtKLMN); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x24() 149 vpKLMN = wasm_f32x4_add(wasm_f32x4_mul(vpKLMN, vtKLMN), vtKLMN); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x24() 156 const v128_t veKLMN = wasm_f32x4_mul(wasm_f32x4_add(vpKLMN, vsKLMN), valpha); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x24()
|
D | velu-wasmsimd-x86-rr2-p6-x24.c | 119 v128_t vpKLMN = wasm_f32x4_add(wasm_f32x4_mul(vc6, vtKLMN), vc5); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_p6_x24() local 126 vpKLMN = wasm_f32x4_add(wasm_f32x4_mul(vpKLMN, vtKLMN), vc4); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_p6_x24() 133 vpKLMN = wasm_f32x4_add(wasm_f32x4_mul(vpKLMN, vtKLMN), vc3); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_p6_x24() 140 vpKLMN = wasm_f32x4_add(wasm_f32x4_mul(vpKLMN, vtKLMN), vc2); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_p6_x24() 147 vpKLMN = wasm_f32x4_mul(vpKLMN, vtKLMN); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_p6_x24() 167 vpKLMN = wasm_f32x4_add(wasm_f32x4_mul(vpKLMN, vtKLMN), vtKLMN); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_p6_x24() 174 const v128_t veKLMN = wasm_f32x4_mul(wasm_f32x4_add(vpKLMN, vsKLMN), valpha); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_p6_x24()
|
D | velu-sse2-rr2-p6-x24.c | 101 __m128 vpKLMN = _mm_add_ps(_mm_mul_ps(vc6, vtKLMN), vc5); in xnn_f32_velu_ukernel__sse2_rr2_p6_x24() local 108 vpKLMN = _mm_add_ps(_mm_mul_ps(vpKLMN, vtKLMN), vc4); in xnn_f32_velu_ukernel__sse2_rr2_p6_x24() 115 vpKLMN = _mm_add_ps(_mm_mul_ps(vpKLMN, vtKLMN), vc3); in xnn_f32_velu_ukernel__sse2_rr2_p6_x24() 122 vpKLMN = _mm_add_ps(_mm_mul_ps(vpKLMN, vtKLMN), vc2); in xnn_f32_velu_ukernel__sse2_rr2_p6_x24() 129 vpKLMN = _mm_mul_ps(vpKLMN, vtKLMN); in xnn_f32_velu_ukernel__sse2_rr2_p6_x24() 149 vpKLMN = _mm_add_ps(_mm_mul_ps(vpKLMN, vtKLMN), vtKLMN); in xnn_f32_velu_ukernel__sse2_rr2_p6_x24() 156 const __m128 veKLMN = _mm_mul_ps(_mm_add_ps(vpKLMN, vsKLMN), valpha); in xnn_f32_velu_ukernel__sse2_rr2_p6_x24()
|
D | velu-wasmsimd-arm-rr2-lut16-p3-x24.c | 156 v128_t vpKLMN = wasm_f32x4_add(wasm_f32x4_mul(vc3, vtKLMN), vc2); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_lut16_p3_x24() local 163 vpKLMN = wasm_f32x4_mul(vpKLMN, vtKLMN); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_lut16_p3_x24() 183 vpKLMN = wasm_f32x4_add(wasm_f32x4_mul(vpKLMN, vtKLMN), vtKLMN); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_lut16_p3_x24() 190 const v128_t veKLMN = wasm_f32x4_mul(wasm_f32x4_add(vpKLMN, vsKLMN), valpha); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_lut16_p3_x24()
|
D | velu-neonfma-rr1-lut16-p3-x24.c | 147 float32x4_t vpKLMN = vfmaq_f32(vc2, vc3, vtKLMN); in xnn_f32_velu_ukernel__neonfma_rr1_lut16_p3_x24() local 154 vpKLMN = vmulq_f32(vpKLMN, vtKLMN); in xnn_f32_velu_ukernel__neonfma_rr1_lut16_p3_x24() 174 vpKLMN = vfmaq_f32(vtKLMN, vpKLMN, vtKLMN); in xnn_f32_velu_ukernel__neonfma_rr1_lut16_p3_x24() 181 const float32x4_t veKLMN = vmulq_f32(vaddq_f32(vpKLMN, vsKLMN), valpha); in xnn_f32_velu_ukernel__neonfma_rr1_lut16_p3_x24()
