/dports/graphics/realsr-ncnn-vulkan/realsr-ncnn-vulkan-20210210/src/ncnn/src/layer/arm/ |
H A D | convolution_3x3_pack8_fp16s.h | 295 float16x8_t _r0tm2 = vsubq_f16(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack8_fp16sa_neon() local 326 vst1q_f16(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack8_fp16sa_neon() 1420 … float16x8_t _r0tm2 = vfmaq_n_f16(vsubq_f16(_tmp04, _tmp03), vsubq_f16(_tmp01, _tmp02), 4.f); in conv3x3s1_winograd42_pack8_fp16sa_neon() local 1427 vst1q_f16(r0_tm_2, _r0tm2); in conv3x3s1_winograd42_pack8_fp16sa_neon()
|
H A D | convolution_3x3_pack8to1_fp16s.h | 307 float16x8_t _r0tm2 = vsubq_f16(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 338 vst1q_f16(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack8to1_fp16sa_neon()
|
H A D | convolution_3x3_pack4_bf16s.h | 191 float32x4_t _r0tm2 = vsubq_f32(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack4_bf16s_neon() local 222 vst1q_f32(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack4_bf16s_neon() 1929 … float32x4_t _r0tm2 = vmlaq_n_f32(vsubq_f32(_tmp04, _tmp03), vsubq_f32(_tmp01, _tmp02), 4.f); in conv3x3s1_winograd42_pack4_bf16s_neon() local 1936 vst1q_f32(r0_tm_2, _r0tm2); in conv3x3s1_winograd42_pack4_bf16s_neon()
|
H A D | convolution_3x3_pack4to1_bf16s.h | 191 float32x4_t _r0tm2 = vsubq_f32(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 222 vst1q_f32(r0_tm_2, _r0tm2); 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_pack8_fp16s.h | 295 float16x8_t _r0tm2 = vsubq_f16(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack8_fp16sa_neon() local 326 vst1q_f16(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack8_fp16sa_neon() 1420 … float16x8_t _r0tm2 = vfmaq_n_f16(vsubq_f16(_tmp04, _tmp03), vsubq_f16(_tmp01, _tmp02), 4.f); in conv3x3s1_winograd42_pack8_fp16sa_neon() local 1427 vst1q_f16(r0_tm_2, _r0tm2); in conv3x3s1_winograd42_pack8_fp16sa_neon()
|
H A D | convolution_3x3_pack8to1_fp16s.h | 307 float16x8_t _r0tm2 = vsubq_f16(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 338 vst1q_f16(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack8to1_fp16sa_neon()
|
H A D | convolution_3x3_pack4_bf16s.h | 191 float32x4_t _r0tm2 = vsubq_f32(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack4_bf16s_neon() local 222 vst1q_f32(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack4_bf16s_neon() 1929 … float32x4_t _r0tm2 = vmlaq_n_f32(vsubq_f32(_tmp04, _tmp03), vsubq_f32(_tmp01, _tmp02), 4.f); in conv3x3s1_winograd42_pack4_bf16s_neon() local 1936 vst1q_f32(r0_tm_2, _r0tm2); in conv3x3s1_winograd42_pack4_bf16s_neon()
|
H A D | convolution_3x3_pack4to1_bf16s.h | 191 float32x4_t _r0tm2 = vsubq_f32(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 222 vst1q_f32(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack4to1_bf16s_neon()
|
H A D | convolution_3x3_pack8to4_fp16s.h | 313 float16x8_t _r0tm2 = vsubq_f16(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 344 vst1q_f16(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack8to4_fp16sa_neon()
|
/dports/graphics/waifu2x-ncnn-vulkan/waifu2x-ncnn-vulkan-20210521/src/ncnn/src/layer/arm/ |
H A D | convolution_3x3_pack8_fp16s.h | 295 float16x8_t _r0tm2 = vsubq_f16(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack8_fp16sa_neon() local 326 vst1q_f16(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack8_fp16sa_neon() 1420 … float16x8_t _r0tm2 = vfmaq_n_f16(vsubq_f16(_tmp04, _tmp03), vsubq_f16(_tmp01, _tmp02), 4.f); in conv3x3s1_winograd42_pack8_fp16sa_neon() local 1427 vst1q_f16(r0_tm_2, _r0tm2); in conv3x3s1_winograd42_pack8_fp16sa_neon()
|
H A D | convolution_3x3_pack8to1_fp16s.h | 307 float16x8_t _r0tm2 = vsubq_f16(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 338 vst1q_f16(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack8to1_fp16sa_neon()
|
H A D | convolution_3x3_pack4_bf16s.h | 191 float32x4_t _r0tm2 = vsubq_f32(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack4_bf16s_neon() local 222 vst1q_f32(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack4_bf16s_neon() 1929 … float32x4_t _r0tm2 = vmlaq_n_f32(vsubq_f32(_tmp04, _tmp03), vsubq_f32(_tmp01, _tmp02), 4.f); in conv3x3s1_winograd42_pack4_bf16s_neon() local 1936 vst1q_f32(r0_tm_2, _r0tm2); in conv3x3s1_winograd42_pack4_bf16s_neon()
