/external/XNNPACK/src/f32-ibilinear/gen/ |
D | neonfma-c8.c | 55 const float32x4_t vtd0123 = vsubq_f32(vtr0123, vtl0123); in xnn_f32_ibilinear_ukernel__neonfma_c8() local 61 const float32x4_t vt0123 = vfmaq_f32(vtl0123, vtd0123, valphah); in xnn_f32_ibilinear_ukernel__neonfma_c8() 66 const float32x4_t vt0123 = vfmaq_lane_f32(vtl0123, vtd0123, valphahv, 0); in xnn_f32_ibilinear_ukernel__neonfma_c8() 92 const float32x4_t vtd0123 = vsubq_f32(vtr0123, vtl0123); in xnn_f32_ibilinear_ukernel__neonfma_c8() local 96 const float32x4_t vt0123 = vfmaq_f32(vtl0123, vtd0123, valphah); in xnn_f32_ibilinear_ukernel__neonfma_c8() 99 const float32x4_t vt0123 = vfmaq_lane_f32(vtl0123, vtd0123, valphahv, 0); in xnn_f32_ibilinear_ukernel__neonfma_c8() 120 const float32x4_t vtd0123 = vsubq_f32(vtr0123, vtl0123); in xnn_f32_ibilinear_ukernel__neonfma_c8() local 124 const float32x4_t vt0123 = vfmaq_f32(vtl0123, vtd0123, valphah); in xnn_f32_ibilinear_ukernel__neonfma_c8() 127 const float32x4_t vt0123 = vfmaq_lane_f32(vtl0123, vtd0123, valphahv, 0); in xnn_f32_ibilinear_ukernel__neonfma_c8()
|
D | neonfma-c4.c | 51 const float32x4_t vtd0123 = vsubq_f32(vtr0123, vtl0123); in xnn_f32_ibilinear_ukernel__neonfma_c4() local 55 const float32x4_t vt0123 = vfmaq_f32(vtl0123, vtd0123, valphah); in xnn_f32_ibilinear_ukernel__neonfma_c4() 58 const float32x4_t vt0123 = vfmaq_lane_f32(vtl0123, vtd0123, valphahv, 0); in xnn_f32_ibilinear_ukernel__neonfma_c4() 78 const float32x4_t vtd0123 = vsubq_f32(vtr0123, vtl0123); in xnn_f32_ibilinear_ukernel__neonfma_c4() local 82 const float32x4_t vt0123 = vfmaq_f32(vtl0123, vtd0123, valphah); in xnn_f32_ibilinear_ukernel__neonfma_c4() 85 const float32x4_t vt0123 = vfmaq_lane_f32(vtl0123, vtd0123, valphahv, 0); in xnn_f32_ibilinear_ukernel__neonfma_c4()
|
D | neon-c8.c | 51 const float32x4_t vtd0123 = vsubq_f32(vtr0123, vtl0123); in xnn_f32_ibilinear_ukernel__neon_c8() local 56 const float32x4_t vt0123 = vmlaq_lane_f32(vtl0123, vtd0123, valphahv, 0); in xnn_f32_ibilinear_ukernel__neon_c8() 76 const float32x4_t vtd0123 = vsubq_f32(vtr0123, vtl0123); in xnn_f32_ibilinear_ukernel__neon_c8() local 79 const float32x4_t vt0123 = vmlaq_lane_f32(vtl0123, vtd0123, valphahv, 0); in xnn_f32_ibilinear_ukernel__neon_c8() 95 const float32x4_t vtd0123 = vsubq_f32(vtr0123, vtl0123); in xnn_f32_ibilinear_ukernel__neon_c8() local 98 const float32x4_t vt0123 = vmlaq_lane_f32(vtl0123, vtd0123, valphahv, 0); in xnn_f32_ibilinear_ukernel__neon_c8()
|
