| /external/XNNPACK/src/f32-vmulcaddc/gen/ |
| D | c8-minmax-sse-2x.c | 57 __m128 vacc1x0123 = _mm_loadu_ps(i1); in xnn_f32_vmulcaddc_minmax_ukernel_c8__sse_2x() local
98 __m128 vacc1x0123 = _mm_loadu_ps(i1); in xnn_f32_vmulcaddc_minmax_ukernel_c8__sse_2x() local
127 __m128 vacc1x0123 = _mm_loadu_ps(i1); in xnn_f32_vmulcaddc_minmax_ukernel_c8__sse_2x() local
|
| D | c8-minmax-neonfma-2x.c | 56 float32x4_t vacc1x0123 = vld1q_f32(i1); i1 += 4; in xnn_f32_vmulcaddc_minmax_ukernel_c8__neonfma_2x() local
87 float32x4_t vacc1x0123 = vld1q_f32(i1); i1 += 4; in xnn_f32_vmulcaddc_minmax_ukernel_c8__neonfma_2x() local
108 float32x4_t vacc1x0123 = vld1q_f32(i1); i1 = (const float*) ((uintptr_t) i1 + c); in xnn_f32_vmulcaddc_minmax_ukernel_c8__neonfma_2x() local
|
| D | c8-minmax-neon-2x.c | 56 float32x4_t vacc1x0123 = vld1q_f32(i1); i1 += 4; in xnn_f32_vmulcaddc_minmax_ukernel_c8__neon_2x() local
91 float32x4_t vacc1x0123 = vld1q_f32(i1); i1 += 4; in xnn_f32_vmulcaddc_minmax_ukernel_c8__neon_2x() local
114 float32x4_t vacc1x0123 = vld1q_f32(i1); i1 = (const float*) ((uintptr_t) i1 + c); in xnn_f32_vmulcaddc_minmax_ukernel_c8__neon_2x() local
|
| D | c4-minmax-sse-2x.c | 55 __m128 vacc1x0123 = _mm_loadu_ps(i1); in xnn_f32_vmulcaddc_minmax_ukernel_c4__sse_2x() local
84 __m128 vacc1x0123 = _mm_loadu_ps(i1); in xnn_f32_vmulcaddc_minmax_ukernel_c4__sse_2x() local
|
| D | c4-minmax-neon-2x.c | 54 float32x4_t vacc1x0123 = vld1q_f32(i1); i1 += 4; in xnn_f32_vmulcaddc_minmax_ukernel_c4__neon_2x() local
77 float32x4_t vacc1x0123 = vld1q_f32(i1); i1 = (const float*) ((uintptr_t) i1 + c); in xnn_f32_vmulcaddc_minmax_ukernel_c4__neon_2x() local
|
| D | c4-minmax-neonfma-2x.c | 54 float32x4_t vacc1x0123 = vld1q_f32(i1); i1 += 4; in xnn_f32_vmulcaddc_minmax_ukernel_c4__neonfma_2x() local
75 float32x4_t vacc1x0123 = vld1q_f32(i1); i1 = (const float*) ((uintptr_t) i1 + c); in xnn_f32_vmulcaddc_minmax_ukernel_c4__neonfma_2x() local
|
| /external/XNNPACK/src/f32-prelu/gen/ |
| D | wasmsimd-minmax-2x8.c | 64 v128_t vacc1x0123 = wasm_i32x4_max(vi1x0123, vzero); in xnn_f32_prelu_ukernel__wasmsimd_minmax_2x8() local
92 v128_t vacc1x0123 = wasm_i32x4_max(vi1x0123, vzero); in xnn_f32_prelu_ukernel__wasmsimd_minmax_2x8() local
114 v128_t vacc1x0123 = wasm_i32x4_max(vi1x0123, vzero); in xnn_f32_prelu_ukernel__wasmsimd_minmax_2x8() local
|
| D | sse-2x8.c | 64 __m128 vacc1x0123 = _mm_max_ps(_mm_setzero_ps(), vi1x0123); in xnn_f32_prelu_ukernel__sse_2x8() local
92 __m128 vacc1x0123 = _mm_max_ps(_mm_setzero_ps(), vi1x0123); in xnn_f32_prelu_ukernel__sse_2x8() local
114 __m128 vacc1x0123 = _mm_max_ps(_mm_setzero_ps(), vi1x0123); in xnn_f32_prelu_ukernel__sse_2x8() local
|
| D | wasmsimd-bitselect-2x8.c | 63 v128_t vacc1x0123 = wasm_f32x4_mul(vi1x0123, vw0123); in xnn_f32_prelu_ukernel__wasmsimd_bitselect_2x8() local
