| /external/XNNPACK/src/f32-velu/gen/ |
| D | velu-neonfma-rr1-p6-x8.c | 62 float32x4_t vp0123 = vfmaq_f32(vc5, vc6, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x8() local
65 vp0123 = vfmaq_f32(vc4, vp0123, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x8()
68 vp0123 = vfmaq_f32(vc3, vp0123, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x8()
71 vp0123 = vfmaq_f32(vc2, vp0123, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x8()
74 vp0123 = vmulq_f32(vp0123, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x8()
82 vp0123 = vfmaq_f32(vt0123, vp0123, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x8()
85 const float32x4_t ve0123 = vmulq_f32(vaddq_f32(vp0123, vs0123), valpha); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x8()
|
| D | velu-neon-rr2-p6-x8.c | 66 float32x4_t vp0123 = vmlaq_f32(vc5, vc6, vt0123); in xnn_f32_velu_ukernel__neon_rr2_p6_x8() local
69 vp0123 = vmlaq_f32(vc4, vp0123, vt0123); in xnn_f32_velu_ukernel__neon_rr2_p6_x8()
72 vp0123 = vmlaq_f32(vc3, vp0123, vt0123); in xnn_f32_velu_ukernel__neon_rr2_p6_x8()
75 vp0123 = vmlaq_f32(vc2, vp0123, vt0123); in xnn_f32_velu_ukernel__neon_rr2_p6_x8()
78 vp0123 = vmulq_f32(vp0123, vt0123); in xnn_f32_velu_ukernel__neon_rr2_p6_x8()
86 vp0123 = vmlaq_f32(vt0123, vp0123, vt0123); in xnn_f32_velu_ukernel__neon_rr2_p6_x8()
89 const float32x4_t ve0123 = vmulq_f32(vaddq_f32(vp0123, vs0123), valpha); in xnn_f32_velu_ukernel__neon_rr2_p6_x8()
|
| D | velu-sse41-rr2-p6-x8.c | 68 __m128 vp0123 = _mm_add_ps(_mm_mul_ps(vc6, vt0123), vc5); in xnn_f32_velu_ukernel__sse41_rr2_p6_x8() local
71 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc4); in xnn_f32_velu_ukernel__sse41_rr2_p6_x8()
74 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc3); in xnn_f32_velu_ukernel__sse41_rr2_p6_x8()
77 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc2); in xnn_f32_velu_ukernel__sse41_rr2_p6_x8()
80 vp0123 = _mm_mul_ps(vp0123, vt0123); in xnn_f32_velu_ukernel__sse41_rr2_p6_x8()
88 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vt0123); in xnn_f32_velu_ukernel__sse41_rr2_p6_x8()
91 const __m128 ve0123 = _mm_mul_ps(_mm_add_ps(vp0123, vs0123), valpha); in xnn_f32_velu_ukernel__sse41_rr2_p6_x8()
|
| D | velu-wasmsimd-arm-rr2-p6-x8.c | 68 v128_t vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vc6, vt0123), vc5); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x8() local
71 vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vp0123, vt0123), vc4); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x8()
74 vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vp0123, vt0123), vc3); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x8()
77 vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vp0123, vt0123), vc2); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x8()
80 vp0123 = wasm_f32x4_mul(vp0123, vt0123); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x8()
88 vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vp0123, vt0123), vt0123); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x8()
91 const v128_t ve0123 = wasm_f32x4_mul(wasm_f32x4_add(vp0123, vs0123), valpha); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x8()
|
| D | velu-neonfma-rr1-p6-x12.c | 68 float32x4_t vp0123 = vfmaq_f32(vc5, vc6, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x12() local
72 vp0123 = vfmaq_f32(vc4, vp0123, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x12()
76 vp0123 = vfmaq_f32(vc3, vp0123, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x12()
80 vp0123 = vfmaq_f32(vc2, vp0123, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x12()
84 vp0123 = vmulq_f32(vp0123, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x12()
95 vp0123 = vfmaq_f32(vt0123, vp0123, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x12()
