/external/XNNPACK/src/f32-ibilinear/gen/ |
D | neonfma-c8.c | 61 const float32x4_t vt0123 = vfmaq_f32(vtl0123, vtd0123, valphah); in xnn_f32_ibilinear_ukernel__neonfma_c8() local 66 const float32x4_t vt0123 = vfmaq_lane_f32(vtl0123, vtd0123, valphahv, 0); 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() local 99 const float32x4_t vt0123 = vfmaq_lane_f32(vtl0123, vtd0123, valphahv, 0); 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() local 127 const float32x4_t vt0123 = vfmaq_lane_f32(vtl0123, vtd0123, valphahv, 0); in xnn_f32_ibilinear_ukernel__neonfma_c8() local
|
D | neonfma-c4.c | 55 const float32x4_t vt0123 = vfmaq_f32(vtl0123, vtd0123, valphah); in xnn_f32_ibilinear_ukernel__neonfma_c4() local 58 const float32x4_t vt0123 = vfmaq_lane_f32(vtl0123, vtd0123, valphahv, 0); 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() local 85 const float32x4_t vt0123 = vfmaq_lane_f32(vtl0123, vtd0123, valphahv, 0); in xnn_f32_ibilinear_ukernel__neonfma_c4() local
|
D | neon-c8.c | 56 const float32x4_t vt0123 = vmlaq_lane_f32(vtl0123, vtd0123, valphahv, 0); 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() local 98 const float32x4_t vt0123 = vmlaq_lane_f32(vtl0123, vtd0123, valphahv, 0); in xnn_f32_ibilinear_ukernel__neon_c8() local
|
D | sse-c8.c | 63 const __m128 vt0123 = _mm_add_ps(vtl0123, _mm_mul_ps(vtd0123, valphah)); 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() local 110 const __m128 vt0123 = _mm_add_ps(vtl0123, _mm_mul_ps(vtd0123, valphah)); in xnn_f32_ibilinear_ukernel__sse_c8() local
|
D | neon-c4.c | 50 const float32x4_t vt0123 = vmlaq_lane_f32(vtl0123, vtd0123, valphahv, 0); 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() local
|
D | sse-c4.c | 57 const __m128 vt0123 = _mm_add_ps(vtl0123, _mm_mul_ps(vtd0123, valphah)); 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() local
|
D | wasmsimd-c8.c | 61 const v128_t vt0123 = wasm_f32x4_add(vtl0123, wasm_f32x4_mul(vtd0123, valphah)); in xnn_f32_ibilinear_ukernel__wasmsimd_c8() local
|
/external/XNNPACK/src/f32-raddstoreexpminusmax/gen/ |
D | sse2-p5-x4.c | 64 __m128 vt0123 = _mm_add_ps(_mm_mul_ps(vn0123, vminus_ln2_hi), vx0123); in xnn_f32_raddstoreexpminusmax_ukernel__sse2_p5_x4() local
|
D | wasmsimd-p5-x4.c | 64 v128_t vt0123 = wasm_f32x4_add(vx0123, wasm_f32x4_mul(vn0123, vminus_ln2_hi)); in xnn_f32_raddstoreexpminusmax_ukernel__wasmsimd_p5_x4() local
|
D | sse2-p5-x8.c | 69 __m128 vt0123 = _mm_add_ps(_mm_mul_ps(vn0123, vminus_ln2_hi), vx0123); in xnn_f32_raddstoreexpminusmax_ukernel__sse2_p5_x8() local
|
D | sse2-p5-x8-acc2.c | 70 __m128 vt0123 = _mm_add_ps(_mm_mul_ps(vn0123, vminus_ln2_hi), vx0123); in xnn_f32_raddstoreexpminusmax_ukernel__sse2_p5_x8_acc2() local
|
D | neonfma-p5-x8-acc2.c | 73 float32x4_t vt0123 = vfmaq_f32(vx0123, vn0123, vminus_ln2_hi); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_p5_x8_acc2() local
|
D | wasmsimd-p5-x8.c | 69 v128_t vt0123 = wasm_f32x4_add(vx0123, wasm_f32x4_mul(vn0123, vminus_ln2_hi)); in xnn_f32_raddstoreexpminusmax_ukernel__wasmsimd_p5_x8() local
|
D | neon-p5-x8.c | 73 float32x4_t vt0123 = vmlaq_f32(vx0123, vn0123, vminus_ln2_hi); in xnn_f32_raddstoreexpminusmax_ukernel__neon_p5_x8() local
|
D | neon-p5-x8-acc2.c | 74 float32x4_t vt0123 = vmlaq_f32(vx0123, vn0123, vminus_ln2_hi); in xnn_f32_raddstoreexpminusmax_ukernel__neon_p5_x8_acc2() local
|
D | neonfma-p5-x8.c | 72 float32x4_t vt0123 = vfmaq_f32(vx0123, vn0123, vminus_ln2_hi); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_p5_x8() local
|
D | wasmsimd-p5-x8-acc2.c | 70 v128_t vt0123 = wasm_f32x4_add(vx0123, wasm_f32x4_mul(vn0123, vminus_ln2_hi)); in xnn_f32_raddstoreexpminusmax_ukernel__wasmsimd_p5_x8_acc2() local
|
D | sse2-p5-x12-acc2.c | 75 __m128 vt0123 = _mm_add_ps(_mm_mul_ps(vn0123, vminus_ln2_hi), vx0123); in xnn_f32_raddstoreexpminusmax_ukernel__sse2_p5_x12_acc2() local
|
D | sse2-p5-x12.c | 74 __m128 vt0123 = _mm_add_ps(_mm_mul_ps(vn0123, vminus_ln2_hi), vx0123); in xnn_f32_raddstoreexpminusmax_ukernel__sse2_p5_x12() local
|
/external/XNNPACK/src/f32-velu/gen/ |
D | velu-sse41-rr2-p6-x8.c | 62 __m128 vt0123 = _mm_add_ps(_mm_mul_ps(vn0123, vminus_ln2_hi), vz0123); in xnn_f32_velu_ukernel__sse41_rr2_p6_x8() local
|
D | velu-neonfma-rr1-p6-x8.c | 59 float32x4_t vt0123 = vfmaq_f32(vz0123, vn0123, vminus_ln2); in xnn_f32_velu_ukernel__neonfma_rr1_p6_x8() local
|
D | velu-sse2-rr2-p6-x8.c | 62 __m128 vt0123 = _mm_add_ps(_mm_mul_ps(vn0123, vminus_ln2_hi), vz0123); in xnn_f32_velu_ukernel__sse2_rr2_p6_x8() local
|
D | velu-wasmsimd-arm-rr2-p6-x8.c | 62 v128_t vt0123 = wasm_f32x4_add(wasm_f32x4_mul(vn0123, vminus_ln2_hi), vz0123); in xnn_f32_velu_ukernel__wasmsimd_arm_rr2_p6_x8() local
|
/external/XNNPACK/src/f32-sigmoid/gen/ |
D | wasmsimd-p5-div-x8.c | 55 v128_t vt0123 = wasm_f32x4_add(vz0123, wasm_f32x4_mul(vn0123, vln2_hi)); in xnn_f32_sigmoid_ukernel__wasmsimd_p5_div_x8() local
|
D | sse41-p5-div-x8.c | 55 __m128 vt0123 = _mm_add_ps(_mm_mul_ps(vn0123, vminus_ln2_hi), vz0123); in xnn_f32_sigmoid_ukernel__sse41_p5_div_x8() local
|