|
D | velu-wasmsimd-x86-rr2-lut16-p3-x24.c | 174 v128_t vpKLMN = wasm_f32x4_add(wasm_f32x4_mul(vc3, vtKLMN), vc2); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_lut16_p3_x24() local 181 vpKLMN = wasm_f32x4_mul(vpKLMN, vtKLMN); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_lut16_p3_x24() 201 vpKLMN = wasm_f32x4_add(wasm_f32x4_mul(vpKLMN, vtKLMN), vtKLMN); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_lut16_p3_x24() 208 const v128_t veKLMN = wasm_f32x4_mul(wasm_f32x4_add(vpKLMN, vsKLMN), valpha); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_lut16_p3_x24()
|
D | velu-neon-rr2-lut16-p3-x24.c | 155 float32x4_t vpKLMN = vmlaq_f32(vc2, vc3, vtKLMN); in xnn_f32_velu_ukernel__neon_rr2_lut16_p3_x24() local 162 vpKLMN = vmulq_f32(vpKLMN, vtKLMN); in xnn_f32_velu_ukernel__neon_rr2_lut16_p3_x24() 182 vpKLMN = vmlaq_f32(vtKLMN, vpKLMN, vtKLMN); in xnn_f32_velu_ukernel__neon_rr2_lut16_p3_x24() 189 const float32x4_t veKLMN = vmulq_f32(vaddq_f32(vpKLMN, vsKLMN), valpha); in xnn_f32_velu_ukernel__neon_rr2_lut16_p3_x24()
|
D | velu-sse41-rr2-lut16-p3-x24.c | 213 __m128 vpKLMN = _mm_add_ps(_mm_mul_ps(vc3, vtKLMN), vc2); in xnn_f32_velu_ukernel__sse41_rr2_lut16_p3_x24() local 220 vpKLMN = _mm_mul_ps(vpKLMN, vtKLMN); in xnn_f32_velu_ukernel__sse41_rr2_lut16_p3_x24() 240 vpKLMN = _mm_add_ps(_mm_mul_ps(vpKLMN, vtKLMN), vtKLMN); in xnn_f32_velu_ukernel__sse41_rr2_lut16_p3_x24() 247 const __m128 veKLMN = _mm_mul_ps(_mm_add_ps(vpKLMN, vsKLMN), valpha); in xnn_f32_velu_ukernel__sse41_rr2_lut16_p3_x24()
|
D | velu-sse2-rr2-lut16-p3-x24.c | 237 __m128 vpKLMN = _mm_add_ps(_mm_mul_ps(vc3, vtKLMN), vc2); in xnn_f32_velu_ukernel__sse2_rr2_lut16_p3_x24() local 244 vpKLMN = _mm_mul_ps(vpKLMN, vtKLMN); in xnn_f32_velu_ukernel__sse2_rr2_lut16_p3_x24() 264 vpKLMN = _mm_add_ps(_mm_mul_ps(vpKLMN, vtKLMN), vtKLMN); in xnn_f32_velu_ukernel__sse2_rr2_lut16_p3_x24() 271 const __m128 veKLMN = _mm_mul_ps(_mm_add_ps(vpKLMN, vsKLMN), valpha); in xnn_f32_velu_ukernel__sse2_rr2_lut16_p3_x24()
|
/external/XNNPACK/src/f32-sigmoid/gen/ |
D | neonfma-rr1-p5-div-x24.c | 85 float32x4_t vpKLMN = vfmaq_f32(vc4, vc5, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_div_x24() local 92 vpKLMN = vfmaq_f32(vc3, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_div_x24() 99 vpKLMN = vfmaq_f32(vc2, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_div_x24() 106 vpKLMN = vfmaq_f32(vc1, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_div_x24() 120 const float32x4_t veKLMN = vfmaq_f32(vsKLMN, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_div_x24()
|
