/external/XNNPACK/src/f32-gemm/gen/ |
D | 4x8-minmax-sse2-dup.c | 102 …const __m128 va3c1111 = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(va3), _MM_SHUFFLE(1, 1… in xnn_f32_gemm_minmax_ukernel_4x8__sse2_dup() local 110 vacc3x0123 = _mm_add_ps(vacc3x0123, _mm_mul_ps(va3c1111, vb0123c1)); in xnn_f32_gemm_minmax_ukernel_4x8__sse2_dup() 114 vacc3x4567 = _mm_add_ps(vacc3x4567, _mm_mul_ps(va3c1111, vb4567c1)); in xnn_f32_gemm_minmax_ukernel_4x8__sse2_dup()
|
D | 4x8-minmax-sse-dup.c | 102 const __m128 va3c1111 = _mm_shuffle_ps(va3, va3, _MM_SHUFFLE(1, 1, 1, 1)); in xnn_f32_gemm_minmax_ukernel_4x8__sse_dup() local 110 vacc3x0123 = _mm_add_ps(vacc3x0123, _mm_mul_ps(va3c1111, vb0123c1)); in xnn_f32_gemm_minmax_ukernel_4x8__sse_dup() 114 vacc3x4567 = _mm_add_ps(vacc3x4567, _mm_mul_ps(va3c1111, vb4567c1)); in xnn_f32_gemm_minmax_ukernel_4x8__sse_dup()
|
D | 5x8-minmax-sse-dup.c | 115 const __m128 va3c1111 = _mm_shuffle_ps(va3, va3, _MM_SHUFFLE(1, 1, 1, 1)); in xnn_f32_gemm_minmax_ukernel_5x8__sse_dup() local 124 vacc3x0123 = _mm_add_ps(vacc3x0123, _mm_mul_ps(va3c1111, vb0123c1)); in xnn_f32_gemm_minmax_ukernel_5x8__sse_dup() 129 vacc3x4567 = _mm_add_ps(vacc3x4567, _mm_mul_ps(va3c1111, vb4567c1)); in xnn_f32_gemm_minmax_ukernel_5x8__sse_dup()
|
D | 5x8-minmax-sse2-dup.c | 115 …const __m128 va3c1111 = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(va3), _MM_SHUFFLE(1, 1… in xnn_f32_gemm_minmax_ukernel_5x8__sse2_dup() local 124 vacc3x0123 = _mm_add_ps(vacc3x0123, _mm_mul_ps(va3c1111, vb0123c1)); in xnn_f32_gemm_minmax_ukernel_5x8__sse2_dup() 129 vacc3x4567 = _mm_add_ps(vacc3x4567, _mm_mul_ps(va3c1111, vb4567c1)); in xnn_f32_gemm_minmax_ukernel_5x8__sse2_dup()
|
/external/XNNPACK/src/f32-igemm/gen/ |
D | 4x8-minmax-sse-dup.c | 124 const __m128 va3c1111 = _mm_shuffle_ps(va3, va3, _MM_SHUFFLE(1, 1, 1, 1)); in xnn_f32_igemm_minmax_ukernel_4x8__sse_dup() local 132 vacc3x0123 = _mm_add_ps(vacc3x0123, _mm_mul_ps(va3c1111, vb0123c1)); in xnn_f32_igemm_minmax_ukernel_4x8__sse_dup() 136 vacc3x4567 = _mm_add_ps(vacc3x4567, _mm_mul_ps(va3c1111, vb4567c1)); in xnn_f32_igemm_minmax_ukernel_4x8__sse_dup()
|
D | 4x8-minmax-sse2-dup.c | 124 …const __m128 va3c1111 = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(va3), _MM_SHUFFLE(1, 1… in xnn_f32_igemm_minmax_ukernel_4x8__sse2_dup() local 132 vacc3x0123 = _mm_add_ps(vacc3x0123, _mm_mul_ps(va3c1111, vb0123c1)); in xnn_f32_igemm_minmax_ukernel_4x8__sse2_dup() 136 vacc3x4567 = _mm_add_ps(vacc3x4567, _mm_mul_ps(va3c1111, vb4567c1)); in xnn_f32_igemm_minmax_ukernel_4x8__sse2_dup()
|
