/dports/graphics/realsr-ncnn-vulkan/realsr-ncnn-vulkan-20210210/src/ncnn/src/layer/arm/ |
H A D | convolution_3x3_pack8_fp16s.h | 318 float16x8_t _r0tm5 = vaddq_f16(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack8_fp16sa_neon() local 329 vst1q_f16(r0_tm_5, _r0tm5); in conv3x3s1_winograd64_pack8_fp16sa_neon() 1423 … float16x8_t _r0tm5 = vfmsq_n_f16(vfmaq_n_f16(_tmp05, _tmp01, 4.f), _tmp03, 5.f); in conv3x3s1_winograd42_pack8_fp16sa_neon() local 1430 vst1q_f16(r0_tm_5, _r0tm5); in conv3x3s1_winograd42_pack8_fp16sa_neon()
|
H A D | convolution_3x3_pack8to1_fp16s.h | 330 float16x8_t _r0tm5 = vaddq_f16(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 341 vst1q_f16(r0_tm_5, _r0tm5); in conv3x3s1_winograd64_pack8to1_fp16sa_neon()
|
H A D | convolution_3x3_pack4_bf16s.h | 214 float32x4_t _r0tm5 = vaddq_f32(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack4_bf16s_neon() local 225 vst1q_f32(r0_tm_5, _r0tm5); in conv3x3s1_winograd64_pack4_bf16s_neon() 1932 … float32x4_t _r0tm5 = vmlsq_n_f32(vmlaq_n_f32(_tmp05, _tmp01, 4.f), _tmp03, 5.f); in conv3x3s1_winograd42_pack4_bf16s_neon() local 1939 vst1q_f32(r0_tm_5, _r0tm5); in conv3x3s1_winograd42_pack4_bf16s_neon()
|
H A D | convolution_3x3_pack4to1_bf16s.h | 214 float32x4_t _r0tm5 = vaddq_f32(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 225 vst1q_f32(r0_tm_5, _r0tm5); 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 | 318 float16x8_t _r0tm5 = vaddq_f16(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack8_fp16sa_neon() local 329 vst1q_f16(r0_tm_5, _r0tm5); in conv3x3s1_winograd64_pack8_fp16sa_neon() 1423 … float16x8_t _r0tm5 = vfmsq_n_f16(vfmaq_n_f16(_tmp05, _tmp01, 4.f), _tmp03, 5.f); in conv3x3s1_winograd42_pack8_fp16sa_neon() local 1430 vst1q_f16(r0_tm_5, _r0tm5); in conv3x3s1_winograd42_pack8_fp16sa_neon()
|
H A D | convolution_3x3_pack8to1_fp16s.h | 330 float16x8_t _r0tm5 = vaddq_f16(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 341 vst1q_f16(r0_tm_5, _r0tm5); in conv3x3s1_winograd64_pack8to1_fp16sa_neon()
|
H A D | convolution_3x3_pack4_bf16s.h | 214 float32x4_t _r0tm5 = vaddq_f32(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack4_bf16s_neon() local 225 vst1q_f32(r0_tm_5, _r0tm5); in conv3x3s1_winograd64_pack4_bf16s_neon() 1932 … float32x4_t _r0tm5 = vmlsq_n_f32(vmlaq_n_f32(_tmp05, _tmp01, 4.f), _tmp03, 5.f); in conv3x3s1_winograd42_pack4_bf16s_neon() local 1939 vst1q_f32(r0_tm_5, _r0tm5); in conv3x3s1_winograd42_pack4_bf16s_neon()
|
H A D | convolution_3x3_pack4to1_bf16s.h | 214 float32x4_t _r0tm5 = vaddq_f32(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 225 vst1q_f32(r0_tm_5, _r0tm5); in conv3x3s1_winograd64_pack4to1_bf16s_neon()
|
H A D | convolution_3x3_pack8to4_fp16s.h | 336 float16x8_t _r0tm5 = vaddq_f16(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 347 vst1q_f16(r0_tm_5, _r0tm5); 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 | 318 float16x8_t _r0tm5 = vaddq_f16(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack8_fp16sa_neon() local 329 vst1q_f16(r0_tm_5, _r0tm5); in conv3x3s1_winograd64_pack8_fp16sa_neon() 1423 … float16x8_t _r0tm5 = vfmsq_n_f16(vfmaq_n_f16(_tmp05, _tmp01, 4.f), _tmp03, 5.f); in conv3x3s1_winograd42_pack8_fp16sa_neon() local 1430 vst1q_f16(r0_tm_5, _r0tm5); in conv3x3s1_winograd42_pack8_fp16sa_neon()
|
H A D | convolution_3x3_pack8to1_fp16s.h | 330 float16x8_t _r0tm5 = vaddq_f16(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 341 vst1q_f16(r0_tm_5, _r0tm5); in conv3x3s1_winograd64_pack8to1_fp16sa_neon()
|
H A D | convolution_3x3_pack4_bf16s.h | 214 float32x4_t _r0tm5 = vaddq_f32(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack4_bf16s_neon() local 225 vst1q_f32(r0_tm_5, _r0tm5); in conv3x3s1_winograd64_pack4_bf16s_neon() 1932 … float32x4_t _r0tm5 = vmlsq_n_f32(vmlaq_n_f32(_tmp05, _tmp01, 4.f), _tmp03, 5.f); in conv3x3s1_winograd42_pack4_bf16s_neon() local 1939 vst1q_f32(r0_tm_5, _r0tm5); in conv3x3s1_winograd42_pack4_bf16s_neon()
|
