| /dports/graphics/vapoursynth-waifu2x-ncnn-vulkan/vapoursynth-waifu2x-ncnn-vulkan-r4/deps/ncnn/src/layer/x86/ |
| H A D | convolution_2x2_pack8.h | 66 __m256 _k03 = _mm256_loadu_ps(kptr + 24); in conv2x2s1_pack8_avx() local
72 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in conv2x2s1_pack8_avx()
100 _sum1 = _mm256_fmadd_ps(_k03, _r03, _sum1); in conv2x2s1_pack8_avx()
109 _k03 = _mm256_loadu_ps(kptr + 24); in conv2x2s1_pack8_avx()
115 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in conv2x2s1_pack8_avx()
140 _sum1 = _mm256_fmadd_ps(_k03, _r03, _sum1); in conv2x2s1_pack8_avx()
259 __m256 _k03 = _mm256_loadu_ps(kptr + 24); in conv2x2s1_pack8_avx() local
265 _sum = _mm256_fmadd_ps(_k03, _r03, _sum); in conv2x2s1_pack8_avx()
292 _k03 = _mm256_loadu_ps(kptr + 24); in conv2x2s1_pack8_avx()
298 _sum = _mm256_fmadd_ps(_k03, _r03, _sum); in conv2x2s1_pack8_avx()
|
| H A D | convolution_2x2_pack8_fp16.h | 247 __m256 _k03 = loadfp16(kptr + 24); in conv2x2s1_fp16_pack8_avx() local
253 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in conv2x2s1_fp16_pack8_avx()
281 _sum1 = _mm256_fmadd_ps(_k03, _r03, _sum1); in conv2x2s1_fp16_pack8_avx()
290 _k03 = loadfp16(kptr + 24); in conv2x2s1_fp16_pack8_avx()
296 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in conv2x2s1_fp16_pack8_avx()
321 _sum1 = _mm256_fmadd_ps(_k03, _r03, _sum1); in conv2x2s1_fp16_pack8_avx()
440 __m256 _k03 = loadfp16(kptr + 24); in conv2x2s1_fp16_pack8_avx() local
446 _sum = _mm256_fmadd_ps(_k03, _r03, _sum); in conv2x2s1_fp16_pack8_avx()
473 _k03 = loadfp16(kptr + 24); in conv2x2s1_fp16_pack8_avx()
479 _sum = _mm256_fmadd_ps(_k03, _r03, _sum); in conv2x2s1_fp16_pack8_avx()
|
| H A D | convolutiondepthwise_5x5_pack8.h | 60 __m256 _k03 = _mm256_loadu_ps(k0 + 24); in convdw5x5s1_pack8_avx() local
67 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in convdw5x5s1_pack8_avx()
214 __m256 _k03 = _mm256_loadu_ps(k0 + 24); in convdw5x5s2_pack8_avx() local
221 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in convdw5x5s2_pack8_avx()
|
| /dports/graphics/waifu2x-ncnn-vulkan/waifu2x-ncnn-vulkan-20210521/src/ncnn/src/layer/x86/ |
| H A D | convolution_2x2_pack8.h | 66 __m256 _k03 = _mm256_loadu_ps(kptr + 24); in conv2x2s1_pack8_avx() local
72 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in conv2x2s1_pack8_avx()
100 _sum1 = _mm256_fmadd_ps(_k03, _r03, _sum1); in conv2x2s1_pack8_avx()
109 _k03 = _mm256_loadu_ps(kptr + 24); in conv2x2s1_pack8_avx()
115 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in conv2x2s1_pack8_avx()
140 _sum1 = _mm256_fmadd_ps(_k03, _r03, _sum1); in conv2x2s1_pack8_avx()