91 v128_t vacc1x0123 = wasm_f32x4_mul(vi1x0123, vw0123); in xnn_f32_prelu_ukernel__wasmsimd_bitselect_2x8() local
113 v128_t vacc1x0123 = wasm_f32x4_mul(vi1x0123, vw0123); in xnn_f32_prelu_ukernel__wasmsimd_bitselect_2x8() local
|
| D | sse41-2x8.c | 66 const __m128 vacc1x0123 = _mm_blendv_ps(vi1x0123, vprod1x0123, vi1x0123); in xnn_f32_prelu_ukernel__sse41_2x8() local
89 __m128 vacc1x0123 = _mm_blendv_ps(vi1x0123, vprod1x0123, vi1x0123); in xnn_f32_prelu_ukernel__sse41_2x8() local
109 __m128 vacc1x0123 = _mm_blendv_ps(vi1x0123, vprod1x0123, vi1x0123); in xnn_f32_prelu_ukernel__sse41_2x8() local
|
| D | wasmsimd-minmax-2x16.c | 74 v128_t vacc1x0123 = wasm_i32x4_max(vi1x0123, vzero); in xnn_f32_prelu_ukernel__wasmsimd_minmax_2x16() local
114 v128_t vacc1x0123 = wasm_i32x4_max(vi1x0123, vzero); in xnn_f32_prelu_ukernel__wasmsimd_minmax_2x16() local
136 v128_t vacc1x0123 = wasm_i32x4_max(vi1x0123, vzero); in xnn_f32_prelu_ukernel__wasmsimd_minmax_2x16() local
|
| D | neon-2x8.c | 60 float32x4_t vacc1x0123 = vmulq_f32(vi1x0123, vw0123); in xnn_f32_prelu_ukernel__neon_2x8() local
85 float32x4_t vacc1x0123 = vmulq_f32(vi1x0123, vw0123); in xnn_f32_prelu_ukernel__neon_2x8() local
104 float32x4_t vacc1x0123 = vmulq_f32(vi1x0123, vw0123); in xnn_f32_prelu_ukernel__neon_2x8() local
|
| D | sse2-2x8.c | 70 …const __m128 vacc1x0123 = _mm_or_ps(_mm_and_ps(vprod1x0123, vmask1x0123), _mm_andnot_ps(vmask1x012… in xnn_f32_prelu_ukernel__sse2_2x8() local
95 …__m128 vacc1x0123 = _mm_or_ps(_mm_and_ps(vprod1x0123, vmask1x0123), _mm_andnot_ps(vmask1x0123, vi1… in xnn_f32_prelu_ukernel__sse2_2x8() local
117 …__m128 vacc1x0123 = _mm_or_ps(_mm_and_ps(vprod1x0123, vmask1x0123), _mm_andnot_ps(vmask1x0123, vi1… in xnn_f32_prelu_ukernel__sse2_2x8() local
|
| D | wasmsimd-bitselect-2x16.c | 73 v128_t vacc1x0123 = wasm_f32x4_mul(vi1x0123, vw0123); in xnn_f32_prelu_ukernel__wasmsimd_bitselect_2x16() local
113 v128_t vacc1x0123 = wasm_f32x4_mul(vi1x0123, vw0123); in xnn_f32_prelu_ukernel__wasmsimd_bitselect_2x16() local
135 v128_t vacc1x0123 = wasm_f32x4_mul(vi1x0123, vw0123); in xnn_f32_prelu_ukernel__wasmsimd_bitselect_2x16() local
|
| D | wasmsimd-minmax-4x8.c | 82 v128_t vacc1x0123 = wasm_i32x4_max(vi1x0123, vzero); in xnn_f32_prelu_ukernel__wasmsimd_minmax_4x8() local
132 v128_t vacc1x0123 = wasm_i32x4_max(vi1x0123, vzero); in xnn_f32_prelu_ukernel__wasmsimd_minmax_4x8() local
168 v128_t vacc1x0123 = wasm_i32x4_max(vi1x0123, vzero); in xnn_f32_prelu_ukernel__wasmsimd_minmax_4x8() local
|
| D | sse-2x4.c | 59 __m128 vacc1x0123 = _mm_max_ps(_mm_setzero_ps(), vi1x0123); in xnn_f32_prelu_ukernel__sse_2x4() local
81 __m128 vacc1x0123 = _mm_max_ps(_mm_setzero_ps(), vi1x0123); in xnn_f32_prelu_ukernel__sse_2x4() local
