/dports/graphics/vapoursynth-waifu2x-ncnn-vulkan/vapoursynth-waifu2x-ncnn-vulkan-r4/deps/ncnn/src/layer/arm/ |
H A D | convolutiondepthwise_5x5_pack4.h | 289 float32x4_t _k41 = vld1q_f32(k0 + 4); in convdw5x5s1_pack4_neon() local 533 float32x4_t _k41 = vld1q_f32(k0 + 4); in convdw5x5s1_pack4_neon() local 699 float32x4_t _k41 = vld1q_f32(k0 + 4); in convdw5x5s1_pack4_neon() local 712 _sum0 = vmlaq_f32(_sum0, _k41, _r41); in convdw5x5s1_pack4_neon() 724 _sum1 = vmlaq_f32(_sum1, _k41, _r51); in convdw5x5s1_pack4_neon() 922 float32x4_t _k41 = vld1q_f32(k0 + 4); in convdw5x5s1_pack4_neon() local 929 _sum0 = vmlaq_f32(_sum0, _k41, _r41); in convdw5x5s1_pack4_neon() 934 _sum1 = vmlaq_f32(_sum1, _k41, _r42); in convdw5x5s1_pack4_neon() 939 _sum2 = vmlaq_f32(_sum2, _k41, _r43); in convdw5x5s1_pack4_neon() 944 _sum3 = vmlaq_f32(_sum3, _k41, _r44); in convdw5x5s1_pack4_neon() [all …]
|
H A D | convolutiondepthwise_5x5_pack8_fp16s.h | 1396 float16x8_t _k41 = vld1q_f16(k0 + 8); in convdw5x5s2_pack8_fp16sa_neon() local 1403 _sum0 = vfmaq_f16(_sum0, _k41, _r41); in convdw5x5s2_pack8_fp16sa_neon()
|
/dports/graphics/waifu2x-ncnn-vulkan/waifu2x-ncnn-vulkan-20210521/src/ncnn/src/layer/arm/ |
H A D | convolutiondepthwise_5x5_pack4.h | 289 float32x4_t _k41 = vld1q_f32(k0 + 4); in convdw5x5s1_pack4_neon() local 533 float32x4_t _k41 = vld1q_f32(k0 + 4); in convdw5x5s1_pack4_neon() local 699 float32x4_t _k41 = vld1q_f32(k0 + 4); in convdw5x5s1_pack4_neon() local 712 _sum0 = vmlaq_f32(_sum0, _k41, _r41); in convdw5x5s1_pack4_neon() 724 _sum1 = vmlaq_f32(_sum1, _k41, _r51); in convdw5x5s1_pack4_neon() 922 float32x4_t _k41 = vld1q_f32(k0 + 4); in convdw5x5s1_pack4_neon() local 929 _sum0 = vmlaq_f32(_sum0, _k41, _r41); in convdw5x5s1_pack4_neon() 934 _sum1 = vmlaq_f32(_sum1, _k41, _r42); in convdw5x5s1_pack4_neon() 939 _sum2 = vmlaq_f32(_sum2, _k41, _r43); in convdw5x5s1_pack4_neon() 944 _sum3 = vmlaq_f32(_sum3, _k41, _r44); in convdw5x5s1_pack4_neon() [all …]
|
H A D | convolutiondepthwise_5x5_pack8_fp16s.h | 1396 float16x8_t _k41 = vld1q_f16(k0 + 8); in convdw5x5s2_pack8_fp16sa_neon() local 1403 _sum0 = vfmaq_f16(_sum0, _k41, _r41); in convdw5x5s2_pack8_fp16sa_neon()
|
/dports/benchmarks/vkpeak/vkpeak-20210430/ncnn/src/layer/arm/ |
H A D | convolutiondepthwise_5x5_pack4.h | 289 float32x4_t _k41 = vld1q_f32(k0 + 4); in convdw5x5s1_pack4_neon() local 533 float32x4_t _k41 = vld1q_f32(k0 + 4); in convdw5x5s1_pack4_neon() local 699 float32x4_t _k41 = vld1q_f32(k0 + 4); in convdw5x5s1_pack4_neon() local 712 _sum0 = vmlaq_f32(_sum0, _k41, _r41); in convdw5x5s1_pack4_neon() 724 _sum1 = vmlaq_f32(_sum1, _k41, _r51); in convdw5x5s1_pack4_neon() 922 float32x4_t _k41 = vld1q_f32(k0 + 4); in convdw5x5s1_pack4_neon() local 929 _sum0 = vmlaq_f32(_sum0, _k41, _r41); in convdw5x5s1_pack4_neon() 934 _sum1 = vmlaq_f32(_sum1, _k41, _r42); in convdw5x5s1_pack4_neon() 939 _sum2 = vmlaq_f32(_sum2, _k41, _r43); in convdw5x5s1_pack4_neon() 944 _sum3 = vmlaq_f32(_sum3, _k41, _r44); in convdw5x5s1_pack4_neon() [all …]