259 __m256 _k03 = _mm256_loadu_ps(kptr + 24); in conv2x2s1_pack8_avx() local
265 _sum = _mm256_fmadd_ps(_k03, _r03, _sum); in conv2x2s1_pack8_avx()
292 _k03 = _mm256_loadu_ps(kptr + 24); in conv2x2s1_pack8_avx()
298 _sum = _mm256_fmadd_ps(_k03, _r03, _sum); in conv2x2s1_pack8_avx()
|
| H A D | convolution_2x2_pack8_fp16.h | 247 __m256 _k03 = loadfp16(kptr + 24); in conv2x2s1_fp16_pack8_avx() local
253 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in conv2x2s1_fp16_pack8_avx()
281 _sum1 = _mm256_fmadd_ps(_k03, _r03, _sum1); in conv2x2s1_fp16_pack8_avx()
290 _k03 = loadfp16(kptr + 24); in conv2x2s1_fp16_pack8_avx()
296 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in conv2x2s1_fp16_pack8_avx()
321 _sum1 = _mm256_fmadd_ps(_k03, _r03, _sum1); in conv2x2s1_fp16_pack8_avx()
440 __m256 _k03 = loadfp16(kptr + 24); in conv2x2s1_fp16_pack8_avx() local
446 _sum = _mm256_fmadd_ps(_k03, _r03, _sum); in conv2x2s1_fp16_pack8_avx()
473 _k03 = loadfp16(kptr + 24); in conv2x2s1_fp16_pack8_avx()
479 _sum = _mm256_fmadd_ps(_k03, _r03, _sum); in conv2x2s1_fp16_pack8_avx()
|
| H A D | convolutiondepthwise_5x5_pack8.h | 60 __m256 _k03 = _mm256_loadu_ps(k0 + 24); in convdw5x5s1_pack8_avx() local
67 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in convdw5x5s1_pack8_avx()
214 __m256 _k03 = _mm256_loadu_ps(k0 + 24); in convdw5x5s2_pack8_avx() local
221 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in convdw5x5s2_pack8_avx()
|
| /dports/misc/ncnn/ncnn-20211208/src/layer/x86/ |
| H A D | convolution_2x2_pack8.h | 66 __m256 _k03 = _mm256_loadu_ps(kptr + 24); in conv2x2s1_pack8_avx() local
72 _sum0 = _mm256_comp_fmadd_ps(_k03, _r03, _sum0); in conv2x2s1_pack8_avx()
100 _sum1 = _mm256_comp_fmadd_ps(_k03, _r03, _sum1); in conv2x2s1_pack8_avx()
109 _k03 = _mm256_loadu_ps(kptr + 24); in conv2x2s1_pack8_avx()
115 _sum0 = _mm256_comp_fmadd_ps(_k03, _r03, _sum0); in conv2x2s1_pack8_avx()
140 _sum1 = _mm256_comp_fmadd_ps(_k03, _r03, _sum1); in conv2x2s1_pack8_avx()
259 __m256 _k03 = _mm256_loadu_ps(kptr + 24); in conv2x2s1_pack8_avx() local
265 _sum = _mm256_comp_fmadd_ps(_k03, _r03, _sum); in conv2x2s1_pack8_avx()
292 _k03 = _mm256_loadu_ps(kptr + 24); in conv2x2s1_pack8_avx()
298 _sum = _mm256_comp_fmadd_ps(_k03, _r03, _sum); in conv2x2s1_pack8_avx()
|
| H A D | convolution_2x2_pack8_fp16.h | 247 __m256 _k03 = loadfp16(kptr + 24); in conv2x2s1_fp16_pack8_avx() local
253 _sum0 = _mm256_comp_fmadd_ps(_k03, _r03, _sum0); in conv2x2s1_fp16_pack8_avx()
281 _sum1 = _mm256_comp_fmadd_ps(_k03, _r03, _sum1); in conv2x2s1_fp16_pack8_avx()
290 _k03 = loadfp16(kptr + 24); in conv2x2s1_fp16_pack8_avx()