99 const float32x4_t ve0123 = vmulq_f32(vaddq_f32(vp0123, vs0123), valpha); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x12()
|
| D | velu-sse2-rr2-p6-x8.c | 68 __m128 vp0123 = _mm_add_ps(_mm_mul_ps(vc6, vt0123), vc5); in xnn_f32_velu_ukernel__sse2_rr2_p6_x8() local
71 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc4); in xnn_f32_velu_ukernel__sse2_rr2_p6_x8()
74 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc3); in xnn_f32_velu_ukernel__sse2_rr2_p6_x8()
77 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc2); in xnn_f32_velu_ukernel__sse2_rr2_p6_x8()
80 vp0123 = _mm_mul_ps(vp0123, vt0123); in xnn_f32_velu_ukernel__sse2_rr2_p6_x8()
88 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vt0123); in xnn_f32_velu_ukernel__sse2_rr2_p6_x8()
91 const __m128 ve0123 = _mm_mul_ps(_mm_add_ps(vp0123, vs0123), valpha); in xnn_f32_velu_ukernel__sse2_rr2_p6_x8()
|
| D | velu-wasmsimd-x86-rr2-p6-x8.c | 73 v128_t vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vc6, vt0123), vc5); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_p6_x8() local
77 vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vp0123, vt0123), vc4); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_p6_x8()
80 vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vp0123, vt0123), vc3); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_p6_x8()
83 vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vp0123, vt0123), vc2); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_p6_x8()
86 vp0123 = wasm_f32x4_mul(vp0123, vt0123); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_p6_x8()
94 vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vp0123, vt0123), vt0123); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_p6_x8()
97 const v128_t ve0123 = wasm_f32x4_mul(wasm_f32x4_add(vp0123, vs0123), valpha); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_p6_x8()
|
| D | velu-neon-rr2-p6-x12.c | 73 float32x4_t vp0123 = vmlaq_f32(vc5, vc6, vt0123); in xnn_f32_velu_ukernel__neon_rr2_p6_x12() local
77 vp0123 = vmlaq_f32(vc4, vp0123, vt0123); in xnn_f32_velu_ukernel__neon_rr2_p6_x12()
81 vp0123 = vmlaq_f32(vc3, vp0123, vt0123); in xnn_f32_velu_ukernel__neon_rr2_p6_x12()
85 vp0123 = vmlaq_f32(vc2, vp0123, vt0123); in xnn_f32_velu_ukernel__neon_rr2_p6_x12()
89 vp0123 = vmulq_f32(vp0123, vt0123); in xnn_f32_velu_ukernel__neon_rr2_p6_x12()
100 vp0123 = vmlaq_f32(vt0123, vp0123, vt0123); in xnn_f32_velu_ukernel__neon_rr2_p6_x12()
104 const float32x4_t ve0123 = vmulq_f32(vaddq_f32(vp0123, vs0123), valpha); in xnn_f32_velu_ukernel__neon_rr2_p6_x12()
|
| D | velu-wasmsimd-arm-rr2-p6-x12.c | 75 v128_t vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vc6, vt0123), vc5); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x12() local
79 vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vp0123, vt0123), vc4); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x12()
83 vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vp0123, vt0123), vc3); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x12()
87 vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vp0123, vt0123), vc2); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x12()
91 vp0123 = wasm_f32x4_mul(vp0123, vt0123); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x12()
102 vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vp0123, vt0123), vt0123); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x12()
106 const v128_t ve0123 = wasm_f32x4_mul(wasm_f32x4_add(vp0123, vs0123), valpha); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x12()
|
| D | velu-sse2-rr2-p6-x12.c | 75 __m128 vp0123 = _mm_add_ps(_mm_mul_ps(vc6, vt0123), vc5); in xnn_f32_velu_ukernel__sse2_rr2_p6_x12() local
79 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc4); in xnn_f32_velu_ukernel__sse2_rr2_p6_x12()