D | wasmsimd-p5-div-x24.c | 94 v128_t vpKLMN = wasm_f32x4_add(vc4, wasm_f32x4_mul(vtKLMN, vc5)); in xnn_f32_sigmoid_ukernel__wasmsimd_p5_div_x24() local 101 vpKLMN = wasm_f32x4_add(vc3, wasm_f32x4_mul(vtKLMN, vpKLMN)); in xnn_f32_sigmoid_ukernel__wasmsimd_p5_div_x24() 108 vpKLMN = wasm_f32x4_add(vc2, wasm_f32x4_mul(vtKLMN, vpKLMN)); in xnn_f32_sigmoid_ukernel__wasmsimd_p5_div_x24() 115 vpKLMN = wasm_f32x4_add(vc1, wasm_f32x4_mul(vtKLMN, vpKLMN)); in xnn_f32_sigmoid_ukernel__wasmsimd_p5_div_x24() 129 const v128_t veKLMN = wasm_f32x4_add(vsKLMN, wasm_f32x4_mul(vtKLMN, vpKLMN)); in xnn_f32_sigmoid_ukernel__wasmsimd_p5_div_x24()
|
D | sse41-p5-div-x24.c | 94 __m128 vpKLMN = _mm_add_ps(_mm_mul_ps(vc5, vtKLMN), vc4); in xnn_f32_sigmoid_ukernel__sse41_p5_div_x24() local 101 vpKLMN = _mm_add_ps(_mm_mul_ps(vpKLMN, vtKLMN), vc3); in xnn_f32_sigmoid_ukernel__sse41_p5_div_x24() 108 vpKLMN = _mm_add_ps(_mm_mul_ps(vpKLMN, vtKLMN), vc2); in xnn_f32_sigmoid_ukernel__sse41_p5_div_x24() 115 vpKLMN = _mm_add_ps(_mm_mul_ps(vpKLMN, vtKLMN), vc1); in xnn_f32_sigmoid_ukernel__sse41_p5_div_x24() 129 __m128 veKLMN = _mm_add_ps(_mm_mul_ps(vtKLMN, vpKLMN), vsKLMN); in xnn_f32_sigmoid_ukernel__sse41_p5_div_x24()
|
D | neonfma-rr1-p5-nr2recps-x24.c | 85 float32x4_t vpKLMN = vfmaq_f32(vc4, vc5, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_nr2recps_x24() local 92 vpKLMN = vfmaq_f32(vc3, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_nr2recps_x24() 99 vpKLMN = vfmaq_f32(vc2, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_nr2recps_x24() 106 vpKLMN = vfmaq_f32(vc1, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_nr2recps_x24() 120 const float32x4_t veKLMN = vfmaq_f32(vsKLMN, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_nr2recps_x24()
|
D | sse2-p5-div-x24.c | 94 __m128 vpKLMN = _mm_add_ps(_mm_mul_ps(vc5, vtKLMN), vc4); in xnn_f32_sigmoid_ukernel__sse2_p5_div_x24() local 101 vpKLMN = _mm_add_ps(_mm_mul_ps(vpKLMN, vtKLMN), vc3); in xnn_f32_sigmoid_ukernel__sse2_p5_div_x24() 108 vpKLMN = _mm_add_ps(_mm_mul_ps(vpKLMN, vtKLMN), vc2); in xnn_f32_sigmoid_ukernel__sse2_p5_div_x24() 115 vpKLMN = _mm_add_ps(_mm_mul_ps(vpKLMN, vtKLMN), vc1); in xnn_f32_sigmoid_ukernel__sse2_p5_div_x24() 129 __m128 veKLMN = _mm_add_ps(_mm_mul_ps(vtKLMN, vpKLMN), vsKLMN); in xnn_f32_sigmoid_ukernel__sse2_p5_div_x24()
|
D | neonfma-rr1-p5-nr1recps1fma-x24.c | 85 float32x4_t vpKLMN = vfmaq_f32(vc4, vc5, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_nr1recps1fma_x24() local 92 vpKLMN = vfmaq_f32(vc3, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_nr1recps1fma_x24() 99 vpKLMN = vfmaq_f32(vc2, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_nr1recps1fma_x24() 106 vpKLMN = vfmaq_f32(vc1, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_nr1recps1fma_x24() 120 const float32x4_t veKLMN = vfmaq_f32(vsKLMN, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_nr1recps1fma_x24()
|