296 _sum0 = _mm256_comp_fmadd_ps(_k03, _r03, _sum0); in conv2x2s1_fp16_pack8_avx()
321 _sum1 = _mm256_comp_fmadd_ps(_k03, _r03, _sum1); in conv2x2s1_fp16_pack8_avx()
440 __m256 _k03 = loadfp16(kptr + 24); in conv2x2s1_fp16_pack8_avx() local
446 _sum = _mm256_comp_fmadd_ps(_k03, _r03, _sum); in conv2x2s1_fp16_pack8_avx()
473 _k03 = loadfp16(kptr + 24); in conv2x2s1_fp16_pack8_avx()
479 _sum = _mm256_comp_fmadd_ps(_k03, _r03, _sum); in conv2x2s1_fp16_pack8_avx()
|
| H A D | convolutiondepthwise_5x5_pack8.h | 60 __m256 _k03 = _mm256_loadu_ps(k0 + 24); in convdw5x5s1_pack8_avx() local
67 _sum0 = _mm256_comp_fmadd_ps(_k03, _r03, _sum0); in convdw5x5s1_pack8_avx()
214 __m256 _k03 = _mm256_loadu_ps(k0 + 24); in convdw5x5s2_pack8_avx() local
221 _sum0 = _mm256_comp_fmadd_ps(_k03, _r03, _sum0); in convdw5x5s2_pack8_avx()
|
| /dports/benchmarks/vkpeak/vkpeak-20210430/ncnn/src/layer/x86/ |
| H A D | convolution_2x2_pack8.h | 66 __m256 _k03 = _mm256_loadu_ps(kptr + 24); in conv2x2s1_pack8_avx() local
72 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in conv2x2s1_pack8_avx()
100 _sum1 = _mm256_fmadd_ps(_k03, _r03, _sum1); in conv2x2s1_pack8_avx()
109 _k03 = _mm256_loadu_ps(kptr + 24); in conv2x2s1_pack8_avx()
115 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in conv2x2s1_pack8_avx()
140 _sum1 = _mm256_fmadd_ps(_k03, _r03, _sum1); in conv2x2s1_pack8_avx()
259 __m256 _k03 = _mm256_loadu_ps(kptr + 24); in conv2x2s1_pack8_avx() local
265 _sum = _mm256_fmadd_ps(_k03, _r03, _sum); in conv2x2s1_pack8_avx()
292 _k03 = _mm256_loadu_ps(kptr + 24); in conv2x2s1_pack8_avx()
298 _sum = _mm256_fmadd_ps(_k03, _r03, _sum); in conv2x2s1_pack8_avx()
|
| H A D | convolution_2x2_pack8_fp16.h | 247 __m256 _k03 = loadfp16(kptr + 24); in conv2x2s1_fp16_pack8_avx() local
253 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in conv2x2s1_fp16_pack8_avx()
281 _sum1 = _mm256_fmadd_ps(_k03, _r03, _sum1); in conv2x2s1_fp16_pack8_avx()
290 _k03 = loadfp16(kptr + 24); in conv2x2s1_fp16_pack8_avx()
296 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in conv2x2s1_fp16_pack8_avx()
321 _sum1 = _mm256_fmadd_ps(_k03, _r03, _sum1); in conv2x2s1_fp16_pack8_avx()
440 __m256 _k03 = loadfp16(kptr + 24); in conv2x2s1_fp16_pack8_avx() local
446 _sum = _mm256_fmadd_ps(_k03, _r03, _sum); in conv2x2s1_fp16_pack8_avx()
473 _k03 = loadfp16(kptr + 24); in conv2x2s1_fp16_pack8_avx()
479 _sum = _mm256_fmadd_ps(_k03, _r03, _sum); in conv2x2s1_fp16_pack8_avx()
|
| H A D | convolutiondepthwise_5x5_pack8.h | 60 __m256 _k03 = _mm256_loadu_ps(k0 + 24); in convdw5x5s1_pack8_avx() local
