| /external/XNNPACK/src/x8-zip/ |
| D | x4-sse2.c | 36 const __m128i vxy_hi = _mm_unpackhi_epi8(vx, vy); in xnn_x8_zip_x4_ukernel__sse2() local
41 const __m128i vxyzw2 = _mm_unpacklo_epi16(vxy_hi, vzw_hi); in xnn_x8_zip_x4_ukernel__sse2()
42 const __m128i vxyzw3 = _mm_unpackhi_epi16(vxy_hi, vzw_hi); in xnn_x8_zip_x4_ukernel__sse2()
57 const __m128i vxy_hi = _mm_unpackhi_epi8(vx, vy); in xnn_x8_zip_x4_ukernel__sse2() local
62 const __m128i vxyzw2 = _mm_unpacklo_epi16(vxy_hi, vzw_hi); in xnn_x8_zip_x4_ukernel__sse2()
63 const __m128i vxyzw3 = _mm_unpackhi_epi16(vxy_hi, vzw_hi); in xnn_x8_zip_x4_ukernel__sse2()
|
| D | x2-sse2.c | 30 const __m128i vxy_hi = _mm_unpackhi_epi8(vx, vy); in xnn_x8_zip_x2_ukernel__sse2() local
32 _mm_storeu_si128((__m128i*) (o + 16), vxy_hi); in xnn_x8_zip_x2_ukernel__sse2()
41 const __m128i vxy_hi = _mm_unpackhi_epi8(vx, vy); in xnn_x8_zip_x2_ukernel__sse2() local
44 _mm_storeu_si128((__m128i*) o + 1, vxy_hi); in xnn_x8_zip_x2_ukernel__sse2()
|
| D | xm-sse2.c | 46 const __m128i vxy_hi = _mm_unpackhi_epi8(vx, vy); in xnn_x8_zip_xm_ukernel__sse2() local
51 __m128i vxyzw2 = _mm_unpacklo_epi16(vxy_hi, vzw_hi); in xnn_x8_zip_xm_ukernel__sse2()
52 __m128i vxyzw3 = _mm_unpackhi_epi16(vxy_hi, vzw_hi); in xnn_x8_zip_xm_ukernel__sse2()
|
| /external/XNNPACK/src/x32-zip/ |
| D | x4-wasmsimd.c | 38 const v128_t vxy_hi = wasm_v32x4_shuffle(vx, vy, 2, 6, 3, 7); in xnn_x32_zip_x4_ukernel__wasmsimd() local
44 const v128_t vxyzw2 = wasm_v32x4_shuffle(vxy_hi, vzw_hi, 0, 1, 4, 5); in xnn_x32_zip_x4_ukernel__wasmsimd()
45 const v128_t vxyzw3 = wasm_v32x4_shuffle(vxy_hi, vzw_hi, 2, 3, 6, 7); in xnn_x32_zip_x4_ukernel__wasmsimd()
|
| D | x4-sse2.c | 38 const __m128i vxy_hi = _mm_unpackhi_epi32(vx, vy); in xnn_x32_zip_x4_ukernel__sse2() local
44 const __m128i vxyzw2 = _mm_unpacklo_epi64(vxy_hi, vzw_hi); in xnn_x32_zip_x4_ukernel__sse2()
45 const __m128i vxyzw3 = _mm_unpackhi_epi64(vxy_hi, vzw_hi); in xnn_x32_zip_x4_ukernel__sse2()
|
| D | xm-wasmsimd.c | 52 const v128_t vxy_hi = wasm_v32x4_shuffle(vx, vy, 2, 6, 3, 7); in xnn_x32_zip_xm_ukernel__wasmsimd() local
58 const v128_t vxyzw2 = wasm_v32x4_shuffle(vxy_hi, vzw_hi, 0, 1, 4, 5); in xnn_x32_zip_xm_ukernel__wasmsimd()
59 const v128_t vxyzw3 = wasm_v32x4_shuffle(vxy_hi, vzw_hi, 2, 3, 6, 7); in xnn_x32_zip_xm_ukernel__wasmsimd()
|
| D | x2-wasmsimd.c | 31 const v128_t vxy_hi = wasm_v32x4_shuffle(vx, vy, 2, 6, 3, 7); in xnn_x32_zip_x2_ukernel__wasmsimd() local
33 wasm_v128_store(o + 4, vxy_hi); in xnn_x32_zip_x2_ukernel__wasmsimd()
|
| D | x2-sse2.c | 31 const __m128i vxy_hi = _mm_unpackhi_epi32(vx, vy); in xnn_x32_zip_x2_ukernel__sse2() local
33 _mm_storeu_si128((__m128i*) (o + 4), vxy_hi); in xnn_x32_zip_x2_ukernel__sse2()
|
| D | xm-sse2.c | 51 const __m128i vxy_hi = _mm_unpackhi_epi32(vx, vy); in xnn_x32_zip_xm_ukernel__sse2() local
57 const __m128i vxyzw2 = _mm_unpacklo_epi64(vxy_hi, vzw_hi); in xnn_x32_zip_xm_ukernel__sse2()
58 const __m128i vxyzw3 = _mm_unpackhi_epi64(vxy_hi, vzw_hi); in xnn_x32_zip_xm_ukernel__sse2()
|