83 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc3); in xnn_f32_velu_ukernel__sse2_rr2_p6_x12()
87 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc2); in xnn_f32_velu_ukernel__sse2_rr2_p6_x12()
91 vp0123 = _mm_mul_ps(vp0123, vt0123); in xnn_f32_velu_ukernel__sse2_rr2_p6_x12()
102 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vt0123); in xnn_f32_velu_ukernel__sse2_rr2_p6_x12()
106 const __m128 ve0123 = _mm_mul_ps(_mm_add_ps(vp0123, vs0123), valpha); in xnn_f32_velu_ukernel__sse2_rr2_p6_x12()
|
| D | velu-sse41-rr2-p6-x12.c | 75 __m128 vp0123 = _mm_add_ps(_mm_mul_ps(vc6, vt0123), vc5); in xnn_f32_velu_ukernel__sse41_rr2_p6_x12() local
79 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc4); in xnn_f32_velu_ukernel__sse41_rr2_p6_x12()
83 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc3); in xnn_f32_velu_ukernel__sse41_rr2_p6_x12()
87 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc2); in xnn_f32_velu_ukernel__sse41_rr2_p6_x12()
91 vp0123 = _mm_mul_ps(vp0123, vt0123); in xnn_f32_velu_ukernel__sse41_rr2_p6_x12()
102 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vt0123); in xnn_f32_velu_ukernel__sse41_rr2_p6_x12()
106 const __m128 ve0123 = _mm_mul_ps(_mm_add_ps(vp0123, vs0123), valpha); in xnn_f32_velu_ukernel__sse41_rr2_p6_x12()
|
| D | velu-neonfma-rr1-p6-x16.c | 74 float32x4_t vp0123 = vfmaq_f32(vc5, vc6, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x16() local
79 vp0123 = vfmaq_f32(vc4, vp0123, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x16()
84 vp0123 = vfmaq_f32(vc3, vp0123, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x16()
89 vp0123 = vfmaq_f32(vc2, vp0123, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x16()
94 vp0123 = vmulq_f32(vp0123, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x16()
108 vp0123 = vfmaq_f32(vt0123, vp0123, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x16()
113 const float32x4_t ve0123 = vmulq_f32(vaddq_f32(vp0123, vs0123), valpha); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x16()
|
| D | velu-wasmsimd-arm-rr2-p6-x16.c | 82 v128_t vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vc6, vt0123), vc5); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x16() local
87 vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vp0123, vt0123), vc4); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x16()
92 vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vp0123, vt0123), vc3); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x16()
97 vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vp0123, vt0123), vc2); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x16()
102 vp0123 = wasm_f32x4_mul(vp0123, vt0123); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x16()
116 vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vp0123, vt0123), vt0123); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x16()
121 const v128_t ve0123 = wasm_f32x4_mul(wasm_f32x4_add(vp0123, vs0123), valpha); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x16()
|
| D | velu-wasmsimd-x86-rr2-p6-x12.c | 82 v128_t vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vc6, vt0123), vc5); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_p6_x12() local
88 vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vp0123, vt0123), vc4); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_p6_x12()
92 vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vp0123, vt0123), vc3); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_p6_x12()
96 vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vp0123, vt0123), vc2); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_p6_x12()
100 vp0123 = wasm_f32x4_mul(vp0123, vt0123); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_p6_x12()