|
| D | wasmsimd-minmax-2x4.c | 59 v128_t vacc1x0123 = wasm_i32x4_max(vi1x0123, vzero); in xnn_f32_prelu_ukernel__wasmsimd_minmax_2x4() local
81 v128_t vacc1x0123 = wasm_i32x4_max(vi1x0123, vzero); in xnn_f32_prelu_ukernel__wasmsimd_minmax_2x4() local
|
| D | neon-2x16.c | 70 float32x4_t vacc1x0123 = vmulq_f32(vi1x0123, vw0123); in xnn_f32_prelu_ukernel__neon_2x16() local
107 float32x4_t vacc1x0123 = vmulq_f32(vi1x0123, vw0123); in xnn_f32_prelu_ukernel__neon_2x16() local
126 float32x4_t vacc1x0123 = vmulq_f32(vi1x0123, vw0123); in xnn_f32_prelu_ukernel__neon_2x16() local
|
| D | wasmsimd-bitselect-2x4.c | 58 v128_t vacc1x0123 = wasm_f32x4_mul(vi1x0123, vw0123); in xnn_f32_prelu_ukernel__wasmsimd_bitselect_2x4() local
80 v128_t vacc1x0123 = wasm_f32x4_mul(vi1x0123, vw0123); in xnn_f32_prelu_ukernel__wasmsimd_bitselect_2x4() local
|
| D | sse41-2x4.c | 60 const __m128 vacc1x0123 = _mm_blendv_ps(vi1x0123, vprod1x0123, vi1x0123); in xnn_f32_prelu_ukernel__sse41_2x4() local
80 __m128 vacc1x0123 = _mm_blendv_ps(vi1x0123, vprod1x0123, vi1x0123); in xnn_f32_prelu_ukernel__sse41_2x4() local
|
| D | neon-2x4.c | 55 float32x4_t vacc1x0123 = vmulq_f32(vi1x0123, vw0123); in xnn_f32_prelu_ukernel__neon_2x4() local
74 float32x4_t vacc1x0123 = vmulq_f32(vi1x0123, vw0123); in xnn_f32_prelu_ukernel__neon_2x4() local
|
| D | sse2-2x4.c | 62 …const __m128 vacc1x0123 = _mm_or_ps(_mm_and_ps(vprod1x0123, vmask1x0123), _mm_andnot_ps(vmask1x012… in xnn_f32_prelu_ukernel__sse2_2x4() local
84 …__m128 vacc1x0123 = _mm_or_ps(_mm_and_ps(vprod1x0123, vmask1x0123), _mm_andnot_ps(vmask1x0123, vi1… in xnn_f32_prelu_ukernel__sse2_2x4() local
|
| D | wasmsimd-minmax-4x16.c | 96 v128_t vacc1x0123 = wasm_i32x4_max(vi1x0123, vzero); in xnn_f32_prelu_ukernel__wasmsimd_minmax_4x16() local
174 v128_t vacc1x0123 = wasm_i32x4_max(vi1x0123, vzero); in xnn_f32_prelu_ukernel__wasmsimd_minmax_4x16() local
210 v128_t vacc1x0123 = wasm_i32x4_max(vi1x0123, vzero); in xnn_f32_prelu_ukernel__wasmsimd_minmax_4x16() local
|
| D | wasmsimd-bitselect-4x8.c | 81 v128_t vacc1x0123 = wasm_f32x4_mul(vi1x0123, vw0123); in xnn_f32_prelu_ukernel__wasmsimd_bitselect_4x8() local
131 v128_t vacc1x0123 = wasm_f32x4_mul(vi1x0123, vw0123); in xnn_f32_prelu_ukernel__wasmsimd_bitselect_4x8() local
167 v128_t vacc1x0123 = wasm_f32x4_mul(vi1x0123, vw0123); in xnn_f32_prelu_ukernel__wasmsimd_bitselect_4x8() local
|
| D | neon-4x8.c | 76 float32x4_t vacc1x0123 = vmulq_f32(vi1x0123, vw0123); in xnn_f32_prelu_ukernel__neon_4x8() local
121 float32x4_t vacc1x0123 = vmulq_f32(vi1x0123, vw0123); in xnn_f32_prelu_ukernel__neon_4x8() local
152 float32x4_t vacc1x0123 = vmulq_f32(vi1x0123, vw0123); in xnn_f32_prelu_ukernel__neon_4x8() local
|