/dports/graphics/waifu2x-ncnn-vulkan/waifu2x-ncnn-vulkan-20210521/src/ncnn/src/layer/arm/ |
H A D | convolution_3x3.h | 2717 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_neon4() local 2724 out0_tm.fill(0.f); in conv3x3s1_winograd64_neon4() 2739 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon4() 3128 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon4() 3699 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon4() 3879 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_neon4() local 3883 out0_tm.fill(0.f); in conv3x3s1_winograd64_neon4() 3891 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon4() 5600 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon5() 6043 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon5() [all …]
|
H A D | convolution_3x3_pack8_fp16s.h | 1059 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_pack8_fp16sa_neon() local 1074 const __fp16* output0_tm_0 = (const __fp16*)out0_tm + (i * w_tm / 8 + j) * 8; in conv3x3s1_winograd64_pack8_fp16sa_neon() 2152 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd42_pack8_fp16sa_neon() local 2167 const __fp16* output0_tm_0 = (const __fp16*)out0_tm + (i * w_tm / 6 + j) * 8; in conv3x3s1_winograd42_pack8_fp16sa_neon()
|
H A D | convolution_3x3_pack8to1_fp16s.h | 1036 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 1051 const __fp16* output0_tm_0 = (const __fp16*)out0_tm + (i * w_tm / 8 + j) * 1; in conv3x3s1_winograd64_pack8to1_fp16sa_neon()
|
H A D | convolution_3x3_pack4_bf16s.h | 1669 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_pack4_bf16s_neon() local 1684 const float* output0_tm_0 = (const float*)out0_tm + (i * w_tm / 8 + j) * 4; in conv3x3s1_winograd64_pack4_bf16s_neon() 3375 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd42_pack4_bf16s_neon() local 3390 const float* output0_tm_0 = (const float*)out0_tm + (i * w_tm / 6 + j) * 4; in conv3x3s1_winograd42_pack4_bf16s_neon()
|
/dports/benchmarks/vkpeak/vkpeak-20210430/ncnn/src/layer/arm/ |
H A D | convolution_3x3.h | 2717 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_neon4() local 2724 out0_tm.fill(0.f); in conv3x3s1_winograd64_neon4() 2739 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon4() 3128 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon4() 3699 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon4() 3879 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_neon4() local 3883 out0_tm.fill(0.f); in conv3x3s1_winograd64_neon4() 3891 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon4() 5600 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon5() 6043 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon5() [all …]
|
H A D | convolution_3x3_pack8_fp16s.h | 1059 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_pack8_fp16sa_neon() local 1074 const __fp16* output0_tm_0 = (const __fp16*)out0_tm + (i * w_tm / 8 + j) * 8; in conv3x3s1_winograd64_pack8_fp16sa_neon() 2152 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd42_pack8_fp16sa_neon() local 2167 const __fp16* output0_tm_0 = (const __fp16*)out0_tm + (i * w_tm / 6 + j) * 8; in conv3x3s1_winograd42_pack8_fp16sa_neon()
|
H A D | convolution_3x3_pack8to1_fp16s.h | 1036 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 1051 const __fp16* output0_tm_0 = (const __fp16*)out0_tm + (i * w_tm / 8 + j) * 1; in conv3x3s1_winograd64_pack8to1_fp16sa_neon()
|
H A D | convolution_3x3_pack4_bf16s.h | 1669 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_pack4_bf16s_neon() local 1684 const float* output0_tm_0 = (const float*)out0_tm + (i * w_tm / 8 + j) * 4; in conv3x3s1_winograd64_pack4_bf16s_neon() 3375 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd42_pack4_bf16s_neon() local 3390 const float* output0_tm_0 = (const float*)out0_tm + (i * w_tm / 6 + j) * 4; in conv3x3s1_winograd42_pack4_bf16s_neon()
|
/dports/misc/ncnn/ncnn-20211208/src/layer/arm/ |