D | neonfma-rr1-p5-nr2fma-x24.c | 85 float32x4_t vpKLMN = vfmaq_f32(vc4, vc5, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_nr2fma_x24() local 92 vpKLMN = vfmaq_f32(vc3, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_nr2fma_x24() 99 vpKLMN = vfmaq_f32(vc2, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_nr2fma_x24() 106 vpKLMN = vfmaq_f32(vc1, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_nr2fma_x24() 120 const float32x4_t veKLMN = vfmaq_f32(vsKLMN, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_p5_nr2fma_x24()
|
D | neon-rr2-p5-nr2recps-x24.c | 93 float32x4_t vpKLMN = vmlaq_f32(vc4, vc5, vtKLMN); in xnn_f32_sigmoid_ukernel__neon_rr2_p5_nr2recps_x24() local 100 vpKLMN = vmlaq_f32(vc3, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neon_rr2_p5_nr2recps_x24() 107 vpKLMN = vmlaq_f32(vc2, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neon_rr2_p5_nr2recps_x24() 114 vpKLMN = vmlaq_f32(vc1, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neon_rr2_p5_nr2recps_x24() 128 const float32x4_t veKLMN = vmlaq_f32(vsKLMN, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neon_rr2_p5_nr2recps_x24()
|
D | neonfma-rr1-lut64-p2-div-x24.c | 143 float32x4_t vpKLMN = vmulq_f32(vtKLMN, vc2); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x24() local 150 vpKLMN = vfmsq_f32(vtKLMN, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x24() 157 const float32x4_t vyKLMN = vfmsq_f32(vsKLMN, vsKLMN, vpKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x24()
|
D | wasmsimd-lut64-p2-div-x24.c | 150 v128_t vpKLMN = wasm_f32x4_mul(vtKLMN, vc2); in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x24() local 157 vpKLMN = wasm_f32x4_sub(vtKLMN, wasm_f32x4_mul(vpKLMN, vtKLMN)); in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x24() 164 const v128_t vyKLMN = wasm_f32x4_sub(vsKLMN, wasm_f32x4_mul(vsKLMN, vpKLMN)); in xnn_f32_sigmoid_ukernel__wasmsimd_lut64_p2_div_x24()
|
D | neonfma-rr1-lut64-p2-nr2fma-x24.c | 143 float32x4_t vpKLMN = vmulq_f32(vtKLMN, vc2); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2fma_x24() local 150 vpKLMN = vfmsq_f32(vtKLMN, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2fma_x24() 157 const float32x4_t vyKLMN = vfmsq_f32(vsKLMN, vsKLMN, vpKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2fma_x24()
|
D | neonfma-rr1-lut64-p2-nr2recps-x24.c | 143 float32x4_t vpKLMN = vmulq_f32(vtKLMN, vc2); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2recps_x24() local 150 vpKLMN = vfmsq_f32(vtKLMN, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2recps_x24() 157 const float32x4_t vyKLMN = vfmsq_f32(vsKLMN, vsKLMN, vpKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2recps_x24()
|
D | neonfma-rr1-lut64-p2-nr1recps1fma-x24.c | 143 float32x4_t vpKLMN = vmulq_f32(vtKLMN, vc2); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr1recps1fma_x24() local 150 vpKLMN = vfmsq_f32(vtKLMN, vpKLMN, vtKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr1recps1fma_x24() 157 const float32x4_t vyKLMN = vfmsq_f32(vsKLMN, vsKLMN, vpKLMN); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr1recps1fma_x24()
|