D | sse-c8.c | 58 const __m128 vtd0123 = _mm_sub_ps(vtr0123, vtl0123); in xnn_f32_ibilinear_ukernel__sse_c8() local 63 const __m128 vt0123 = _mm_add_ps(vtl0123, _mm_mul_ps(vtd0123, valphah)); in xnn_f32_ibilinear_ukernel__sse_c8() 88 const __m128 vtd0123 = _mm_sub_ps(vtr0123, vtl0123); in xnn_f32_ibilinear_ukernel__sse_c8() local 91 const __m128 vt0123 = _mm_add_ps(vtl0123, _mm_mul_ps(vtd0123, valphah)); in xnn_f32_ibilinear_ukernel__sse_c8() 107 const __m128 vtd0123 = _mm_sub_ps(vtr0123, vtl0123); in xnn_f32_ibilinear_ukernel__sse_c8() local 110 const __m128 vt0123 = _mm_add_ps(vtl0123, _mm_mul_ps(vtd0123, valphah)); in xnn_f32_ibilinear_ukernel__sse_c8()
|
D | neon-c4.c | 47 const float32x4_t vtd0123 = vsubq_f32(vtr0123, vtl0123); in xnn_f32_ibilinear_ukernel__neon_c4() local 50 const float32x4_t vt0123 = vmlaq_lane_f32(vtl0123, vtd0123, valphahv, 0); in xnn_f32_ibilinear_ukernel__neon_c4() 65 const float32x4_t vtd0123 = vsubq_f32(vtr0123, vtl0123); in xnn_f32_ibilinear_ukernel__neon_c4() local 68 const float32x4_t vt0123 = vmlaq_lane_f32(vtl0123, vtd0123, valphahv, 0); in xnn_f32_ibilinear_ukernel__neon_c4()
|
D | sse-c4.c | 54 const __m128 vtd0123 = _mm_sub_ps(vtr0123, vtl0123); in xnn_f32_ibilinear_ukernel__sse_c4() local 57 const __m128 vt0123 = _mm_add_ps(vtl0123, _mm_mul_ps(vtd0123, valphah)); in xnn_f32_ibilinear_ukernel__sse_c4() 73 const __m128 vtd0123 = _mm_sub_ps(vtr0123, vtl0123); in xnn_f32_ibilinear_ukernel__sse_c4() local 76 const __m128 vt0123 = _mm_add_ps(vtl0123, _mm_mul_ps(vtd0123, valphah)); in xnn_f32_ibilinear_ukernel__sse_c4()
|
D | wasmsimd-c8.c | 56 const v128_t vtd0123 = wasm_f32x4_sub(vtr0123, vtl0123); in xnn_f32_ibilinear_ukernel__wasmsimd_c8() local 61 const v128_t vt0123 = wasm_f32x4_add(vtl0123, wasm_f32x4_mul(vtd0123, valphah)); in xnn_f32_ibilinear_ukernel__wasmsimd_c8()
|
/external/XNNPACK/src/f32-ibilinear/ |
D | neon.c.in | 98 const float32x4_t vtd0123 = vsubq_f32(vtr0123, vtl0123); 103 const float32x4_t vt0123 = vfmaq_f32(vtl0123, vtd0123, valphah); 106 const float32x4_t vt0123 = vfmaq_lane_f32(vtl0123, vtd0123, valphahv, 0); 110 const float32x4_t vt0123 = vmlaq_lane_f32(vtl0123, vtd0123, valphahv, 0); 133 const float32x4_t vtd0123 = vsubq_f32(vtr0123, vtl0123); 138 const float32x4_t vt0123 = vfmaq_f32(vtl0123, vtd0123, valphah); 141 const float32x4_t vt0123 = vfmaq_lane_f32(vtl0123, vtd0123, valphahv, 0); 145 const float32x4_t vt0123 = vmlaq_lane_f32(vtl0123, vtd0123, valphahv, 0);
|
D | sse.c.in | 89 const __m128 vtd0123 = _mm_sub_ps(vtr0123, vtl0123); 92 const __m128 vt0123 = _mm_add_ps(vtl0123, _mm_mul_ps(vtd0123, valphah)); 108 const __m128 vtd0123 = _mm_sub_ps(vtr0123, vtl0123); 111 const __m128 vt0123 = _mm_add_ps(vtl0123, _mm_mul_ps(vtd0123, valphah));
|