/dports/graphics/vapoursynth-waifu2x-ncnn-vulkan/vapoursynth-waifu2x-ncnn-vulkan-r4/deps/ncnn/src/layer/arm/ |
H A D | convolution_3x3_pack4to1_bf16s.h | 127 … float32x4_t _tmp34a = vmlsq_n_f32(vmlaq_n_f32(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 196 … float32x4_t _tmp34a = vmlsq_n_f32(vmlaq_n_f32(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local
|
H A D | convolution_3x3_pack8to1_fp16s.h | 243 … float16x8_t _tmp34a = vfmsq_n_f16(vfmaq_n_f16(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 312 … float16x8_t _tmp34a = vfmsq_n_f16(vfmaq_n_f16(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local
|
H A D | convolution_3x3_pack8to4_fp16s.h | 249 … float16x8_t _tmp34a = vfmsq_n_f16(vfmaq_n_f16(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 318 … float16x8_t _tmp34a = vfmsq_n_f16(vfmaq_n_f16(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4_fp16s.h | 349 … float16x4_t _tmp34a = vfms_n_f16(vfma_n_f16(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local 418 … float16x4_t _tmp34a = vfms_n_f16(vfma_n_f16(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4to1.h | 384 … float32x4_t _tmp34a = vmlsq_n_f32(vmlaq_n_f32(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack4to1_neon() local 453 … float32x4_t _tmp34a = vmlsq_n_f32(vmlaq_n_f32(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); 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 | 127 … float32x4_t _tmp34a = vmlsq_n_f32(vmlaq_n_f32(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 196 … float32x4_t _tmp34a = vmlsq_n_f32(vmlaq_n_f32(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local
|
H A D | convolution_3x3_pack8to1_fp16s.h | 243 … float16x8_t _tmp34a = vfmsq_n_f16(vfmaq_n_f16(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 312 … float16x8_t _tmp34a = vfmsq_n_f16(vfmaq_n_f16(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local
|
H A D | convolution_3x3_pack8to4_fp16s.h | 249 … float16x8_t _tmp34a = vfmsq_n_f16(vfmaq_n_f16(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 318 … float16x8_t _tmp34a = vfmsq_n_f16(vfmaq_n_f16(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4_fp16s.h | 349 … float16x4_t _tmp34a = vfms_n_f16(vfma_n_f16(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local 418 … float16x4_t _tmp34a = vfms_n_f16(vfma_n_f16(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4to1.h | 384 … float32x4_t _tmp34a = vmlsq_n_f32(vmlaq_n_f32(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack4to1_neon() local 453 … float32x4_t _tmp34a = vmlsq_n_f32(vmlaq_n_f32(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); in conv3x3s1_winograd64_pack4to1_neon() local
|
/dports/benchmarks/vkpeak/vkpeak-20210430/ncnn/src/layer/arm/ |
H A D | convolution_3x3_pack4to1_bf16s.h | 127 … float32x4_t _tmp34a = vmlsq_n_f32(vmlaq_n_f32(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 196 … float32x4_t _tmp34a = vmlsq_n_f32(vmlaq_n_f32(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local
|
H A D | convolution_3x3_pack8to1_fp16s.h | 243 … float16x8_t _tmp34a = vfmsq_n_f16(vfmaq_n_f16(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 312 … float16x8_t _tmp34a = vfmsq_n_f16(vfmaq_n_f16(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local
|
H A D | convolution_3x3_pack8to4_fp16s.h | 249 … float16x8_t _tmp34a = vfmsq_n_f16(vfmaq_n_f16(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 318 … float16x8_t _tmp34a = vfmsq_n_f16(vfmaq_n_f16(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4_fp16s.h | 349 … float16x4_t _tmp34a = vfms_n_f16(vfma_n_f16(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local 418 … float16x4_t _tmp34a = vfms_n_f16(vfma_n_f16(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4to1.h | 384 … float32x4_t _tmp34a = vmlsq_n_f32(vmlaq_n_f32(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack4to1_neon() local 453 … float32x4_t _tmp34a = vmlsq_n_f32(vmlaq_n_f32(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); in conv3x3s1_winograd64_pack4to1_neon() local
|
/dports/misc/ncnn/ncnn-20211208/src/layer/arm/ |
H A D | convolution_3x3_pack4to1_bf16s.h | 127 … float32x4_t _tmp34a = vmlsq_n_f32(vmlaq_n_f32(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 196 … float32x4_t _tmp34a = vmlsq_n_f32(vmlaq_n_f32(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local
|
H A D | convolution_3x3_pack8to1_fp16s.h | 243 … float16x8_t _tmp34a = vfmsq_n_f16(vfmaq_n_f16(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 312 … float16x8_t _tmp34a = vfmsq_n_f16(vfmaq_n_f16(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local
|
H A D | convolution_3x3_pack8to4_fp16s.h | 249 … float16x8_t _tmp34a = vfmsq_n_f16(vfmaq_n_f16(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 318 … float16x8_t _tmp34a = vfmsq_n_f16(vfmaq_n_f16(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4_fp16s.h | 349 … float16x4_t _tmp34a = vfms_n_f16(vfma_n_f16(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local 418 … float16x4_t _tmp34a = vfms_n_f16(vfma_n_f16(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4to1.h | 384 … float32x4_t _tmp34a = vmlsq_n_f32(vmlaq_n_f32(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack4to1_neon() local 453 … float32x4_t _tmp34a = vmlsq_n_f32(vmlaq_n_f32(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); 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 | 127 … float32x4_t _tmp34a = vmlsq_n_f32(vmlaq_n_f32(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 196 … float32x4_t _tmp34a = vmlsq_n_f32(vmlaq_n_f32(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local
|
H A D | convolution_3x3_pack8to1_fp16s.h | 243 … float16x8_t _tmp34a = vfmsq_n_f16(vfmaq_n_f16(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 312 … float16x8_t _tmp34a = vfmsq_n_f16(vfmaq_n_f16(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local
|
H A D | convolution_3x3_pack8to4_fp16s.h | 249 … float16x8_t _tmp34a = vfmsq_n_f16(vfmaq_n_f16(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 318 … float16x8_t _tmp34a = vfmsq_n_f16(vfmaq_n_f16(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4_fp16s.h | 349 … float16x4_t _tmp34a = vfms_n_f16(vfma_n_f16(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local 418 … float16x4_t _tmp34a = vfms_n_f16(vfma_n_f16(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); 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 | 350 … __m256 _tmp34a = _mm256_fmrsub_1_ps(_mm256_fmadd_1_ps(_r06, _r02, 0.25f), _r04, 1.25f); in conv3x3s1_winograd64_pack8_avx() local 419 … __m256 _tmp34a = _mm256_fmrsub_1_ps(_mm256_fmadd_1_ps(_tmp06, _tmp02, 0.25f), _tmp04, 1.25f); in conv3x3s1_winograd64_pack8_avx() local
|