/aosp_15_r20/external/XNNPACK/src/x32-transposec/gen/ |
H A D | 2x2-reuse-multi-zip-neon.c | 46 const uint32x2_t v1_0 = vld1_u32(i0); i0 = (uint32_t*) ((uintptr_t) i0 + input_stride); in xnn_x32_transposec_ukernel__2x2_reuse_multi_zip_neon() local 57 const uint32x2_t v1_0 = vld1_u32(i0); in xnn_x32_transposec_ukernel__2x2_reuse_multi_zip_neon() local
|
H A D | 2x2-multi-multi-zip-neon.c | 48 const uint32x2_t v1_0 = vld1_u32(i0); i0 = (uint32_t*) ((uintptr_t) i0 + input_offset); in xnn_x32_transposec_ukernel__2x2_multi_multi_zip_neon() local 59 const uint32x2_t v1_0 = vld1_u32(i0); in xnn_x32_transposec_ukernel__2x2_multi_multi_zip_neon() local
|
H A D | 2x2-reuse-switch-zip-neon.c | 44 const uint32x2_t v1_0 = vld1_u32(i0); i0 = (uint32_t*) ((uintptr_t) i0 + input_stride); in xnn_x32_transposec_ukernel__2x2_reuse_switch_zip_neon() local 63 const uint32x2_t v1_0 = vld1_u32(i0); in xnn_x32_transposec_ukernel__2x2_reuse_switch_zip_neon() local
|
H A D | 2x2-reuse-dec-zip-neon.c | 46 const uint32x2_t v1_0 = vld1_u32(i0); i0 = (uint32_t*) ((uintptr_t) i0 + input_stride); in xnn_x32_transposec_ukernel__2x2_reuse_dec_zip_neon() local 62 const uint32x2_t v1_0 = vld1_u32(i0); in xnn_x32_transposec_ukernel__2x2_reuse_dec_zip_neon() local
|
H A D | 2x2-multi-switch-zip-neon.c | 46 const uint32x2_t v1_0 = vld1_u32(i0); i0 = (uint32_t*) ((uintptr_t) i0 + input_offset); in xnn_x32_transposec_ukernel__2x2_multi_switch_zip_neon() local 65 const uint32x2_t v1_0 = vld1_u32(i0); in xnn_x32_transposec_ukernel__2x2_multi_switch_zip_neon() local
|
H A D | 2x2-reuse-mov-zip-neon.c | 46 const uint32x2_t v1_0 = vld1_u32(i0); i0 = (uint32_t*) ((uintptr_t) i0 + input_stride); in xnn_x32_transposec_ukernel__2x2_reuse_mov_zip_neon() local 63 const uint32x2_t v1_0 = vld1_u32(i0); in xnn_x32_transposec_ukernel__2x2_reuse_mov_zip_neon() local
|
H A D | 2x2-multi-dec-zip-neon.c | 48 const uint32x2_t v1_0 = vld1_u32(i0); i0 = (uint32_t*) ((uintptr_t) i0 + input_offset); in xnn_x32_transposec_ukernel__2x2_multi_dec_zip_neon() local 64 const uint32x2_t v1_0 = vld1_u32(i0); in xnn_x32_transposec_ukernel__2x2_multi_dec_zip_neon() local
|
H A D | 2x2-multi-mov-zip-neon.c | 48 const uint32x2_t v1_0 = vld1_u32(i0); i0 = (uint32_t*) ((uintptr_t) i0 + input_offset); in xnn_x32_transposec_ukernel__2x2_multi_mov_zip_neon() local 65 const uint32x2_t v1_0 = vld1_u32(i0); in xnn_x32_transposec_ukernel__2x2_multi_mov_zip_neon() local
|
H A D | 4x4-reuse-multi-zip-neon.c | 59 const uint32x4x2_t v1_0 = vzipq_u32(v2_0, v2_2); in xnn_x32_transposec_ukernel__4x4_reuse_multi_zip_neon() local 85 const uint32x4x2_t v1_0 = vzipq_u32(v2_0, v2_2); in xnn_x32_transposec_ukernel__4x4_reuse_multi_zip_neon() local
|
H A D | 4x4-multi-multi-zip-neon.c | 63 const uint32x4x2_t v1_0 = vzipq_u32(v2_0, v2_2); in xnn_x32_transposec_ukernel__4x4_multi_multi_zip_neon() local 87 const uint32x4x2_t v1_0 = vzipq_u32(v2_0, v2_2); in xnn_x32_transposec_ukernel__4x4_multi_multi_zip_neon() local
|
H A D | 4x4-reuse-dec-zip-neon.c | 51 const uint32x4x2_t v1_0 = vzipq_u32(v2_0, v2_2); in xnn_x32_transposec_ukernel__4x4_reuse_dec_zip_neon() local 88 const uint32x4x2_t v1_0 = vzipq_u32(v2_0, v2_2); in xnn_x32_transposec_ukernel__4x4_reuse_dec_zip_neon() local
|
H A D | 4x4-reuse-multi-wasmsimd.c | 63 const v128_t v1_0 = wasm_v32x4_shuffle(v2_0, v2_2, 0, 4, 1, 5); in xnn_x32_transposec_ukernel__4x4_reuse_multi_wasmsimd() local 96 const v128_t v1_0 = wasm_v32x4_shuffle(v2_0, v2_2, 0, 4, 1, 5); in xnn_x32_transposec_ukernel__4x4_reuse_multi_wasmsimd() local
|
