| /dports/misc/ncnn/ncnn-20211208/src/layer/riscv/ |
| H A D | convolution_7x7_pack1ton.h | 432 _sum0 = vfmacc_vf_f32m1(_sum0, r5[2], _k52, vl); in conv7x7s2_pack1ton_rvv()
433 _sum1 = vfmacc_vf_f32m1(_sum1, r5[4], _k52, vl); in conv7x7s2_pack1ton_rvv()
434 _sum2 = vfmacc_vf_f32m1(_sum2, r5[6], _k52, vl); in conv7x7s2_pack1ton_rvv()
435 _sum3 = vfmacc_vf_f32m1(_sum3, r5[8], _k52, vl); in conv7x7s2_pack1ton_rvv()
436 _sum4 = vfmacc_vf_f32m1(_sum4, r5[10], _k52, vl); in conv7x7s2_pack1ton_rvv()
437 _sum5 = vfmacc_vf_f32m1(_sum5, r5[12], _k52, vl); in conv7x7s2_pack1ton_rvv()
438 _sum6 = vfmacc_vf_f32m1(_sum6, r5[14], _k52, vl); in conv7x7s2_pack1ton_rvv()
779 _sum0 = vfmacc_vf_f32m1(_sum0, r5[2], _k52, vl); in conv7x7s2_pack1ton_rvv()
780 _sum1 = vfmacc_vf_f32m1(_sum1, r5[4], _k52, vl); in conv7x7s2_pack1ton_rvv()
781 _sum2 = vfmacc_vf_f32m1(_sum2, r5[6], _k52, vl); in conv7x7s2_pack1ton_rvv()
[all …]
|
| H A D | convolution_7x7_pack1ton_fp16s.h | 432 _sum0 = vfmacc_vf_f16m1(_sum0, r5[2], _k52, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
433 _sum1 = vfmacc_vf_f16m1(_sum1, r5[4], _k52, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
434 _sum2 = vfmacc_vf_f16m1(_sum2, r5[6], _k52, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
435 _sum3 = vfmacc_vf_f16m1(_sum3, r5[8], _k52, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
436 _sum4 = vfmacc_vf_f16m1(_sum4, r5[10], _k52, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
437 _sum5 = vfmacc_vf_f16m1(_sum5, r5[12], _k52, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
438 _sum6 = vfmacc_vf_f16m1(_sum6, r5[14], _k52, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
779 _sum0 = vfmacc_vf_f16m1(_sum0, r5[2], _k52, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
780 _sum1 = vfmacc_vf_f16m1(_sum1, r5[4], _k52, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
781 _sum2 = vfmacc_vf_f16m1(_sum2, r5[6], _k52, vl); in conv7x7s2_pack1ton_fp16sa_rvv()
[all …]
|
| /dports/misc/ncnn/ncnn-20211208/src/layer/mips/ |
| H A D | convolution_7x7_pack1to4.h | 351 v4f32 _k52 = (v4f32)__msa_ld_w(kptr + 4 * 2, 0); in conv7x7s2_pack1to4_msa() local
385 _sum0 = __msa_fmadd_w(_sum0, _r52, _k52); in conv7x7s2_pack1to4_msa()
386 _sum1 = __msa_fmadd_w(_sum1, _r54, _k52); in conv7x7s2_pack1to4_msa()
387 _sum2 = __msa_fmadd_w(_sum2, _r56, _k52); in conv7x7s2_pack1to4_msa()
388 _sum3 = __msa_fmadd_w(_sum3, _r58, _k52); in conv7x7s2_pack1to4_msa()
589 v4f32 _k52 = (v4f32)__msa_ld_w(kptr + 4 * 2, 0); in conv7x7s2_pack1to4_msa() local
602 _sum0 = __msa_fmadd_w(_sum0, (v4f32)__msa_splati_w(_r5, 2), _k52); 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 | 266 __m256 _k52 = _mm256_loadu_ps(kptr + 16); in conv3x3s1_pack8_avx() local
277 _sum00 = _mm256_fmadd_ps(_r122, _k52, _sum00); in conv3x3s1_pack8_avx()
295 _sum10 = _mm256_fmadd_ps(_r132, _k52, _sum10); in conv3x3s1_pack8_avx()
597 __m256 _k52 = _mm256_loadu_ps(kptr + 16); in conv3x3s1_pack8_avx() local
608 _sum0 = _mm256_fmadd_ps(_r122, _k52, _sum0); 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 | 266 __m256 _k52 = _mm256_loadu_ps(kptr + 16); in conv3x3s1_pack8_avx() local
277 _sum00 = _mm256_fmadd_ps(_r122, _k52, _sum00); in conv3x3s1_pack8_avx()
295 _sum10 = _mm256_fmadd_ps(_r132, _k52, _sum10); in conv3x3s1_pack8_avx()
597 __m256 _k52 = _mm256_loadu_ps(kptr + 16); in conv3x3s1_pack8_avx() local
608 _sum0 = _mm256_fmadd_ps(_r122, _k52, _sum0); in conv3x3s1_pack8_avx()
|
| /dports/benchmarks/vkpeak/vkpeak-20210430/ncnn/src/layer/x86/ |
| H A D | convolution_3x3_pack8.h | 266 __m256 _k52 = _mm256_loadu_ps(kptr + 16); in conv3x3s1_pack8_avx() local
277 _sum00 = _mm256_fmadd_ps(_r122, _k52, _sum00); in conv3x3s1_pack8_avx()
295 _sum10 = _mm256_fmadd_ps(_r132, _k52, _sum10); in conv3x3s1_pack8_avx()
597 __m256 _k52 = _mm256_loadu_ps(kptr + 16); in conv3x3s1_pack8_avx() local
608 _sum0 = _mm256_fmadd_ps(_r122, _k52, _sum0); in conv3x3s1_pack8_avx()
|
| /dports/misc/ncnn/ncnn-20211208/src/layer/x86/ |
| H A D | convolution_3x3_pack8.h | 266 __m256 _k52 = _mm256_loadu_ps(kptr + 16); in conv3x3s1_pack8_avx() local
277 _sum00 = _mm256_comp_fmadd_ps(_r122, _k52, _sum00); in conv3x3s1_pack8_avx()
295 _sum10 = _mm256_comp_fmadd_ps(_r132, _k52, _sum10); in conv3x3s1_pack8_avx()
597 __m256 _k52 = _mm256_loadu_ps(kptr + 16); in conv3x3s1_pack8_avx() local
608 _sum0 = _mm256_comp_fmadd_ps(_r122, _k52, _sum0); in conv3x3s1_pack8_avx()
|