/external/XNNPACK/src/x16-transposec/gen/ |
D | 8x8-reuse-dec-zip-neon.c | 53 const uint16x8_t v3_7 = vld1q_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_stride); in xnn_x16_transposec_ukernel__8x8_reuse_dec_zip_neon() local 134 const uint16x8_t v3_7 = vmovq_n_u16(0); in xnn_x16_transposec_ukernel__8x8_reuse_dec_zip_neon() local
|
D | 8x8-multi-dec-zip-neon.c | 61 const uint16x8_t v3_7 = vld1q_u16(i7); i7 = (uint16_t*) ((uintptr_t) i7 + input_offset); in xnn_x16_transposec_ukernel__8x8_multi_dec_zip_neon() local 136 const uint16x8_t v3_7 = vmovq_n_u16(0); in xnn_x16_transposec_ukernel__8x8_multi_dec_zip_neon() local
|
D | 8x8-reuse-switch-zip-neon.c | 52 const uint16x8_t v3_7 = vld1q_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_stride); in xnn_x16_transposec_ukernel__8x8_reuse_switch_zip_neon() local 124 const uint16x8_t v3_7 = vmovq_n_u16(0); in xnn_x16_transposec_ukernel__8x8_reuse_switch_zip_neon() local
|
D | 8x8-reuse-mov-zip-neon.c | 53 const uint16x8_t v3_7 = vld1q_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_stride); in xnn_x16_transposec_ukernel__8x8_reuse_mov_zip_neon() local 141 const uint16x8_t v3_7 = vmovq_n_u16(0); in xnn_x16_transposec_ukernel__8x8_reuse_mov_zip_neon() local
|
D | 8x8-reuse-multi-zip-neon.c | 77 const uint16x8_t v3_7 = vld1q_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_stride); in xnn_x16_transposec_ukernel__8x8_reuse_multi_zip_neon() local 135 const uint16x8_t v3_7 = vmovq_n_u16(0); in xnn_x16_transposec_ukernel__8x8_reuse_multi_zip_neon() local
|
D | 8x8-multi-switch-zip-neon.c | 60 const uint16x8_t v3_7 = vld1q_u16(i7); i7 = (uint16_t*) ((uintptr_t) i7 + input_offset); in xnn_x16_transposec_ukernel__8x8_multi_switch_zip_neon() local 126 const uint16x8_t v3_7 = vmovq_n_u16(0); in xnn_x16_transposec_ukernel__8x8_multi_switch_zip_neon() local
|
D | 8x8-multi-mov-zip-neon.c | 61 const uint16x8_t v3_7 = vld1q_u16(i7); i7 = (uint16_t*) ((uintptr_t) i7 + input_offset); in xnn_x16_transposec_ukernel__8x8_multi_mov_zip_neon() local 143 const uint16x8_t v3_7 = vmovq_n_u16(0); in xnn_x16_transposec_ukernel__8x8_multi_mov_zip_neon() local
|
D | 8x8-reuse-mov-wasmsimd.c | 60 const v128_t v3_7 = wasm_v128_load(i0); in xnn_x16_transposec_ukernel__8x8_reuse_mov_wasmsimd() local 160 const v128_t v3_7 = wasm_v128_xor(v3_0, v3_0); in xnn_x16_transposec_ukernel__8x8_reuse_mov_wasmsimd() local
|
D | 8x8-multi-switch-wasmsimd.c | 67 const v128_t v3_7 = wasm_v128_load(i7); in xnn_x16_transposec_ukernel__8x8_multi_switch_wasmsimd() local 152 const v128_t v3_7 = wasm_v128_xor(v3_0, v3_0); in xnn_x16_transposec_ukernel__8x8_multi_switch_wasmsimd() local
|
D | 8x8-reuse-multi-sse2.c | 86 const __m128i v3_7 = _mm_loadu_si128((const __m128i*) i0); in xnn_x16_transposec_ukernel__8x8_reuse_multi_sse2() local 166 const __m128i v3_7 = _mm_undefined_si128(); in xnn_x16_transposec_ukernel__8x8_reuse_multi_sse2() local
|
D | 8x8-reuse-switch-sse2.c | 61 const __m128i v3_7 = _mm_loadu_si128((const __m128i*) i0); in xnn_x16_transposec_ukernel__8x8_reuse_switch_sse2() local 154 const __m128i v3_7 = _mm_undefined_si128(); in xnn_x16_transposec_ukernel__8x8_reuse_switch_sse2() local
|
D | 8x8-reuse-multi-wasmsimd.c | 84 const v128_t v3_7 = wasm_v128_load(i0); in xnn_x16_transposec_ukernel__8x8_reuse_multi_wasmsimd() local 162 const v128_t v3_7 = wasm_v128_xor(v3_0, v3_0); in xnn_x16_transposec_ukernel__8x8_reuse_multi_wasmsimd() local
|