111 vp0123 = wasm_f32x4_add(wasm_f32x4_mul(vp0123, vt0123), vt0123); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_p6_x12()
115 const v128_t ve0123 = wasm_f32x4_mul(wasm_f32x4_add(vp0123, vs0123), valpha); in xnn_f32_velu_ukernel__wasmsimd_x86_rr2_p6_x12()
|
| D | velu-sse41-rr2-p6-x16.c | 82 __m128 vp0123 = _mm_add_ps(_mm_mul_ps(vc6, vt0123), vc5); in xnn_f32_velu_ukernel__sse41_rr2_p6_x16() local
87 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc4); in xnn_f32_velu_ukernel__sse41_rr2_p6_x16()
92 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc3); in xnn_f32_velu_ukernel__sse41_rr2_p6_x16()
97 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc2); in xnn_f32_velu_ukernel__sse41_rr2_p6_x16()
102 vp0123 = _mm_mul_ps(vp0123, vt0123); in xnn_f32_velu_ukernel__sse41_rr2_p6_x16()
116 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vt0123); in xnn_f32_velu_ukernel__sse41_rr2_p6_x16()
121 const __m128 ve0123 = _mm_mul_ps(_mm_add_ps(vp0123, vs0123), valpha); in xnn_f32_velu_ukernel__sse41_rr2_p6_x16()
|
| D | velu-neon-rr2-p6-x16.c | 80 float32x4_t vp0123 = vmlaq_f32(vc5, vc6, vt0123); in xnn_f32_velu_ukernel__neon_rr2_p6_x16() local
85 vp0123 = vmlaq_f32(vc4, vp0123, vt0123); in xnn_f32_velu_ukernel__neon_rr2_p6_x16()
90 vp0123 = vmlaq_f32(vc3, vp0123, vt0123); in xnn_f32_velu_ukernel__neon_rr2_p6_x16()
95 vp0123 = vmlaq_f32(vc2, vp0123, vt0123); in xnn_f32_velu_ukernel__neon_rr2_p6_x16()
100 vp0123 = vmulq_f32(vp0123, vt0123); in xnn_f32_velu_ukernel__neon_rr2_p6_x16()
114 vp0123 = vmlaq_f32(vt0123, vp0123, vt0123); in xnn_f32_velu_ukernel__neon_rr2_p6_x16()
119 const float32x4_t ve0123 = vmulq_f32(vaddq_f32(vp0123, vs0123), valpha); in xnn_f32_velu_ukernel__neon_rr2_p6_x16()
|
| D | velu-neonfma-rr1-p6-x20.c | 80 float32x4_t vp0123 = vfmaq_f32(vc5, vc6, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x20() local
86 vp0123 = vfmaq_f32(vc4, vp0123, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x20()
92 vp0123 = vfmaq_f32(vc3, vp0123, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x20()
98 vp0123 = vfmaq_f32(vc2, vp0123, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x20()
104 vp0123 = vmulq_f32(vp0123, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x20()
121 vp0123 = vfmaq_f32(vt0123, vp0123, vt0123); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x20()
127 const float32x4_t ve0123 = vmulq_f32(vaddq_f32(vp0123, vs0123), valpha); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x20()
|
| D | velu-sse2-rr2-p6-x16.c | 82 __m128 vp0123 = _mm_add_ps(_mm_mul_ps(vc6, vt0123), vc5); in xnn_f32_velu_ukernel__sse2_rr2_p6_x16() local
87 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc4); in xnn_f32_velu_ukernel__sse2_rr2_p6_x16()
92 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc3); in xnn_f32_velu_ukernel__sse2_rr2_p6_x16()
97 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc2); in xnn_f32_velu_ukernel__sse2_rr2_p6_x16()
102 vp0123 = _mm_mul_ps(vp0123, vt0123); in xnn_f32_velu_ukernel__sse2_rr2_p6_x16()
116 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vt0123); in xnn_f32_velu_ukernel__sse2_rr2_p6_x16()
121 const __m128 ve0123 = _mm_mul_ps(_mm_add_ps(vp0123, vs0123), valpha); in xnn_f32_velu_ukernel__sse2_rr2_p6_x16()
|
| /external/XNNPACK/src/f32-raddstoreexpminusmax/gen/ |
| D | wasmsimd-p5-x4.c | 69 v128_t vp0123 = wasm_f32x4_add(vc4, wasm_f32x4_mul(vc5, vt0123)); in xnn_f32_raddstoreexpminusmax_ukernel__wasmsimd_p5_x4() local
71 vp0123 = wasm_f32x4_add(vc3, wasm_f32x4_mul(vp0123, vt0123)); in xnn_f32_raddstoreexpminusmax_ukernel__wasmsimd_p5_x4()
73 vp0123 = wasm_f32x4_add(vc2, wasm_f32x4_mul(vp0123, vt0123)); in xnn_f32_raddstoreexpminusmax_ukernel__wasmsimd_p5_x4()
75 vp0123 = wasm_f32x4_add(vc1, wasm_f32x4_mul(vp0123, vt0123)); in xnn_f32_raddstoreexpminusmax_ukernel__wasmsimd_p5_x4()