67 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in convdw5x5s1_pack8_avx()
214 __m256 _k03 = _mm256_loadu_ps(k0 + 24); in convdw5x5s2_pack8_avx() local
221 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in convdw5x5s2_pack8_avx()
|
| /dports/graphics/realsr-ncnn-vulkan/realsr-ncnn-vulkan-20210210/src/ncnn/src/layer/x86/ |
| H A D | convolution_2x2_pack8.h | 66 __m256 _k03 = _mm256_loadu_ps(kptr + 24); in conv2x2s1_pack8_avx() local
72 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in conv2x2s1_pack8_avx()
100 _sum1 = _mm256_fmadd_ps(_k03, _r03, _sum1); in conv2x2s1_pack8_avx()
109 _k03 = _mm256_loadu_ps(kptr + 24); in conv2x2s1_pack8_avx()
115 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in conv2x2s1_pack8_avx()
140 _sum1 = _mm256_fmadd_ps(_k03, _r03, _sum1); in conv2x2s1_pack8_avx()
259 __m256 _k03 = _mm256_loadu_ps(kptr + 24); in conv2x2s1_pack8_avx() local
265 _sum = _mm256_fmadd_ps(_k03, _r03, _sum); in conv2x2s1_pack8_avx()
292 _k03 = _mm256_loadu_ps(kptr + 24); in conv2x2s1_pack8_avx()
298 _sum = _mm256_fmadd_ps(_k03, _r03, _sum); in conv2x2s1_pack8_avx()
|
| H A D | convolution_2x2_pack8_fp16.h | 247 __m256 _k03 = loadfp16(kptr + 24); in conv2x2s1_fp16_pack8_avx() local
253 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in conv2x2s1_fp16_pack8_avx()
281 _sum1 = _mm256_fmadd_ps(_k03, _r03, _sum1); in conv2x2s1_fp16_pack8_avx()
290 _k03 = loadfp16(kptr + 24); in conv2x2s1_fp16_pack8_avx()
296 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in conv2x2s1_fp16_pack8_avx()
321 _sum1 = _mm256_fmadd_ps(_k03, _r03, _sum1); in conv2x2s1_fp16_pack8_avx()
440 __m256 _k03 = loadfp16(kptr + 24); in conv2x2s1_fp16_pack8_avx() local
446 _sum = _mm256_fmadd_ps(_k03, _r03, _sum); in conv2x2s1_fp16_pack8_avx()
473 _k03 = loadfp16(kptr + 24); in conv2x2s1_fp16_pack8_avx()
479 _sum = _mm256_fmadd_ps(_k03, _r03, _sum); in conv2x2s1_fp16_pack8_avx()
|
| H A D | convolutiondepthwise_5x5_pack8.h | 60 __m256 _k03 = _mm256_loadu_ps(k0 + 24); in convdw5x5s1_pack8_avx() local
67 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in convdw5x5s1_pack8_avx()
214 __m256 _k03 = _mm256_loadu_ps(k0 + 24); in convdw5x5s2_pack8_avx() local
221 _sum0 = _mm256_fmadd_ps(_k03, _r03, _sum0); in convdw5x5s2_pack8_avx()
|
| /dports/graphics/vapoursynth-waifu2x-ncnn-vulkan/vapoursynth-waifu2x-ncnn-vulkan-r4/deps/ncnn/src/layer/arm/ |
| H A D | convolutiondepthwise_5x5_pack4.h | 80 float32x4_t _k03 = vld1q_f32(k0 + 12); in convdw5x5s1_pack4_neon() local
87 _sum00 = vmlaq_f32(_sum00, _k03, _r03); in convdw5x5s1_pack4_neon()
92 _sum01 = vmlaq_f32(_sum01, _k03, _r04); in convdw5x5s1_pack4_neon()
