/dports/graphics/waifu2x-ncnn-vulkan/waifu2x-ncnn-vulkan-20210521/src/ncnn/src/layer/arm/ |
H A D | convolution_3x3.h | 5963 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm0, _bb2p0, 0); in conv3x3s1_winograd64_neon5() 5965 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm2, _bb2p0, 1); in conv3x3s1_winograd64_neon5() 5975 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm4, _bb2p0, 2); in conv3x3s1_winograd64_neon5() 5977 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm6, _bb2p0, 3); in conv3x3s1_winograd64_neon5() 5987 _sum0123 = vmlaq_f32(_sum0123, _bb2p0, _ktm0123); in conv3x3s1_winograd64_neon5() 6703 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm0, _bb2p0, 0); in conv3x3s1_winograd64_neon5() 6704 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm1, _bb2p0, 1); in conv3x3s1_winograd64_neon5() 6705 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm2, _bb2p0, 2); in conv3x3s1_winograd64_neon5() 6706 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm3, _bb2p0, 3); in conv3x3s1_winograd64_neon5() 6708 _sum0123 = vmlaq_lane_f32(_sum0123, _ktm0, vget_low_f32(_bb2p0), 0); in conv3x3s1_winograd64_neon5() [all …]
|
/dports/benchmarks/vkpeak/vkpeak-20210430/ncnn/src/layer/arm/ |
H A D | convolution_3x3.h | 5963 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm0, _bb2p0, 0); in conv3x3s1_winograd64_neon5() 5965 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm2, _bb2p0, 1); in conv3x3s1_winograd64_neon5() 5975 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm4, _bb2p0, 2); in conv3x3s1_winograd64_neon5() 5977 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm6, _bb2p0, 3); in conv3x3s1_winograd64_neon5() 5987 _sum0123 = vmlaq_f32(_sum0123, _bb2p0, _ktm0123); in conv3x3s1_winograd64_neon5() 6703 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm0, _bb2p0, 0); in conv3x3s1_winograd64_neon5() 6704 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm1, _bb2p0, 1); in conv3x3s1_winograd64_neon5() 6705 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm2, _bb2p0, 2); in conv3x3s1_winograd64_neon5() 6706 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm3, _bb2p0, 3); in conv3x3s1_winograd64_neon5() 6708 _sum0123 = vmlaq_lane_f32(_sum0123, _ktm0, vget_low_f32(_bb2p0), 0); in conv3x3s1_winograd64_neon5() [all …]
|
/dports/misc/ncnn/ncnn-20211208/src/layer/arm/ |
H A D | convolution_3x3.h | 5963 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm0, _bb2p0, 0); in conv3x3s1_winograd64_neon5() 5965 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm2, _bb2p0, 1); in conv3x3s1_winograd64_neon5() 5975 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm4, _bb2p0, 2); in conv3x3s1_winograd64_neon5() 5977 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm6, _bb2p0, 3); in conv3x3s1_winograd64_neon5() 5987 _sum0123 = vmlaq_f32(_sum0123, _bb2p0, _ktm0123); in conv3x3s1_winograd64_neon5() 6703 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm0, _bb2p0, 0); in conv3x3s1_winograd64_neon5() 6704 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm1, _bb2p0, 1); in conv3x3s1_winograd64_neon5() 6705 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm2, _bb2p0, 2); in conv3x3s1_winograd64_neon5() 6706 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm3, _bb2p0, 3); in conv3x3s1_winograd64_neon5() 6708 _sum0123 = vmlaq_lane_f32(_sum0123, _ktm0, vget_low_f32(_bb2p0), 0); in conv3x3s1_winograd64_neon5() [all …]
