/dports/graphics/waifu2x-ncnn-vulkan/waifu2x-ncnn-vulkan-20210521/src/ncnn/src/layer/arm/ |
H A D | convolution_3x3.h | 1940 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_neon4() local 4669 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_neon5() local 4822 r0_tm0 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4823 r0_tm1 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4824 r0_tm2 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4825 r0_tm3 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4841 r0_tm0 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4842 r0_tm1 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4843 r0_tm2 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4844 r0_tm3 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() [all …]
|
H A D | convolution_3x3_pack8_fp16s.h | 187 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_pack8_fp16sa_neon() local 262 __fp16* r0_tm_0 = (__fp16*)img0_tm + (i * w_tm / 8 + j) * 8; in conv3x3s1_winograd64_pack8_fp16sa_neon() 1365 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd42_pack8_fp16sa_neon() local 1402 __fp16* r0_tm_0 = (__fp16*)img0_tm + (i * w_tm / 6 + j) * 8; in conv3x3s1_winograd42_pack8_fp16sa_neon()
|
H A D | convolution_3x3_pack8to1_fp16s.h | 199 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 274 __fp16* r0_tm_0 = (__fp16*)img0_tm + (i * w_tm / 8 + j) * 8; in conv3x3s1_winograd64_pack8to1_fp16sa_neon()
|
H A D | convolution_3x3_pack4_bf16s.h | 83 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_pack4_bf16s_neon() local 158 float* r0_tm_0 = (float*)img0_tm + (i * w_tm / 8 + j) * 4; in conv3x3s1_winograd64_pack4_bf16s_neon() 1874 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd42_pack4_bf16s_neon() local 1911 float* r0_tm_0 = (float*)img0_tm + (i * w_tm / 6 + j) * 4; in conv3x3s1_winograd42_pack4_bf16s_neon()
|
H A D | convolution_3x3_pack4to1_bf16s.h | 83 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 158 float* r0_tm_0 = (float*)img0_tm + (i * w_tm / 8 + j) * 4; in conv3x3s1_winograd64_pack4to1_bf16s_neon()
|
/dports/benchmarks/vkpeak/vkpeak-20210430/ncnn/src/layer/arm/ |
H A D | convolution_3x3.h | 1940 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_neon4() local 4669 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_neon5() local 4822 r0_tm0 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4823 r0_tm1 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4824 r0_tm2 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4825 r0_tm3 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4841 r0_tm0 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4842 r0_tm1 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4843 r0_tm2 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4844 r0_tm3 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() [all …]
|
H A D | convolution_3x3_pack8_fp16s.h | 187 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_pack8_fp16sa_neon() local 262 __fp16* r0_tm_0 = (__fp16*)img0_tm + (i * w_tm / 8 + j) * 8; in conv3x3s1_winograd64_pack8_fp16sa_neon() 1365 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd42_pack8_fp16sa_neon() local 1402 __fp16* r0_tm_0 = (__fp16*)img0_tm + (i * w_tm / 6 + j) * 8; in conv3x3s1_winograd42_pack8_fp16sa_neon()
|
H A D | convolution_3x3_pack8to1_fp16s.h | 199 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 274 __fp16* r0_tm_0 = (__fp16*)img0_tm + (i * w_tm / 8 + j) * 8; in conv3x3s1_winograd64_pack8to1_fp16sa_neon()
|
H A D | convolution_3x3_pack4_bf16s.h | 83 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_pack4_bf16s_neon() local 158 float* r0_tm_0 = (float*)img0_tm + (i * w_tm / 8 + j) * 4; in conv3x3s1_winograd64_pack4_bf16s_neon() 1874 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd42_pack4_bf16s_neon() local 1911 float* r0_tm_0 = (float*)img0_tm + (i * w_tm / 6 + j) * 4; in conv3x3s1_winograd42_pack4_bf16s_neon()
|
H A D | convolution_3x3_pack4to1_bf16s.h | 83 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 158 float* r0_tm_0 = (float*)img0_tm + (i * w_tm / 8 + j) * 4; in conv3x3s1_winograd64_pack4to1_bf16s_neon()
|
/dports/misc/ncnn/ncnn-20211208/src/layer/arm/ |
H A D | convolution_3x3.h | 1940 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_neon4() local 4669 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_neon5() local 4822 r0_tm0 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4823 r0_tm1 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4824 r0_tm2 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4825 r0_tm3 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4841 r0_tm0 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4842 r0_tm1 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4843 r0_tm2 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4844 r0_tm3 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() [all …]
|
H A D | convolution_3x3_pack8_fp16s.h | 187 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_pack8_fp16sa_neon() local 262 __fp16* r0_tm_0 = (__fp16*)img0_tm + (i * w_tm / 8 + j) * 8; in conv3x3s1_winograd64_pack8_fp16sa_neon() 1365 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd42_pack8_fp16sa_neon() local 1402 __fp16* r0_tm_0 = (__fp16*)img0_tm + (i * w_tm / 6 + j) * 8; in conv3x3s1_winograd42_pack8_fp16sa_neon()