83 v128_t vf0123 = wasm_f32x4_add(vs0123, wasm_f32x4_mul(vt0123, vp0123)); in xnn_f32_raddstoreexpminusmax_ukernel__wasmsimd_p5_x4()
|
| D | sse2-p5-x4.c | 69 __m128 vp0123 = _mm_add_ps(_mm_mul_ps(vc5, vt0123), vc4); in xnn_f32_raddstoreexpminusmax_ukernel__sse2_p5_x4() local
71 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc3); in xnn_f32_raddstoreexpminusmax_ukernel__sse2_p5_x4()
73 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc2); in xnn_f32_raddstoreexpminusmax_ukernel__sse2_p5_x4()
75 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc1); in xnn_f32_raddstoreexpminusmax_ukernel__sse2_p5_x4()
83 __m128 vf0123 = _mm_add_ps(_mm_mul_ps(vt0123, vp0123), vs0123); in xnn_f32_raddstoreexpminusmax_ukernel__sse2_p5_x4()
|
| D | sse2-p5-x8-acc2.c | 77 __m128 vp0123 = _mm_add_ps(_mm_mul_ps(vc5, vt0123), vc4); in xnn_f32_raddstoreexpminusmax_ukernel__sse2_p5_x8_acc2() local
80 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc3); in xnn_f32_raddstoreexpminusmax_ukernel__sse2_p5_x8_acc2()
83 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc2); in xnn_f32_raddstoreexpminusmax_ukernel__sse2_p5_x8_acc2()
86 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc1); in xnn_f32_raddstoreexpminusmax_ukernel__sse2_p5_x8_acc2()
96 __m128 vf0123 = _mm_add_ps(_mm_mul_ps(vt0123, vp0123), vs0123); in xnn_f32_raddstoreexpminusmax_ukernel__sse2_p5_x8_acc2()
|
| D | neonfma-p5-x8-acc2.c | 80 float32x4_t vp0123 = vfmaq_f32(vc4, vc5, vt0123); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_p5_x8_acc2() local
83 vp0123 = vfmaq_f32(vc3, vp0123, vt0123); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_p5_x8_acc2()
86 vp0123 = vfmaq_f32(vc2, vp0123, vt0123); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_p5_x8_acc2()
89 vp0123 = vfmaq_f32(vc1, vp0123, vt0123); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_p5_x8_acc2()
99 float32x4_t vf0123 = vfmaq_f32(vs0123, vp0123, vt0123); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_p5_x8_acc2()
|
| D | wasmsimd-p5-x8.c | 76 v128_t vp0123 = wasm_f32x4_add(vc4, wasm_f32x4_mul(vc5, vt0123)); in xnn_f32_raddstoreexpminusmax_ukernel__wasmsimd_p5_x8() local
79 vp0123 = wasm_f32x4_add(vc3, wasm_f32x4_mul(vp0123, vt0123)); in xnn_f32_raddstoreexpminusmax_ukernel__wasmsimd_p5_x8()
82 vp0123 = wasm_f32x4_add(vc2, wasm_f32x4_mul(vp0123, vt0123)); in xnn_f32_raddstoreexpminusmax_ukernel__wasmsimd_p5_x8()
85 vp0123 = wasm_f32x4_add(vc1, wasm_f32x4_mul(vp0123, vt0123)); in xnn_f32_raddstoreexpminusmax_ukernel__wasmsimd_p5_x8()
95 v128_t vf0123 = wasm_f32x4_add(vs0123, wasm_f32x4_mul(vt0123, vp0123)); in xnn_f32_raddstoreexpminusmax_ukernel__wasmsimd_p5_x8()
|
| D | neon-p5-x8.c | 80 float32x4_t vp0123 = vmlaq_f32(vc4, vc5, vt0123); in xnn_f32_raddstoreexpminusmax_ukernel__neon_p5_x8() local
83 vp0123 = vmlaq_f32(vc3, vp0123, vt0123); in xnn_f32_raddstoreexpminusmax_ukernel__neon_p5_x8()
86 vp0123 = vmlaq_f32(vc2, vp0123, vt0123); in xnn_f32_raddstoreexpminusmax_ukernel__neon_p5_x8()
89 vp0123 = vmlaq_f32(vc1, vp0123, vt0123); in xnn_f32_raddstoreexpminusmax_ukernel__neon_p5_x8()
99 float32x4_t vf0123 = vmlaq_f32(vs0123, vp0123, vt0123); in xnn_f32_raddstoreexpminusmax_ukernel__neon_p5_x8()
|
| D | sse2-p5-x8.c | 76 __m128 vp0123 = _mm_add_ps(_mm_mul_ps(vc5, vt0123), vc4); in xnn_f32_raddstoreexpminusmax_ukernel__sse2_p5_x8() local
79 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc3); in xnn_f32_raddstoreexpminusmax_ukernel__sse2_p5_x8()
82 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc2); in xnn_f32_raddstoreexpminusmax_ukernel__sse2_p5_x8()
85 vp0123 = _mm_add_ps(_mm_mul_ps(vp0123, vt0123), vc1); in xnn_f32_raddstoreexpminusmax_ukernel__sse2_p5_x8()
95 __m128 vf0123 = _mm_add_ps(_mm_mul_ps(vt0123, vp0123), vs0123); in xnn_f32_raddstoreexpminusmax_ukernel__sse2_p5_x8()
|