97 _sum02 = vmlaq_f32(_sum02, _k03, _r05); in convdw5x5s1_pack4_neon()
102 _sum03 = vmlaq_f32(_sum03, _k03, _r06); in convdw5x5s1_pack4_neon()
124 _sum10 = vmlaq_f32(_sum10, _k03, _r13); in convdw5x5s1_pack4_neon()
129 _sum11 = vmlaq_f32(_sum11, _k03, _r14); in convdw5x5s1_pack4_neon()
134 _sum12 = vmlaq_f32(_sum12, _k03, _r15); in convdw5x5s1_pack4_neon()
139 _sum13 = vmlaq_f32(_sum13, _k03, _r16); in convdw5x5s1_pack4_neon()
614 _sum0 = vmlaq_f32(_sum0, _k03, _r03); in convdw5x5s1_pack4_neon()
[all …]
|
| /dports/graphics/waifu2x-ncnn-vulkan/waifu2x-ncnn-vulkan-20210521/src/ncnn/src/layer/arm/ |
| H A D | convolutiondepthwise_5x5_pack4.h | 80 float32x4_t _k03 = vld1q_f32(k0 + 12); in convdw5x5s1_pack4_neon() local
87 _sum00 = vmlaq_f32(_sum00, _k03, _r03); in convdw5x5s1_pack4_neon()
92 _sum01 = vmlaq_f32(_sum01, _k03, _r04); in convdw5x5s1_pack4_neon()
97 _sum02 = vmlaq_f32(_sum02, _k03, _r05); in convdw5x5s1_pack4_neon()
102 _sum03 = vmlaq_f32(_sum03, _k03, _r06); in convdw5x5s1_pack4_neon()
124 _sum10 = vmlaq_f32(_sum10, _k03, _r13); in convdw5x5s1_pack4_neon()
129 _sum11 = vmlaq_f32(_sum11, _k03, _r14); in convdw5x5s1_pack4_neon()
134 _sum12 = vmlaq_f32(_sum12, _k03, _r15); in convdw5x5s1_pack4_neon()
139 _sum13 = vmlaq_f32(_sum13, _k03, _r16); in convdw5x5s1_pack4_neon()
614 _sum0 = vmlaq_f32(_sum0, _k03, _r03); in convdw5x5s1_pack4_neon()
[all …]
|
| /dports/benchmarks/vkpeak/vkpeak-20210430/ncnn/src/layer/arm/ |
| H A D | convolutiondepthwise_5x5_pack4.h | 80 float32x4_t _k03 = vld1q_f32(k0 + 12); in convdw5x5s1_pack4_neon() local
87 _sum00 = vmlaq_f32(_sum00, _k03, _r03); in convdw5x5s1_pack4_neon()
92 _sum01 = vmlaq_f32(_sum01, _k03, _r04); in convdw5x5s1_pack4_neon()
97 _sum02 = vmlaq_f32(_sum02, _k03, _r05); in convdw5x5s1_pack4_neon()
102 _sum03 = vmlaq_f32(_sum03, _k03, _r06); in convdw5x5s1_pack4_neon()
124 _sum10 = vmlaq_f32(_sum10, _k03, _r13); in convdw5x5s1_pack4_neon()
129 _sum11 = vmlaq_f32(_sum11, _k03, _r14); in convdw5x5s1_pack4_neon()
134 _sum12 = vmlaq_f32(_sum12, _k03, _r15); in convdw5x5s1_pack4_neon()
139 _sum13 = vmlaq_f32(_sum13, _k03, _r16); in convdw5x5s1_pack4_neon()
614 _sum0 = vmlaq_f32(_sum0, _k03, _r03); in convdw5x5s1_pack4_neon()
[all …]
|
| /dports/misc/ncnn/ncnn-20211208/src/layer/arm/ |
| H A D | convolutiondepthwise_5x5_pack4.h | 80 float32x4_t _k03 = vld1q_f32(k0 + 12); in convdw5x5s1_pack4_neon() local
87 _sum00 = vmlaq_f32(_sum00, _k03, _r03); in convdw5x5s1_pack4_neon()
92 _sum01 = vmlaq_f32(_sum01, _k03, _r04); in convdw5x5s1_pack4_neon()
97 _sum02 = vmlaq_f32(_sum02, _k03, _r05); in convdw5x5s1_pack4_neon()
