| /external/XNNPACK/src/qc8-dwconv/gen/ |
| D | up16x3-minmax-fp32-avx2-mul32.c | 61 …const __m256i vk0x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w … in xnn_qc8_dwconv_minmax_fp32_ukernel_up16x3__avx2_mul32() local
|
| D | up16x3-minmax-fp32-wasmsimd-mul16-add16.c | 62 …const v128_t vk0x89ABCDEF = wasm_i16x8_load8x8((const void*) ((uintptr_t) w + 16 * sizeof(int32_t)… in xnn_qc8_dwconv_minmax_fp32_ukernel_up16x3__wasmsimd_mul16_add16() local
|
| D | up16x3-minmax-fp32-neon-mla8-ld64.c | 65 const int8x8_t vk0x89ABCDEF = vld1_s8(w); w = (const void*) ((const int8_t*) w + 8); in xnn_qc8_dwconv_minmax_fp32_ukernel_up16x3__neon_mla8_ld64() local
|
| D | up16x3-minmax-fp32-neonv8-mla8-ld64.c | 65 const int8x8_t vk0x89ABCDEF = vld1_s8(w); w = (const void*) ((const int8_t*) w + 8); in xnn_qc8_dwconv_minmax_fp32_ukernel_up16x3__neonv8_mla8_ld64() local
|
| /external/XNNPACK/src/f32-dwconv/gen/ |
| D | up16x3-minmax-avx.c | 64 const __m256 vk0x89ABCDEF = _mm256_load_ps(w + 24); in xnn_f32_dwconv_minmax_ukernel_up16x3__avx() local
|
| D | up16x3-minmax-fma3.c | 64 const __m256 vk0x89ABCDEF = _mm256_load_ps(w + 24); in xnn_f32_dwconv_minmax_ukernel_up16x3__fma3() local
|
| D | up16x3-minmax-avx-acc2.c | 64 const __m256 vk0x89ABCDEF = _mm256_load_ps(w + 24); in xnn_f32_dwconv_minmax_ukernel_up16x3__avx_acc2() local
|
| D | up16x3-minmax-fma3-acc2.c | 64 const __m256 vk0x89ABCDEF = _mm256_load_ps(w + 24); in xnn_f32_dwconv_minmax_ukernel_up16x3__fma3_acc2() local
|
| D | up16x4-minmax-avx.c | 69 const __m256 vk0x89ABCDEF = _mm256_load_ps(w + 24); in xnn_f32_dwconv_minmax_ukernel_up16x4__avx() local
|
| D | up16x4-minmax-fma3.c | 69 const __m256 vk0x89ABCDEF = _mm256_load_ps(w + 24); in xnn_f32_dwconv_minmax_ukernel_up16x4__fma3() local
|
| D | up16x4-minmax-fma3-acc2.c | 69 const __m256 vk0x89ABCDEF = _mm256_load_ps(w + 24); in xnn_f32_dwconv_minmax_ukernel_up16x4__fma3_acc2() local
|
| D | up16x4-minmax-avx-acc2.c | 69 const __m256 vk0x89ABCDEF = _mm256_load_ps(w + 24); in xnn_f32_dwconv_minmax_ukernel_up16x4__avx_acc2() local
|
| /external/XNNPACK/src/f16-dwconv/gen/ |
| D | up16x3-minmax-fma3.c | 67 const __m256 vk0x89ABCDEF = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (w + 24))); in xnn_f16_dwconv_minmax_ukernel_up16x3__fma3() local
|
| D | up16x3-minmax-neonfp16arith.c | 64 const float16x8_t vk0x89ABCDEF = vld1q_f16(w); w += 8; in xnn_f16_dwconv_minmax_ukernel_up16x3__neonfp16arith() local
|
| D | up16x3-minmax-fma3-acc2.c | 67 const __m256 vk0x89ABCDEF = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (w + 24))); in xnn_f16_dwconv_minmax_ukernel_up16x3__fma3_acc2() local
|
| D | up16x3-minmax-neonfp16arith-acc2.c | 64 const float16x8_t vk0x89ABCDEF = vld1q_f16(w); w += 8; in xnn_f16_dwconv_minmax_ukernel_up16x3__neonfp16arith_acc2() local
|
| D | up16x4-minmax-fma3.c | 72 const __m256 vk0x89ABCDEF = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (w + 24))); in xnn_f16_dwconv_minmax_ukernel_up16x4__fma3() local
|
| D | up16x4-minmax-neonfp16arith.c | 69 const float16x8_t vk0x89ABCDEF = vld1q_f16(w); w += 8; in xnn_f16_dwconv_minmax_ukernel_up16x4__neonfp16arith() local
|
| D | up16x4-minmax-neonfp16arith-acc2.c | 69 const float16x8_t vk0x89ABCDEF = vld1q_f16(w); w += 8; in xnn_f16_dwconv_minmax_ukernel_up16x4__neonfp16arith_acc2() local
|
| D | up16x4-minmax-fma3-acc2.c | 72 const __m256 vk0x89ABCDEF = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (w + 24))); in xnn_f16_dwconv_minmax_ukernel_up16x4__fma3_acc2() local
|
| D | up32x3-minmax-neonfp16arith.c | 68 const float16x8_t vk0x89ABCDEF = vld1q_f16(w); w += 8; in xnn_f16_dwconv_minmax_ukernel_up32x3__neonfp16arith() local
|
| D | up32x3-minmax-fma3.c | 71 const __m256 vk0x89ABCDEF = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (w + 40))); in xnn_f16_dwconv_minmax_ukernel_up32x3__fma3() local
|
| D | up32x3-minmax-fma3-acc2.c | 71 const __m256 vk0x89ABCDEF = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (w + 40))); in xnn_f16_dwconv_minmax_ukernel_up32x3__fma3_acc2() local
|
| D | up32x3-minmax-neonfp16arith-acc2.c | 68 const float16x8_t vk0x89ABCDEF = vld1q_f16(w); w += 8; in xnn_f16_dwconv_minmax_ukernel_up32x3__neonfp16arith_acc2() local
|
| D | up32x4-minmax-neonfp16arith.c | 73 const float16x8_t vk0x89ABCDEF = vld1q_f16(w); w += 8; in xnn_f16_dwconv_minmax_ukernel_up32x4__neonfp16arith() local
|