H A D | 4x4-reuse-mov-zip-neon.c | 51 const uint32x4x2_t v1_0 = vzipq_u32(v2_0, v2_2); in xnn_x32_transposec_ukernel__4x4_reuse_mov_zip_neon() local 91 const uint32x4x2_t v1_0 = vzipq_u32(v2_0, v2_2); in xnn_x32_transposec_ukernel__4x4_reuse_mov_zip_neon() local
|
H A D | 4x4-multi-dec-zip-neon.c | 55 const uint32x4x2_t v1_0 = vzipq_u32(v2_0, v2_2); in xnn_x32_transposec_ukernel__4x4_multi_dec_zip_neon() local 90 const uint32x4x2_t v1_0 = vzipq_u32(v2_0, v2_2); in xnn_x32_transposec_ukernel__4x4_multi_dec_zip_neon() local
|
H A D | 4x4-reuse-multi-sse2.c | 65 const __m128i v1_0 = _mm_unpacklo_epi32(v2_0, v2_1); in xnn_x32_transposec_ukernel__4x4_reuse_multi_sse2() local 100 const __m128i v1_0 = _mm_unpacklo_epi32(v2_0, v2_1); in xnn_x32_transposec_ukernel__4x4_reuse_multi_sse2() local
|
/aosp_15_r20/external/XNNPACK/src/x64-transposec/gen/ |
H A D | 2x2-reuse-multi-sse2.c | 48 const __m128i v1_0 = _mm_loadu_si128((const __m128i*) i0); in xnn_x64_transposec_ukernel__2x2_reuse_multi_sse2() local 65 const __m128i v1_0 = _mm_loadu_si128((const __m128i*) i0); in xnn_x64_transposec_ukernel__2x2_reuse_multi_sse2() local
|
H A D | 2x2-multi-multi-sse2.c | 50 const __m128i v1_0 = _mm_loadu_si128((const __m128i*) i0); in xnn_x64_transposec_ukernel__2x2_multi_multi_sse2() local 67 const __m128i v1_0 = _mm_loadu_si128((const __m128i*) i0); in xnn_x64_transposec_ukernel__2x2_multi_multi_sse2() local
|
H A D | 2x2-reuse-switch-sse2.c | 46 const __m128i v1_0 = _mm_loadu_si128((const __m128i*) i0); in xnn_x64_transposec_ukernel__2x2_reuse_switch_sse2() local 70 const __m128i v1_0 = _mm_loadu_si128((const __m128i*) i0); in xnn_x64_transposec_ukernel__2x2_reuse_switch_sse2() local
|
H A D | 2x2-multi-switch-sse2.c | 48 const __m128i v1_0 = _mm_loadu_si128((const __m128i*) i0); in xnn_x64_transposec_ukernel__2x2_multi_switch_sse2() local 72 const __m128i v1_0 = _mm_loadu_si128((const __m128i*) i0); in xnn_x64_transposec_ukernel__2x2_multi_switch_sse2() local
|
H A D | 2x2-reuse-mov-sse2.c | 48 const __m128i v1_0 = _mm_loadu_si128((const __m128i*) i0); in xnn_x64_transposec_ukernel__2x2_reuse_mov_sse2() local 69 const __m128i v1_0 = _mm_loadu_si128((const __m128i*) i0); in xnn_x64_transposec_ukernel__2x2_reuse_mov_sse2() local
|
H A D | 2x2-multi-mov-sse2.c | 50 const __m128i v1_0 = _mm_loadu_si128((const __m128i*) i0); in xnn_x64_transposec_ukernel__2x2_multi_mov_sse2() local 71 const __m128i v1_0 = _mm_loadu_si128((const __m128i*) i0); in xnn_x64_transposec_ukernel__2x2_multi_mov_sse2() local
|
/aosp_15_r20/external/XNNPACK/src/x16-transposec/gen/ |
H A D | 4x4-reuse-multi-zip-neon.c | 59 const uint16x4x2_t v1_0 = vzip_u16(v2_0, v2_2); in xnn_x16_transposec_ukernel__4x4_reuse_multi_zip_neon() local 85 const uint16x4x2_t v1_0 = vzip_u16(v2_0, v2_2); in xnn_x16_transposec_ukernel__4x4_reuse_multi_zip_neon() local
|
H A D | 4x4-multi-multi-zip-neon.c | 63 const uint16x4x2_t v1_0 = vzip_u16(v2_0, v2_2); in xnn_x16_transposec_ukernel__4x4_multi_multi_zip_neon() local 87 const uint16x4x2_t v1_0 = vzip_u16(v2_0, v2_2); in xnn_x16_transposec_ukernel__4x4_multi_multi_zip_neon() local
|
H A D | 4x4-reuse-dec-zip-neon.c | 51 const uint16x4x2_t v1_0 = vzip_u16(v2_0, v2_2); in xnn_x16_transposec_ukernel__4x4_reuse_dec_zip_neon() local 88 const uint16x4x2_t v1_0 = vzip_u16(v2_0, v2_2); in xnn_x16_transposec_ukernel__4x4_reuse_dec_zip_neon() local
|
H A D | 4x4-reuse-switch-zip-neon.c | 50 const uint16x4x2_t v1_0 = vzip_u16(v2_0, v2_2); in xnn_x16_transposec_ukernel__4x4_reuse_switch_zip_neon() local 86 const uint16x4x2_t v1_0 = vzip_u16(v2_0, v2_2); in xnn_x16_transposec_ukernel__4x4_reuse_switch_zip_neon() local
|