102 _sum03 = vmlaq_f32(_sum03, _k03, _r06); in convdw5x5s1_pack4_neon()
124 _sum10 = vmlaq_f32(_sum10, _k03, _r13); in convdw5x5s1_pack4_neon()
129 _sum11 = vmlaq_f32(_sum11, _k03, _r14); in convdw5x5s1_pack4_neon()
134 _sum12 = vmlaq_f32(_sum12, _k03, _r15); in convdw5x5s1_pack4_neon()
139 _sum13 = vmlaq_f32(_sum13, _k03, _r16); in convdw5x5s1_pack4_neon()
614 _sum0 = vmlaq_f32(_sum0, _k03, _r03); in convdw5x5s1_pack4_neon()
[all …]
|
| /dports/graphics/realsr-ncnn-vulkan/realsr-ncnn-vulkan-20210210/src/ncnn/src/layer/arm/ |
| H A D | convolutiondepthwise_5x5_pack4.h | 80 float32x4_t _k03 = vld1q_f32(k0 + 12); in convdw5x5s1_pack4_neon() local
87 _sum00 = vmlaq_f32(_sum00, _k03, _r03); in convdw5x5s1_pack4_neon()
92 _sum01 = vmlaq_f32(_sum01, _k03, _r04); in convdw5x5s1_pack4_neon()
97 _sum02 = vmlaq_f32(_sum02, _k03, _r05); in convdw5x5s1_pack4_neon()
102 _sum03 = vmlaq_f32(_sum03, _k03, _r06); in convdw5x5s1_pack4_neon()
124 _sum10 = vmlaq_f32(_sum10, _k03, _r13); in convdw5x5s1_pack4_neon()
129 _sum11 = vmlaq_f32(_sum11, _k03, _r14); in convdw5x5s1_pack4_neon()
134 _sum12 = vmlaq_f32(_sum12, _k03, _r15); in convdw5x5s1_pack4_neon()
139 _sum13 = vmlaq_f32(_sum13, _k03, _r16); in convdw5x5s1_pack4_neon()
614 _sum0 = vmlaq_f32(_sum0, _k03, _r03); in convdw5x5s1_pack4_neon()
[all …]
|
| /dports/misc/ncnn/ncnn-20211208/src/layer/mips/ |
| H A D | convolutiondepthwise_5x5_pack4.h | 74 v4f32 _k03 = (v4f32)__msa_ld_w(k0 + 4 * 3, 0); in convdw5x5s1_pack4_msa() local
81 _sum0 = __msa_fmadd_w(_sum0, _k03, _r03); in convdw5x5s1_pack4_msa()
93 _sum1 = __msa_fmadd_w(_sum1, _k03, _r13); in convdw5x5s1_pack4_msa()
244 v4f32 _k03 = (v4f32)__msa_ld_w(k0 + 4 * 3, 0); in convdw5x5s1_pack4_msa() local
251 _sum0 = __msa_fmadd_w(_sum0, _k03, _r03); in convdw5x5s1_pack4_msa()
407 v4f32 _k03 = (v4f32)__msa_ld_w(k0 + 4 * 3, 0); in convdw5x5s2_pack4_msa() local
414 _sum0 = __msa_fmadd_w(_sum0, _k03, _r03); in convdw5x5s2_pack4_msa()
|
| /dports/misc/ncnn/ncnn-20211208/src/layer/riscv/ |
| H A D | convolutiondepthwise_5x5_packn.h | 68 vfloat32m1_t _k03 = vle32_v_f32m1(k0 + packn * 3, vl); in convdw5x5s1_packn_rvv() local
75 _sum0 = vfmacc_vv_f32m1(_sum0, _k03, _r03, vl); in convdw5x5s1_packn_rvv()
87 _sum1 = vfmacc_vv_f32m1(_sum1, _k03, _r13, vl); in convdw5x5s1_packn_rvv()
230 vfloat32m1_t _k03 = vle32_v_f32m1(k0 + packn * 3, vl); in convdw5x5s1_packn_rvv() local
237 _sum0 = vfmacc_vv_f32m1(_sum0, _k03, _r03, vl); in convdw5x5s1_packn_rvv()
388 vfloat32m1_t _k03 = vle32_v_f32m1(k0 + packn * 3, vl); in convdw5x5s2_packn_rvv() local
395 _sum0 = vfmacc_vv_f32m1(_sum0, _k03, _r03, vl); in convdw5x5s2_packn_rvv()
|