H A D | convolution_3x3.h | 2717 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_neon4() local 2724 out0_tm.fill(0.f); in conv3x3s1_winograd64_neon4() 2739 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon4() 3128 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon4() 3699 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon4() 3879 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_neon4() local 3883 out0_tm.fill(0.f); in conv3x3s1_winograd64_neon4() 3891 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon4() 5600 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon5() 6043 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon5() [all …]
|
H A D | convolution_3x3_pack8_fp16s.h | 1059 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_pack8_fp16sa_neon() local 1074 const __fp16* output0_tm_0 = (const __fp16*)out0_tm + (i * w_tm / 8 + j) * 8; in conv3x3s1_winograd64_pack8_fp16sa_neon() 2152 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd42_pack8_fp16sa_neon() local 2167 const __fp16* output0_tm_0 = (const __fp16*)out0_tm + (i * w_tm / 6 + j) * 8; in conv3x3s1_winograd42_pack8_fp16sa_neon()
|
H A D | convolution_3x3_pack8to1_fp16s.h | 1036 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 1051 const __fp16* output0_tm_0 = (const __fp16*)out0_tm + (i * w_tm / 8 + j) * 1; in conv3x3s1_winograd64_pack8to1_fp16sa_neon()
|
H A D | convolution_3x3_pack4_bf16s.h | 1669 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_pack4_bf16s_neon() local 1684 const float* output0_tm_0 = (const float*)out0_tm + (i * w_tm / 8 + j) * 4; in conv3x3s1_winograd64_pack4_bf16s_neon() 3375 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd42_pack4_bf16s_neon() local 3390 const float* output0_tm_0 = (const float*)out0_tm + (i * w_tm / 6 + j) * 4; in conv3x3s1_winograd42_pack4_bf16s_neon()
|
/dports/graphics/vapoursynth-waifu2x-ncnn-vulkan/vapoursynth-waifu2x-ncnn-vulkan-r4/deps/ncnn/src/layer/arm/ |
H A D | convolution_3x3.h | 2717 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_neon4() local 2724 out0_tm.fill(0.f); in conv3x3s1_winograd64_neon4() 2739 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon4() 3128 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon4() 3699 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon4() 3879 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_neon4() local 3883 out0_tm.fill(0.f); in conv3x3s1_winograd64_neon4() 3891 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon4() 5600 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon5() 6043 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon5() [all …]
|
H A D | convolution_3x3_pack8_fp16s.h | 1059 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_pack8_fp16sa_neon() local 1074 const __fp16* output0_tm_0 = (const __fp16*)out0_tm + (i * w_tm / 8 + j) * 8; in conv3x3s1_winograd64_pack8_fp16sa_neon() 2152 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd42_pack8_fp16sa_neon() local 2167 const __fp16* output0_tm_0 = (const __fp16*)out0_tm + (i * w_tm / 6 + j) * 8; in conv3x3s1_winograd42_pack8_fp16sa_neon()
|
H A D | convolution_3x3_pack8to1_fp16s.h | 1036 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 1051 const __fp16* output0_tm_0 = (const __fp16*)out0_tm + (i * w_tm / 8 + j) * 1; in conv3x3s1_winograd64_pack8to1_fp16sa_neon()
|
H A D | convolution_3x3_pack4_bf16s.h | 1669 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_pack4_bf16s_neon() local 1684 const float* output0_tm_0 = (const float*)out0_tm + (i * w_tm / 8 + j) * 4; in conv3x3s1_winograd64_pack4_bf16s_neon() 3375 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd42_pack4_bf16s_neon() local 3390 const float* output0_tm_0 = (const float*)out0_tm + (i * w_tm / 6 + j) * 4; in conv3x3s1_winograd42_pack4_bf16s_neon()
|
/dports/graphics/realsr-ncnn-vulkan/realsr-ncnn-vulkan-20210210/src/ncnn/src/layer/arm/ |
H A D | convolution_3x3.h | 2717 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_neon4() local 2724 out0_tm.fill(0.f); in conv3x3s1_winograd64_neon4() 2739 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon4() 3128 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon4() 3699 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon4() 3879 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_neon4() local 3883 out0_tm.fill(0.f); in conv3x3s1_winograd64_neon4() 3891 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon4() 5600 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon5() 6043 float* output0_tm = out0_tm; in conv3x3s1_winograd64_neon5() [all …]