H A D | convolution_3x3_pack4to1_bf16s.h | 214 float32x4_t _r0tm5 = vaddq_f32(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 225 vst1q_f32(r0_tm_5, _r0tm5); in conv3x3s1_winograd64_pack4to1_bf16s_neon()
|
H A D | convolution_3x3_pack8to4_fp16s.h | 336 float16x8_t _r0tm5 = vaddq_f16(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 347 vst1q_f16(r0_tm_5, _r0tm5); in conv3x3s1_winograd64_pack8to4_fp16sa_neon()
|
/dports/benchmarks/vkpeak/vkpeak-20210430/ncnn/src/layer/arm/ |
H A D | convolution_3x3_pack8_fp16s.h | 318 float16x8_t _r0tm5 = vaddq_f16(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack8_fp16sa_neon() local 329 vst1q_f16(r0_tm_5, _r0tm5); in conv3x3s1_winograd64_pack8_fp16sa_neon() 1423 … float16x8_t _r0tm5 = vfmsq_n_f16(vfmaq_n_f16(_tmp05, _tmp01, 4.f), _tmp03, 5.f); in conv3x3s1_winograd42_pack8_fp16sa_neon() local 1430 vst1q_f16(r0_tm_5, _r0tm5); in conv3x3s1_winograd42_pack8_fp16sa_neon()
|
H A D | convolution_3x3_pack8to1_fp16s.h | 330 float16x8_t _r0tm5 = vaddq_f16(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 341 vst1q_f16(r0_tm_5, _r0tm5); in conv3x3s1_winograd64_pack8to1_fp16sa_neon()
|
H A D | convolution_3x3_pack4_bf16s.h | 214 float32x4_t _r0tm5 = vaddq_f32(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack4_bf16s_neon() local 225 vst1q_f32(r0_tm_5, _r0tm5); in conv3x3s1_winograd64_pack4_bf16s_neon() 1932 … float32x4_t _r0tm5 = vmlsq_n_f32(vmlaq_n_f32(_tmp05, _tmp01, 4.f), _tmp03, 5.f); in conv3x3s1_winograd42_pack4_bf16s_neon() local 1939 vst1q_f32(r0_tm_5, _r0tm5); in conv3x3s1_winograd42_pack4_bf16s_neon()
|
H A D | convolution_3x3_pack4to1_bf16s.h | 214 float32x4_t _r0tm5 = vaddq_f32(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 225 vst1q_f32(r0_tm_5, _r0tm5); in conv3x3s1_winograd64_pack4to1_bf16s_neon()
|
H A D | convolution_3x3_pack8to4_fp16s.h | 336 float16x8_t _r0tm5 = vaddq_f16(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 347 vst1q_f16(r0_tm_5, _r0tm5); in conv3x3s1_winograd64_pack8to4_fp16sa_neon()
|
/dports/misc/ncnn/ncnn-20211208/src/layer/arm/ |
H A D | convolution_3x3_pack8_fp16s.h | 318 float16x8_t _r0tm5 = vaddq_f16(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack8_fp16sa_neon() local 329 vst1q_f16(r0_tm_5, _r0tm5); in conv3x3s1_winograd64_pack8_fp16sa_neon() 1423 … float16x8_t _r0tm5 = vfmsq_n_f16(vfmaq_n_f16(_tmp05, _tmp01, 4.f), _tmp03, 5.f); in conv3x3s1_winograd42_pack8_fp16sa_neon() local 1430 vst1q_f16(r0_tm_5, _r0tm5); in conv3x3s1_winograd42_pack8_fp16sa_neon()
|
H A D | convolution_3x3_pack8to1_fp16s.h | 330 float16x8_t _r0tm5 = vaddq_f16(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 341 vst1q_f16(r0_tm_5, _r0tm5); in conv3x3s1_winograd64_pack8to1_fp16sa_neon()
|
H A D | convolution_3x3_pack4_bf16s.h | 214 float32x4_t _r0tm5 = vaddq_f32(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack4_bf16s_neon() local 225 vst1q_f32(r0_tm_5, _r0tm5); in conv3x3s1_winograd64_pack4_bf16s_neon() 1932 … float32x4_t _r0tm5 = vmlsq_n_f32(vmlaq_n_f32(_tmp05, _tmp01, 4.f), _tmp03, 5.f); in conv3x3s1_winograd42_pack4_bf16s_neon() local 1939 vst1q_f32(r0_tm_5, _r0tm5); in conv3x3s1_winograd42_pack4_bf16s_neon()
|
H A D | convolution_3x3_pack4to1_bf16s.h | 214 float32x4_t _r0tm5 = vaddq_f32(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 225 vst1q_f32(r0_tm_5, _r0tm5); in conv3x3s1_winograd64_pack4to1_bf16s_neon()
|
H A D | convolution_3x3_pack8to4_fp16s.h | 336 float16x8_t _r0tm5 = vaddq_f16(_tmp56a, _tmp56b); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 347 vst1q_f16(r0_tm_5, _r0tm5); in conv3x3s1_winograd64_pack8to4_fp16sa_neon()
|
/dports/misc/ncnn/ncnn-20211208/src/layer/riscv/ |
H A D | convolution_3x3_packn.h | 262 vfloat32m1_t _r0tm5 = vfadd_vv_f32m1(_tmp56a, _tmp56b, vl); in conv3x3s1_winograd64_packn_rvv() local 270 vse32_v_f32m1(r0_tm_5, _r0tm5, vl); in conv3x3s1_winograd64_packn_rvv() 942 … vfloat32m1_t _r0tm5 = vfmacc_vf_f32m1(vfmacc_vf_f32m1(_tmp05, 4.f, _tmp01, vl), -5.f, _tmp03, vl); in conv3x3s1_winograd42_packn_rvv() local 949 vse32_v_f32m1(r0_tm_5, _r0tm5, vl); in conv3x3s1_winograd42_packn_rvv()
|