D | 8x8-reuse-switch-wasmsimd.c | 59 const v128_t v3_7 = wasm_v128_load(i0); in xnn_x16_transposec_ukernel__8x8_reuse_switch_wasmsimd() local 150 const v128_t v3_7 = wasm_v128_xor(v3_0, v3_0); in xnn_x16_transposec_ukernel__8x8_reuse_switch_wasmsimd() local
|
D | 8x8-multi-switch-sse2.c | 69 const __m128i v3_7 = _mm_loadu_si128((const __m128i*) i7); in xnn_x16_transposec_ukernel__8x8_multi_switch_sse2() local 156 const __m128i v3_7 = _mm_undefined_si128(); in xnn_x16_transposec_ukernel__8x8_multi_switch_sse2() local
|
D | 8x8-reuse-mov-sse2.c | 62 const __m128i v3_7 = _mm_loadu_si128((const __m128i*) i0); in xnn_x16_transposec_ukernel__8x8_reuse_mov_sse2() local 164 const __m128i v3_7 = _mm_undefined_si128(); in xnn_x16_transposec_ukernel__8x8_reuse_mov_sse2() local
|
D | 8x8-multi-mov-sse2.c | 70 const __m128i v3_7 = _mm_loadu_si128((const __m128i*) i7); in xnn_x16_transposec_ukernel__8x8_multi_mov_sse2() local 166 const __m128i v3_7 = _mm_undefined_si128(); in xnn_x16_transposec_ukernel__8x8_multi_mov_sse2() local
|
D | 8x8-multi-mov-wasmsimd.c | 68 const v128_t v3_7 = wasm_v128_load(i7); in xnn_x16_transposec_ukernel__8x8_multi_mov_wasmsimd() local 162 const v128_t v3_7 = wasm_v128_xor(v3_0, v3_0); in xnn_x16_transposec_ukernel__8x8_multi_mov_wasmsimd() local
|
/external/XNNPACK/src/x8-transposec/gen/ |
D | 8x8-reuse-dec-zip-neon.c | 53 const uint8x8_t v3_7 = vld1_u8(i0); i0 = (uint8_t*) ((uintptr_t) i0 + input_stride); in xnn_x8_transposec_ukernel__8x8_reuse_dec_zip_neon() local 134 const uint8x8_t v3_7 = vmov_n_u8(0); in xnn_x8_transposec_ukernel__8x8_reuse_dec_zip_neon() local
|
D | 8x8-multi-switch-zip-neon.c | 60 const uint8x8_t v3_7 = vld1_u8(i7); i7 = (uint8_t*) ((uintptr_t) i7 + input_offset); in xnn_x8_transposec_ukernel__8x8_multi_switch_zip_neon() local 126 const uint8x8_t v3_7 = vmov_n_u8(0); in xnn_x8_transposec_ukernel__8x8_multi_switch_zip_neon() local
|
D | 8x8-multi-dec-zip-neon.c | 61 const uint8x8_t v3_7 = vld1_u8(i7); i7 = (uint8_t*) ((uintptr_t) i7 + input_offset); in xnn_x8_transposec_ukernel__8x8_multi_dec_zip_neon() local 136 const uint8x8_t v3_7 = vmov_n_u8(0); in xnn_x8_transposec_ukernel__8x8_multi_dec_zip_neon() local
|
D | 8x8-reuse-switch-zip-neon.c | 52 const uint8x8_t v3_7 = vld1_u8(i0); i0 = (uint8_t*) ((uintptr_t) i0 + input_stride); in xnn_x8_transposec_ukernel__8x8_reuse_switch_zip_neon() local 124 const uint8x8_t v3_7 = vmov_n_u8(0); in xnn_x8_transposec_ukernel__8x8_reuse_switch_zip_neon() local
|
D | 8x8-reuse-mov-zip-neon.c | 53 const uint8x8_t v3_7 = vld1_u8(i0); i0 = (uint8_t*) ((uintptr_t) i0 + input_stride); in xnn_x8_transposec_ukernel__8x8_reuse_mov_zip_neon() local 141 const uint8x8_t v3_7 = vmov_n_u8(0); in xnn_x8_transposec_ukernel__8x8_reuse_mov_zip_neon() local
|
D | 8x8-reuse-multi-zip-neon.c | 77 const uint8x8_t v3_7 = vld1_u8(i0); i0 = (uint8_t*) ((uintptr_t) i0 + input_stride); in xnn_x8_transposec_ukernel__8x8_reuse_multi_zip_neon() local 135 const uint8x8_t v3_7 = vmov_n_u8(0); in xnn_x8_transposec_ukernel__8x8_reuse_multi_zip_neon() local
|
D | 8x8-multi-mov-zip-neon.c | 61 const uint8x8_t v3_7 = vld1_u8(i7); i7 = (uint8_t*) ((uintptr_t) i7 + input_offset); in xnn_x8_transposec_ukernel__8x8_multi_mov_zip_neon() local 143 const uint8x8_t v3_7 = vmov_n_u8(0); in xnn_x8_transposec_ukernel__8x8_multi_mov_zip_neon() local
|
D | 16x16-reuse-mov-zip-neon.c | 70 const uint8x16x2_t v3_7 = vzipq_u8(v4_7, v4_15); in xnn_x8_transposec_ukernel__16x16_reuse_mov_zip_neon() local 258 const uint8x16x2_t v3_7 = vzipq_u8(v4_7, v4_15); in xnn_x8_transposec_ukernel__16x16_reuse_mov_zip_neon() local
|