| H A D | convolutiondepthwise_5x5_packn_fp16s.h | 68 vfloat16m1_t _k03 = vle16_v_f16m1(k0 + packn * 3, vl); in convdw5x5s1_packn_fp16sa_rvv() local
75 _sum0 = vfmacc_vv_f16m1(_sum0, _k03, _r03, vl); in convdw5x5s1_packn_fp16sa_rvv()
87 _sum1 = vfmacc_vv_f16m1(_sum1, _k03, _r13, vl); in convdw5x5s1_packn_fp16sa_rvv()
230 vfloat16m1_t _k03 = vle16_v_f16m1(k0 + packn * 3, vl); in convdw5x5s1_packn_fp16sa_rvv() local
237 _sum0 = vfmacc_vv_f16m1(_sum0, _k03, _r03, vl); in convdw5x5s1_packn_fp16sa_rvv()
388 vfloat16m1_t _k03 = vle16_v_f16m1(k0 + packn * 3, vl); in convdw5x5s2_packn_fp16sa_rvv() local
395 _sum0 = vfmacc_vv_f16m1(_sum0, _k03, _r03, vl); in convdw5x5s2_packn_fp16sa_rvv()
|
| H A D | convolution_7x7_pack1ton.h | 105 _sum0 = vfmacc_vf_f32m1(_sum0, r0[3], _k03, vl); in conv7x7s2_pack1ton_rvv()
106 _sum1 = vfmacc_vf_f32m1(_sum1, r0[5], _k03, vl); in conv7x7s2_pack1ton_rvv()
107 _sum2 = vfmacc_vf_f32m1(_sum2, r0[7], _k03, vl); in conv7x7s2_pack1ton_rvv()
108 _sum3 = vfmacc_vf_f32m1(_sum3, r0[9], _k03, vl); in conv7x7s2_pack1ton_rvv()
109 _sum4 = vfmacc_vf_f32m1(_sum4, r0[11], _k03, vl); in conv7x7s2_pack1ton_rvv()
110 _sum5 = vfmacc_vf_f32m1(_sum5, r0[13], _k03, vl); in conv7x7s2_pack1ton_rvv()
111 _sum6 = vfmacc_vf_f32m1(_sum6, r0[15], _k03, vl); in conv7x7s2_pack1ton_rvv()
112 _sum7 = vfmacc_vf_f32m1(_sum7, r0[17], _k03, vl); in conv7x7s2_pack1ton_rvv()
588 _sum0 = vfmacc_vf_f32m1(_sum0, r0[3], _k03, vl); in conv7x7s2_pack1ton_rvv()
589 _sum1 = vfmacc_vf_f32m1(_sum1, r0[5], _k03, vl); in conv7x7s2_pack1ton_rvv()
[all …]
|
| H A D | convolution_7x7_pack1ton_fp16s.h | 105 _sum0 = vfmacc_vf_f16m1(_sum0, r0[3], _k03, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
106 _sum1 = vfmacc_vf_f16m1(_sum1, r0[5], _k03, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
107 _sum2 = vfmacc_vf_f16m1(_sum2, r0[7], _k03, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
108 _sum3 = vfmacc_vf_f16m1(_sum3, r0[9], _k03, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
109 _sum4 = vfmacc_vf_f16m1(_sum4, r0[11], _k03, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
110 _sum5 = vfmacc_vf_f16m1(_sum5, r0[13], _k03, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
111 _sum6 = vfmacc_vf_f16m1(_sum6, r0[15], _k03, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
112 _sum7 = vfmacc_vf_f16m1(_sum7, r0[17], _k03, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
588 _sum0 = vfmacc_vf_f16m1(_sum0, r0[3], _k03, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
589 _sum1 = vfmacc_vf_f16m1(_sum1, r0[5], _k03, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
[all …]
|