|
H A D | convolution_3x3_pack8_fp16s.h | 1059 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_pack8_fp16sa_neon() local 1074 const __fp16* output0_tm_0 = (const __fp16*)out0_tm + (i * w_tm / 8 + j) * 8; in conv3x3s1_winograd64_pack8_fp16sa_neon() 2152 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd42_pack8_fp16sa_neon() local 2167 const __fp16* output0_tm_0 = (const __fp16*)out0_tm + (i * w_tm / 6 + j) * 8; in conv3x3s1_winograd42_pack8_fp16sa_neon()
|
H A D | convolution_3x3_pack8to1_fp16s.h | 1036 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 1051 const __fp16* output0_tm_0 = (const __fp16*)out0_tm + (i * w_tm / 8 + j) * 1; in conv3x3s1_winograd64_pack8to1_fp16sa_neon()
|
H A D | convolution_3x3_pack4_bf16s.h | 1669 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd64_pack4_bf16s_neon() local 1684 const float* output0_tm_0 = (const float*)out0_tm + (i * w_tm / 8 + j) * 4; in conv3x3s1_winograd64_pack4_bf16s_neon() 3375 const Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd42_pack4_bf16s_neon() local 3390 const float* output0_tm_0 = (const float*)out0_tm + (i * w_tm / 6 + j) * 4; in conv3x3s1_winograd42_pack4_bf16s_neon()
|
/dports/graphics/vapoursynth-waifu2x-ncnn-vulkan/vapoursynth-waifu2x-ncnn-vulkan-r4/deps/ncnn/src/layer/x86/ |
H A D | convolution_3x3_int8.h | 275 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd23_int8_sse() local 287 int* output0_tm = out0_tm.row<int>(i); in conv3x3s1_winograd23_int8_sse() 381 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd23_int8_sse() local 386 int* output0_tm = out0_tm.row<int>(i); in conv3x3s1_winograd23_int8_sse() 758 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd43_int8_sse() local 763 int* output0_tm = out0_tm.row<int>(i); in conv3x3s1_winograd43_int8_sse()
|
/dports/graphics/waifu2x-ncnn-vulkan/waifu2x-ncnn-vulkan-20210521/src/ncnn/src/layer/x86/ |
H A D | convolution_3x3_int8.h | 275 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd23_int8_sse() local 287 int* output0_tm = out0_tm.row<int>(i); in conv3x3s1_winograd23_int8_sse() 381 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd23_int8_sse() local 386 int* output0_tm = out0_tm.row<int>(i); in conv3x3s1_winograd23_int8_sse() 758 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd43_int8_sse() local 763 int* output0_tm = out0_tm.row<int>(i); in conv3x3s1_winograd43_int8_sse()
|
/dports/misc/ncnn/ncnn-20211208/src/layer/x86/ |
H A D | convolution_3x3_int8.h | 275 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd23_int8_sse() local 287 int* output0_tm = out0_tm.row<int>(i); in conv3x3s1_winograd23_int8_sse() 381 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd23_int8_sse() local 386 int* output0_tm = out0_tm.row<int>(i); in conv3x3s1_winograd23_int8_sse() 758 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd43_int8_sse() local 763 int* output0_tm = out0_tm.row<int>(i); in conv3x3s1_winograd43_int8_sse()
|
/dports/benchmarks/vkpeak/vkpeak-20210430/ncnn/src/layer/x86/ |
H A D | convolution_3x3_int8.h | 275 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd23_int8_sse() local 287 int* output0_tm = out0_tm.row<int>(i); in conv3x3s1_winograd23_int8_sse() 381 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd23_int8_sse() local 386 int* output0_tm = out0_tm.row<int>(i); in conv3x3s1_winograd23_int8_sse() 758 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd43_int8_sse() local 763 int* output0_tm = out0_tm.row<int>(i); in conv3x3s1_winograd43_int8_sse()
|
/dports/graphics/realsr-ncnn-vulkan/realsr-ncnn-vulkan-20210210/src/ncnn/src/layer/x86/ |
H A D | convolution_3x3_int8.h | 275 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd23_int8_sse() local 287 int* output0_tm = out0_tm.row<int>(i); in conv3x3s1_winograd23_int8_sse() 381 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd23_int8_sse() local 386 int* output0_tm = out0_tm.row<int>(i); in conv3x3s1_winograd23_int8_sse() 758 Mat out0_tm = top_blob_tm.channel(p); in conv3x3s1_winograd43_int8_sse() local 763 int* output0_tm = out0_tm.row<int>(i); in conv3x3s1_winograd43_int8_sse()
|