|
/dports/graphics/vapoursynth-waifu2x-ncnn-vulkan/vapoursynth-waifu2x-ncnn-vulkan-r4/deps/ncnn/src/layer/arm/ |
H A D | convolution_3x3.h | 5963 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm0, _bb2p0, 0); in conv3x3s1_winograd64_neon5() 5965 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm2, _bb2p0, 1); in conv3x3s1_winograd64_neon5() 5975 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm4, _bb2p0, 2); in conv3x3s1_winograd64_neon5() 5977 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm6, _bb2p0, 3); in conv3x3s1_winograd64_neon5() 5987 _sum0123 = vmlaq_f32(_sum0123, _bb2p0, _ktm0123); in conv3x3s1_winograd64_neon5() 6703 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm0, _bb2p0, 0); in conv3x3s1_winograd64_neon5() 6704 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm1, _bb2p0, 1); in conv3x3s1_winograd64_neon5() 6705 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm2, _bb2p0, 2); in conv3x3s1_winograd64_neon5() 6706 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm3, _bb2p0, 3); in conv3x3s1_winograd64_neon5() 6708 _sum0123 = vmlaq_lane_f32(_sum0123, _ktm0, vget_low_f32(_bb2p0), 0); in conv3x3s1_winograd64_neon5() [all …]
|
/dports/graphics/realsr-ncnn-vulkan/realsr-ncnn-vulkan-20210210/src/ncnn/src/layer/arm/ |
H A D | convolution_3x3.h | 5963 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm0, _bb2p0, 0); in conv3x3s1_winograd64_neon5() 5965 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm2, _bb2p0, 1); in conv3x3s1_winograd64_neon5() 5975 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm4, _bb2p0, 2); in conv3x3s1_winograd64_neon5() 5977 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm6, _bb2p0, 3); in conv3x3s1_winograd64_neon5() 5987 _sum0123 = vmlaq_f32(_sum0123, _bb2p0, _ktm0123); in conv3x3s1_winograd64_neon5() 6703 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm0, _bb2p0, 0); in conv3x3s1_winograd64_neon5() 6704 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm1, _bb2p0, 1); in conv3x3s1_winograd64_neon5() 6705 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm2, _bb2p0, 2); in conv3x3s1_winograd64_neon5() 6706 _sum0123 = vmlaq_laneq_f32(_sum0123, _ktm3, _bb2p0, 3); in conv3x3s1_winograd64_neon5() 6708 _sum0123 = vmlaq_lane_f32(_sum0123, _ktm0, vget_low_f32(_bb2p0), 0); in conv3x3s1_winograd64_neon5() [all …]
|
/dports/graphics/vapoursynth-waifu2x-ncnn-vulkan/vapoursynth-waifu2x-ncnn-vulkan-r4/deps/ncnn/src/layer/x86/ |
H A D | x86_usability.h | 263 __m256 _sum0123 = _mm256_add_ps(_sum01, _sum23); in _mm256_fmadd_ps4() local 264 _sum = _mm256_add_ps(_sum, _sum0123); in _mm256_fmadd_ps4()
|
/dports/graphics/waifu2x-ncnn-vulkan/waifu2x-ncnn-vulkan-20210521/src/ncnn/src/layer/x86/ |
H A D | x86_usability.h | 263 __m256 _sum0123 = _mm256_add_ps(_sum01, _sum23); in _mm256_fmadd_ps4() local 264 _sum = _mm256_add_ps(_sum, _sum0123); in _mm256_fmadd_ps4()
|
/dports/benchmarks/vkpeak/vkpeak-20210430/ncnn/src/layer/x86/ |
H A D | x86_usability.h | 263 __m256 _sum0123 = _mm256_add_ps(_sum01, _sum23); in _mm256_fmadd_ps4() local 264 _sum = _mm256_add_ps(_sum, _sum0123); in _mm256_fmadd_ps4()
|
/dports/misc/ncnn/ncnn-20211208/src/layer/x86/ |
H A D | x86_usability.h | 316 __m256 _sum0123 = _mm256_add_ps(_sum01, _sum23); in _mm256_comp_fmadd_ps4() local 317 _sum = _mm256_add_ps(_sum, _sum0123); in _mm256_comp_fmadd_ps4()
|