D | 5x8-minmax-sse-dup.c | 140 const __m128 va3c1111 = _mm_shuffle_ps(va3, va3, _MM_SHUFFLE(1, 1, 1, 1)); in xnn_f32_igemm_minmax_ukernel_5x8__sse_dup() local 149 vacc3x0123 = _mm_add_ps(vacc3x0123, _mm_mul_ps(va3c1111, vb0123c1)); in xnn_f32_igemm_minmax_ukernel_5x8__sse_dup() 154 vacc3x4567 = _mm_add_ps(vacc3x4567, _mm_mul_ps(va3c1111, vb4567c1)); in xnn_f32_igemm_minmax_ukernel_5x8__sse_dup()
|
D | 5x8-minmax-sse2-dup.c | 140 …const __m128 va3c1111 = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(va3), _MM_SHUFFLE(1, 1… in xnn_f32_igemm_minmax_ukernel_5x8__sse2_dup() local 149 vacc3x0123 = _mm_add_ps(vacc3x0123, _mm_mul_ps(va3c1111, vb0123c1)); in xnn_f32_igemm_minmax_ukernel_5x8__sse2_dup() 154 vacc3x4567 = _mm_add_ps(vacc3x4567, _mm_mul_ps(va3c1111, vb4567c1)); in xnn_f32_igemm_minmax_ukernel_5x8__sse2_dup()
|
/external/XNNPACK/src/f32-gemm/gen-inc/ |
D | 4x8inc-minmax-sse-dup.c | 104 const __m128 va3c1111 = _mm_shuffle_ps(va3, va3, _MM_SHUFFLE(1, 1, 1, 1)); in xnn_f32_gemminc_minmax_ukernel_4x8__sse_dup() local 112 vacc3x0123 = _mm_add_ps(vacc3x0123, _mm_mul_ps(va3c1111, vb0123c1)); in xnn_f32_gemminc_minmax_ukernel_4x8__sse_dup() 116 vacc3x4567 = _mm_add_ps(vacc3x4567, _mm_mul_ps(va3c1111, vb4567c1)); in xnn_f32_gemminc_minmax_ukernel_4x8__sse_dup()
|
D | 4x8inc-minmax-sse2-dup.c | 104 …const __m128 va3c1111 = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(va3), _MM_SHUFFLE(1, 1… in xnn_f32_gemminc_minmax_ukernel_4x8__sse2_dup() local 112 vacc3x0123 = _mm_add_ps(vacc3x0123, _mm_mul_ps(va3c1111, vb0123c1)); in xnn_f32_gemminc_minmax_ukernel_4x8__sse2_dup() 116 vacc3x4567 = _mm_add_ps(vacc3x4567, _mm_mul_ps(va3c1111, vb4567c1)); in xnn_f32_gemminc_minmax_ukernel_4x8__sse2_dup()
|
D | 5x8inc-minmax-sse2-dup.c | 117 …const __m128 va3c1111 = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(va3), _MM_SHUFFLE(1, 1… in xnn_f32_gemminc_minmax_ukernel_5x8__sse2_dup() local 126 vacc3x0123 = _mm_add_ps(vacc3x0123, _mm_mul_ps(va3c1111, vb0123c1)); in xnn_f32_gemminc_minmax_ukernel_5x8__sse2_dup() 131 vacc3x4567 = _mm_add_ps(vacc3x4567, _mm_mul_ps(va3c1111, vb4567c1)); in xnn_f32_gemminc_minmax_ukernel_5x8__sse2_dup()
|
D | 5x8inc-minmax-sse-dup.c | 117 const __m128 va3c1111 = _mm_shuffle_ps(va3, va3, _MM_SHUFFLE(1, 1, 1, 1)); in xnn_f32_gemminc_minmax_ukernel_5x8__sse_dup() local 126 vacc3x0123 = _mm_add_ps(vacc3x0123, _mm_mul_ps(va3c1111, vb0123c1)); in xnn_f32_gemminc_minmax_ukernel_5x8__sse_dup() 131 vacc3x4567 = _mm_add_ps(vacc3x4567, _mm_mul_ps(va3c1111, vb4567c1)); in xnn_f32_gemminc_minmax_ukernel_5x8__sse_dup()
|