|
H A D | convolution_3x3_pack8to1_fp16s.h | 199 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 274 __fp16* r0_tm_0 = (__fp16*)img0_tm + (i * w_tm / 8 + j) * 8; in conv3x3s1_winograd64_pack8to1_fp16sa_neon()
|
H A D | convolution_3x3_pack4_bf16s.h | 83 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_pack4_bf16s_neon() local 158 float* r0_tm_0 = (float*)img0_tm + (i * w_tm / 8 + j) * 4; in conv3x3s1_winograd64_pack4_bf16s_neon() 1874 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd42_pack4_bf16s_neon() local 1911 float* r0_tm_0 = (float*)img0_tm + (i * w_tm / 6 + j) * 4; in conv3x3s1_winograd42_pack4_bf16s_neon()
|
H A D | convolution_3x3_pack4to1_bf16s.h | 83 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 158 float* r0_tm_0 = (float*)img0_tm + (i * w_tm / 8 + j) * 4; in conv3x3s1_winograd64_pack4to1_bf16s_neon()
|
/dports/graphics/vapoursynth-waifu2x-ncnn-vulkan/vapoursynth-waifu2x-ncnn-vulkan-r4/deps/ncnn/src/layer/arm/ |
H A D | convolution_3x3.h | 1940 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_neon4() local 4669 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_neon5() local 4822 r0_tm0 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4823 r0_tm1 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4824 r0_tm2 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4825 r0_tm3 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4841 r0_tm0 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4842 r0_tm1 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4843 r0_tm2 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4844 r0_tm3 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() [all …]
|
H A D | convolution_3x3_pack8_fp16s.h | 187 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_pack8_fp16sa_neon() local 262 __fp16* r0_tm_0 = (__fp16*)img0_tm + (i * w_tm / 8 + j) * 8; in conv3x3s1_winograd64_pack8_fp16sa_neon() 1365 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd42_pack8_fp16sa_neon() local 1402 __fp16* r0_tm_0 = (__fp16*)img0_tm + (i * w_tm / 6 + j) * 8; in conv3x3s1_winograd42_pack8_fp16sa_neon()
|
H A D | convolution_3x3_pack8to1_fp16s.h | 199 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 274 __fp16* r0_tm_0 = (__fp16*)img0_tm + (i * w_tm / 8 + j) * 8; in conv3x3s1_winograd64_pack8to1_fp16sa_neon()
|
H A D | convolution_3x3_pack4_bf16s.h | 83 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_pack4_bf16s_neon() local 158 float* r0_tm_0 = (float*)img0_tm + (i * w_tm / 8 + j) * 4; in conv3x3s1_winograd64_pack4_bf16s_neon() 1874 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd42_pack4_bf16s_neon() local 1911 float* r0_tm_0 = (float*)img0_tm + (i * w_tm / 6 + j) * 4; in conv3x3s1_winograd42_pack4_bf16s_neon()
|
H A D | convolution_3x3_pack4to1_bf16s.h | 83 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 158 float* r0_tm_0 = (float*)img0_tm + (i * w_tm / 8 + j) * 4; in conv3x3s1_winograd64_pack4to1_bf16s_neon()
|
/dports/graphics/realsr-ncnn-vulkan/realsr-ncnn-vulkan-20210210/src/ncnn/src/layer/arm/ |
H A D | convolution_3x3.h | 1940 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_neon4() local 4669 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_neon5() local 4822 r0_tm0 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4823 r0_tm1 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4824 r0_tm2 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4825 r0_tm3 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4841 r0_tm0 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4842 r0_tm1 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4843 r0_tm2 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() 4844 r0_tm3 += img0_tm.w * tiles; in conv3x3s1_winograd64_neon5() [all …]
|
H A D | convolution_3x3_pack8_fp16s.h | 187 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_pack8_fp16sa_neon() local 262 __fp16* r0_tm_0 = (__fp16*)img0_tm + (i * w_tm / 8 + j) * 8; in conv3x3s1_winograd64_pack8_fp16sa_neon() 1365 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd42_pack8_fp16sa_neon() local 1402 __fp16* r0_tm_0 = (__fp16*)img0_tm + (i * w_tm / 6 + j) * 8; in conv3x3s1_winograd42_pack8_fp16sa_neon()
|
H A D | convolution_3x3_pack8to1_fp16s.h | 199 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 274 __fp16* r0_tm_0 = (__fp16*)img0_tm + (i * w_tm / 8 + j) * 8; in conv3x3s1_winograd64_pack8to1_fp16sa_neon()
|
H A D | convolution_3x3_pack4_bf16s.h | 83 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_pack4_bf16s_neon() local 158 float* r0_tm_0 = (float*)img0_tm + (i * w_tm / 8 + j) * 4; in conv3x3s1_winograd64_pack4_bf16s_neon() 1874 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd42_pack4_bf16s_neon() local 1911 float* r0_tm_0 = (float*)img0_tm + (i * w_tm / 6 + j) * 4; in conv3x3s1_winograd42_pack4_bf16s_neon()
|
H A D | convolution_3x3_pack4to1_bf16s.h | 83 Mat img0_tm = bottom_blob_tm.channel(q); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 158 float* r0_tm_0 = (float*)img0_tm + (i * w_tm / 8 + j) * 4; in conv3x3s1_winograd64_pack4to1_bf16s_neon()
|