/dports/graphics/vapoursynth-waifu2x-ncnn-vulkan/vapoursynth-waifu2x-ncnn-vulkan-r4/deps/ncnn/src/layer/arm/ |
H A D | convolution_3x3_pack4to1_bf16s.h | 113 float32x4_t _tmp12a = vmlsq_n_f32(vaddq_f32(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 184 float32x4_t _tmp12a = vmlsq_n_f32(vaddq_f32(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local
|
H A D | convolution_3x3_pack8to1_fp16s.h | 229 float16x8_t _tmp12a = vfmsq_n_f16(vaddq_f16(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 300 float16x8_t _tmp12a = vfmsq_n_f16(vaddq_f16(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local
|
H A D | convolution_3x3_pack8to4_fp16s.h | 235 float16x8_t _tmp12a = vfmsq_n_f16(vaddq_f16(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 306 float16x8_t _tmp12a = vfmsq_n_f16(vaddq_f16(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4_fp16s.h | 335 float16x4_t _tmp12a = vfms_n_f16(vadd_f16(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local 406 float16x4_t _tmp12a = vfms_n_f16(vadd_f16(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4to1.h | 370 float32x4_t _tmp12a = vmlsq_n_f32(vaddq_f32(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack4to1_neon() local 441 float32x4_t _tmp12a = vmlsq_n_f32(vaddq_f32(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack4to1_neon() local
|
/dports/graphics/waifu2x-ncnn-vulkan/waifu2x-ncnn-vulkan-20210521/src/ncnn/src/layer/arm/ |
H A D | convolution_3x3_pack4to1_bf16s.h | 113 float32x4_t _tmp12a = vmlsq_n_f32(vaddq_f32(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 184 float32x4_t _tmp12a = vmlsq_n_f32(vaddq_f32(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local
|
H A D | convolution_3x3_pack8to1_fp16s.h | 229 float16x8_t _tmp12a = vfmsq_n_f16(vaddq_f16(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 300 float16x8_t _tmp12a = vfmsq_n_f16(vaddq_f16(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local
|
H A D | convolution_3x3_pack8to4_fp16s.h | 235 float16x8_t _tmp12a = vfmsq_n_f16(vaddq_f16(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 306 float16x8_t _tmp12a = vfmsq_n_f16(vaddq_f16(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4_fp16s.h | 335 float16x4_t _tmp12a = vfms_n_f16(vadd_f16(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local 406 float16x4_t _tmp12a = vfms_n_f16(vadd_f16(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4to1.h | 370 float32x4_t _tmp12a = vmlsq_n_f32(vaddq_f32(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack4to1_neon() local 441 float32x4_t _tmp12a = vmlsq_n_f32(vaddq_f32(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack4to1_neon() local
|
/dports/benchmarks/vkpeak/vkpeak-20210430/ncnn/src/layer/arm/ |
H A D | convolution_3x3_pack4to1_bf16s.h | 113 float32x4_t _tmp12a = vmlsq_n_f32(vaddq_f32(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 184 float32x4_t _tmp12a = vmlsq_n_f32(vaddq_f32(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local
|
H A D | convolution_3x3_pack8to1_fp16s.h | 229 float16x8_t _tmp12a = vfmsq_n_f16(vaddq_f16(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 300 float16x8_t _tmp12a = vfmsq_n_f16(vaddq_f16(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local
|
H A D | convolution_3x3_pack8to4_fp16s.h | 235 float16x8_t _tmp12a = vfmsq_n_f16(vaddq_f16(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 306 float16x8_t _tmp12a = vfmsq_n_f16(vaddq_f16(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4_fp16s.h | 335 float16x4_t _tmp12a = vfms_n_f16(vadd_f16(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local 406 float16x4_t _tmp12a = vfms_n_f16(vadd_f16(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4to1.h | 370 float32x4_t _tmp12a = vmlsq_n_f32(vaddq_f32(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack4to1_neon() local 441 float32x4_t _tmp12a = vmlsq_n_f32(vaddq_f32(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack4to1_neon() local
|
/dports/misc/ncnn/ncnn-20211208/src/layer/arm/ |
H A D | convolution_3x3_pack4to1_bf16s.h | 113 float32x4_t _tmp12a = vmlsq_n_f32(vaddq_f32(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 184 float32x4_t _tmp12a = vmlsq_n_f32(vaddq_f32(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local
|
H A D | convolution_3x3_pack8to1_fp16s.h | 229 float16x8_t _tmp12a = vfmsq_n_f16(vaddq_f16(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 300 float16x8_t _tmp12a = vfmsq_n_f16(vaddq_f16(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local
|
H A D | convolution_3x3_pack8to4_fp16s.h | 235 float16x8_t _tmp12a = vfmsq_n_f16(vaddq_f16(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 306 float16x8_t _tmp12a = vfmsq_n_f16(vaddq_f16(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4_fp16s.h | 335 float16x4_t _tmp12a = vfms_n_f16(vadd_f16(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local 406 float16x4_t _tmp12a = vfms_n_f16(vadd_f16(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4to1.h | 370 float32x4_t _tmp12a = vmlsq_n_f32(vaddq_f32(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack4to1_neon() local 441 float32x4_t _tmp12a = vmlsq_n_f32(vaddq_f32(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack4to1_neon() local
|
/dports/graphics/realsr-ncnn-vulkan/realsr-ncnn-vulkan-20210210/src/ncnn/src/layer/arm/ |
H A D | convolution_3x3_pack4to1_bf16s.h | 113 float32x4_t _tmp12a = vmlsq_n_f32(vaddq_f32(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 184 float32x4_t _tmp12a = vmlsq_n_f32(vaddq_f32(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local
|
H A D | convolution_3x3_pack8to1_fp16s.h | 229 float16x8_t _tmp12a = vfmsq_n_f16(vaddq_f16(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 300 float16x8_t _tmp12a = vfmsq_n_f16(vaddq_f16(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local
|
H A D | convolution_3x3_pack8to4_fp16s.h | 235 float16x8_t _tmp12a = vfmsq_n_f16(vaddq_f16(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 306 float16x8_t _tmp12a = vfmsq_n_f16(vaddq_f16(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4_fp16s.h | 335 float16x4_t _tmp12a = vfms_n_f16(vadd_f16(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local 406 float16x4_t _tmp12a = vfms_n_f16(vadd_f16(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local
|
/dports/graphics/realsr-ncnn-vulkan/realsr-ncnn-vulkan-20210210/src/ncnn/src/layer/x86/ |
H A D | convolution_3x3_pack8.h | 336 __m256 _tmp12a = _mm256_fmrsub_1_ps(_mm256_add_ps(_r02, _r06), _r04, 4.25f); in conv3x3s1_winograd64_pack8_avx() local 407 … __m256 _tmp12a = _mm256_fmrsub_1_ps(_mm256_add_ps(_tmp02, _tmp06), _tmp04, 4.25f); in conv3x3s1_winograd64_pack8_avx() local
|