|
H A D | convolution_3x3_pack4to1_bf16s.h | 191 float32x4_t _r0tm2 = vsubq_f32(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 222 vst1q_f32(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack4to1_bf16s_neon()
|
H A D | convolution_3x3_pack8to4_fp16s.h | 313 float16x8_t _r0tm2 = vsubq_f16(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 344 vst1q_f16(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack8to4_fp16sa_neon()
|
/dports/benchmarks/vkpeak/vkpeak-20210430/ncnn/src/layer/arm/ |
H A D | convolution_3x3_pack8_fp16s.h | 295 float16x8_t _r0tm2 = vsubq_f16(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack8_fp16sa_neon() local 326 vst1q_f16(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack8_fp16sa_neon() 1420 … float16x8_t _r0tm2 = vfmaq_n_f16(vsubq_f16(_tmp04, _tmp03), vsubq_f16(_tmp01, _tmp02), 4.f); in conv3x3s1_winograd42_pack8_fp16sa_neon() local 1427 vst1q_f16(r0_tm_2, _r0tm2); in conv3x3s1_winograd42_pack8_fp16sa_neon()
|
H A D | convolution_3x3_pack8to1_fp16s.h | 307 float16x8_t _r0tm2 = vsubq_f16(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 338 vst1q_f16(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack8to1_fp16sa_neon()
|
H A D | convolution_3x3_pack4_bf16s.h | 191 float32x4_t _r0tm2 = vsubq_f32(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack4_bf16s_neon() local 222 vst1q_f32(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack4_bf16s_neon() 1929 … float32x4_t _r0tm2 = vmlaq_n_f32(vsubq_f32(_tmp04, _tmp03), vsubq_f32(_tmp01, _tmp02), 4.f); in conv3x3s1_winograd42_pack4_bf16s_neon() local 1936 vst1q_f32(r0_tm_2, _r0tm2); in conv3x3s1_winograd42_pack4_bf16s_neon()
|
H A D | convolution_3x3_pack4to1_bf16s.h | 191 float32x4_t _r0tm2 = vsubq_f32(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 222 vst1q_f32(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack4to1_bf16s_neon()
|
H A D | convolution_3x3_pack8to4_fp16s.h | 313 float16x8_t _r0tm2 = vsubq_f16(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 344 vst1q_f16(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack8to4_fp16sa_neon()
|
/dports/misc/ncnn/ncnn-20211208/src/layer/arm/ |
H A D | convolution_3x3_pack8_fp16s.h | 295 float16x8_t _r0tm2 = vsubq_f16(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack8_fp16sa_neon() local 326 vst1q_f16(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack8_fp16sa_neon() 1420 … float16x8_t _r0tm2 = vfmaq_n_f16(vsubq_f16(_tmp04, _tmp03), vsubq_f16(_tmp01, _tmp02), 4.f); in conv3x3s1_winograd42_pack8_fp16sa_neon() local 1427 vst1q_f16(r0_tm_2, _r0tm2); in conv3x3s1_winograd42_pack8_fp16sa_neon()
|
H A D | convolution_3x3_pack8to1_fp16s.h | 307 float16x8_t _r0tm2 = vsubq_f16(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 338 vst1q_f16(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack8to1_fp16sa_neon()
|
H A D | convolution_3x3_pack4_bf16s.h | 191 float32x4_t _r0tm2 = vsubq_f32(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack4_bf16s_neon() local 222 vst1q_f32(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack4_bf16s_neon() 1929 … float32x4_t _r0tm2 = vmlaq_n_f32(vsubq_f32(_tmp04, _tmp03), vsubq_f32(_tmp01, _tmp02), 4.f); in conv3x3s1_winograd42_pack4_bf16s_neon() local 1936 vst1q_f32(r0_tm_2, _r0tm2); in conv3x3s1_winograd42_pack4_bf16s_neon()
|
H A D | convolution_3x3_pack4to1_bf16s.h | 191 float32x4_t _r0tm2 = vsubq_f32(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 222 vst1q_f32(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack4to1_bf16s_neon()
|
H A D | convolution_3x3_pack8to4_fp16s.h | 313 float16x8_t _r0tm2 = vsubq_f16(_tmp12a, _tmp12b); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 344 vst1q_f16(r0_tm_2, _r0tm2); in conv3x3s1_winograd64_pack8to4_fp16sa_neon()
|
/dports/misc/ncnn/ncnn-20211208/src/layer/riscv/ |
H A D | convolution_3x3_packn.h | 251 vfloat32m1_t _r0tm2 = vfsub_vv_f32m1(_tmp12a, _tmp12b, vl); in conv3x3s1_winograd64_packn_rvv() local 267 vse32_v_f32m1(r0_tm_2, _r0tm2, vl); in conv3x3s1_winograd64_packn_rvv() 939 …vfloat32m1_t _r0tm2 = vfmacc_vf_f32m1(vfsub_vv_f32m1(_tmp04, _tmp03, vl), 4.f, vfsub_vv_f32m1(_tmp… in conv3x3s1_winograd42_packn_rvv() local 946 vse32_v_f32m1(r0_tm_2, _r0tm2, vl); in conv3x3s1_winograd42_packn_rvv()
|