| /dports/misc/ncnn/ncnn-20211208/src/layer/riscv/ |
| H A D | convolution_7x7_pack1ton.h | 440 _sum0 = vfmacc_vf_f32m1(_sum0, r5[3], _k53, vl); in conv7x7s2_pack1ton_rvv()
441 _sum1 = vfmacc_vf_f32m1(_sum1, r5[5], _k53, vl); in conv7x7s2_pack1ton_rvv()
442 _sum2 = vfmacc_vf_f32m1(_sum2, r5[7], _k53, vl); in conv7x7s2_pack1ton_rvv()
443 _sum3 = vfmacc_vf_f32m1(_sum3, r5[9], _k53, vl); in conv7x7s2_pack1ton_rvv()
444 _sum4 = vfmacc_vf_f32m1(_sum4, r5[11], _k53, vl); in conv7x7s2_pack1ton_rvv()
445 _sum5 = vfmacc_vf_f32m1(_sum5, r5[13], _k53, vl); in conv7x7s2_pack1ton_rvv()
446 _sum6 = vfmacc_vf_f32m1(_sum6, r5[15], _k53, vl); in conv7x7s2_pack1ton_rvv()
783 _sum0 = vfmacc_vf_f32m1(_sum0, r5[3], _k53, vl); in conv7x7s2_pack1ton_rvv()
784 _sum1 = vfmacc_vf_f32m1(_sum1, r5[5], _k53, vl); in conv7x7s2_pack1ton_rvv()
785 _sum2 = vfmacc_vf_f32m1(_sum2, r5[7], _k53, vl); in conv7x7s2_pack1ton_rvv()
[all …]
|
| H A D | convolution_7x7_pack1ton_fp16s.h | 440 _sum0 = vfmacc_vf_f16m1(_sum0, r5[3], _k53, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
441 _sum1 = vfmacc_vf_f16m1(_sum1, r5[5], _k53, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
442 _sum2 = vfmacc_vf_f16m1(_sum2, r5[7], _k53, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
443 _sum3 = vfmacc_vf_f16m1(_sum3, r5[9], _k53, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
444 _sum4 = vfmacc_vf_f16m1(_sum4, r5[11], _k53, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
445 _sum5 = vfmacc_vf_f16m1(_sum5, r5[13], _k53, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
446 _sum6 = vfmacc_vf_f16m1(_sum6, r5[15], _k53, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
783 _sum0 = vfmacc_vf_f16m1(_sum0, r5[3], _k53, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
784 _sum1 = vfmacc_vf_f16m1(_sum1, r5[5], _k53, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
785 _sum2 = vfmacc_vf_f16m1(_sum2, r5[7], _k53, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
[all …]
|
| /dports/misc/ncnn/ncnn-20211208/src/layer/mips/ |
| H A D | convolution_7x7_pack1to4.h | 352 v4f32 _k53 = (v4f32)__msa_ld_w(kptr + 4 * 3, 0); in conv7x7s2_pack1to4_msa() local
389 _sum0 = __msa_fmadd_w(_sum0, _r53, _k53); in conv7x7s2_pack1to4_msa()
390 _sum1 = __msa_fmadd_w(_sum1, _r55, _k53); in conv7x7s2_pack1to4_msa()
391 _sum2 = __msa_fmadd_w(_sum2, _r57, _k53); in conv7x7s2_pack1to4_msa()
392 _sum3 = __msa_fmadd_w(_sum3, _r59, _k53); in conv7x7s2_pack1to4_msa()
590 v4f32 _k53 = (v4f32)__msa_ld_w(kptr + 4 * 3, 0); in conv7x7s2_pack1to4_msa() local
603 _sum0 = __msa_fmadd_w(_sum0, (v4f32)__msa_splati_w(_r5, 3), _k53); in conv7x7s2_pack1to4_msa()
|
| /dports/graphics/vapoursynth-waifu2x-ncnn-vulkan/vapoursynth-waifu2x-ncnn-vulkan-r4/deps/ncnn/src/layer/x86/ |
| H A D | convolution_3x3_pack8.h | 267 __m256 _k53 = _mm256_loadu_ps(kptr + 24); in conv3x3s1_pack8_avx() local
278 _sum01 = _mm256_fmadd_ps(_r123, _k53, _sum01); in conv3x3s1_pack8_avx()
296 _sum11 = _mm256_fmadd_ps(_r133, _k53, _sum11); in conv3x3s1_pack8_avx()
598 __m256 _k53 = _mm256_loadu_ps(kptr + 24); in conv3x3s1_pack8_avx() local
609 _sum1 = _mm256_fmadd_ps(_r123, _k53, _sum1); in conv3x3s1_pack8_avx()
|
| /dports/graphics/waifu2x-ncnn-vulkan/waifu2x-ncnn-vulkan-20210521/src/ncnn/src/layer/x86/ |
| H A D | convolution_3x3_pack8.h | 267 __m256 _k53 = _mm256_loadu_ps(kptr + 24); in conv3x3s1_pack8_avx() local
278 _sum01 = _mm256_fmadd_ps(_r123, _k53, _sum01); in conv3x3s1_pack8_avx()
296 _sum11 = _mm256_fmadd_ps(_r133, _k53, _sum11); in conv3x3s1_pack8_avx()
598 __m256 _k53 = _mm256_loadu_ps(kptr + 24); in conv3x3s1_pack8_avx() local
609 _sum1 = _mm256_fmadd_ps(_r123, _k53, _sum1); in conv3x3s1_pack8_avx()
|
| /dports/benchmarks/vkpeak/vkpeak-20210430/ncnn/src/layer/x86/ |
| H A D | convolution_3x3_pack8.h | 267 __m256 _k53 = _mm256_loadu_ps(kptr + 24); in conv3x3s1_pack8_avx() local
278 _sum01 = _mm256_fmadd_ps(_r123, _k53, _sum01); in conv3x3s1_pack8_avx()
296 _sum11 = _mm256_fmadd_ps(_r133, _k53, _sum11); in conv3x3s1_pack8_avx()
598 __m256 _k53 = _mm256_loadu_ps(kptr + 24); in conv3x3s1_pack8_avx() local
609 _sum1 = _mm256_fmadd_ps(_r123, _k53, _sum1); in conv3x3s1_pack8_avx()
|
| /dports/misc/ncnn/ncnn-20211208/src/layer/x86/ |
| H A D | convolution_3x3_pack8.h | 267 __m256 _k53 = _mm256_loadu_ps(kptr + 24); in conv3x3s1_pack8_avx() local
278 _sum01 = _mm256_comp_fmadd_ps(_r123, _k53, _sum01); in conv3x3s1_pack8_avx()
296 _sum11 = _mm256_comp_fmadd_ps(_r133, _k53, _sum11); in conv3x3s1_pack8_avx()
598 __m256 _k53 = _mm256_loadu_ps(kptr + 24); in conv3x3s1_pack8_avx() local
609 _sum1 = _mm256_comp_fmadd_ps(_r123, _k53, _sum1); in conv3x3s1_pack8_avx()
|