|
H A D | convolutiondepthwise_5x5_pack8_fp16s.h | 1396 float16x8_t _k41 = vld1q_f16(k0 + 8); in convdw5x5s2_pack8_fp16sa_neon() local 1403 _sum0 = vfmaq_f16(_sum0, _k41, _r41); in convdw5x5s2_pack8_fp16sa_neon()
|
/dports/misc/ncnn/ncnn-20211208/src/layer/arm/ |
H A D | convolutiondepthwise_5x5_pack4.h | 289 float32x4_t _k41 = vld1q_f32(k0 + 4); in convdw5x5s1_pack4_neon() local 533 float32x4_t _k41 = vld1q_f32(k0 + 4); in convdw5x5s1_pack4_neon() local 699 float32x4_t _k41 = vld1q_f32(k0 + 4); in convdw5x5s1_pack4_neon() local 712 _sum0 = vmlaq_f32(_sum0, _k41, _r41); in convdw5x5s1_pack4_neon() 724 _sum1 = vmlaq_f32(_sum1, _k41, _r51); in convdw5x5s1_pack4_neon() 922 float32x4_t _k41 = vld1q_f32(k0 + 4); in convdw5x5s1_pack4_neon() local 929 _sum0 = vmlaq_f32(_sum0, _k41, _r41); in convdw5x5s1_pack4_neon() 934 _sum1 = vmlaq_f32(_sum1, _k41, _r42); in convdw5x5s1_pack4_neon() 939 _sum2 = vmlaq_f32(_sum2, _k41, _r43); in convdw5x5s1_pack4_neon() 944 _sum3 = vmlaq_f32(_sum3, _k41, _r44); in convdw5x5s1_pack4_neon() [all …]
|
H A D | convolutiondepthwise_5x5_pack8_fp16s.h | 1396 float16x8_t _k41 = vld1q_f16(k0 + 8); in convdw5x5s2_pack8_fp16sa_neon() local 1403 _sum0 = vfmaq_f16(_sum0, _k41, _r41); in convdw5x5s2_pack8_fp16sa_neon()
|
/dports/graphics/realsr-ncnn-vulkan/realsr-ncnn-vulkan-20210210/src/ncnn/src/layer/arm/ |
H A D | convolutiondepthwise_5x5_pack4.h | 289 float32x4_t _k41 = vld1q_f32(k0 + 4); in convdw5x5s1_pack4_neon() local 533 float32x4_t _k41 = vld1q_f32(k0 + 4); in convdw5x5s1_pack4_neon() local 699 float32x4_t _k41 = vld1q_f32(k0 + 4); in convdw5x5s1_pack4_neon() local 712 _sum0 = vmlaq_f32(_sum0, _k41, _r41); in convdw5x5s1_pack4_neon() 724 _sum1 = vmlaq_f32(_sum1, _k41, _r51); in convdw5x5s1_pack4_neon() 922 float32x4_t _k41 = vld1q_f32(k0 + 4); in convdw5x5s1_pack4_neon() local 929 _sum0 = vmlaq_f32(_sum0, _k41, _r41); in convdw5x5s1_pack4_neon() 934 _sum1 = vmlaq_f32(_sum1, _k41, _r42); in convdw5x5s1_pack4_neon() 939 _sum2 = vmlaq_f32(_sum2, _k41, _r43); in convdw5x5s1_pack4_neon() 944 _sum3 = vmlaq_f32(_sum3, _k41, _r44); in convdw5x5s1_pack4_neon() [all …]
|
H A D | convolutiondepthwise_5x5_pack8_fp16s.h | 1396 float16x8_t _k41 = vld1q_f16(k0 + 8); in convdw5x5s2_pack8_fp16sa_neon() local 1403 _sum0 = vfmaq_f16(_sum0, _k41, _r41); in convdw5x5s2_pack8_fp16sa_neon()
|
/dports/misc/ncnn/ncnn-20211208/src/layer/mips/ |
H A D | convolutiondepthwise_5x5_pack4.h | 172 v4f32 _k41 = (v4f32)__msa_ld_w(k0 + 4, 0); in convdw5x5s1_pack4_msa() local 179 _sum0 = __msa_fmadd_w(_sum0, _k41, _r41); in convdw5x5s1_pack4_msa() 191 _sum1 = __msa_fmadd_w(_sum1, _k41, _r51); in convdw5x5s1_pack4_msa() 318 v4f32 _k41 = (v4f32)__msa_ld_w(k0 + 4, 0); in convdw5x5s1_pack4_msa() local 325 _sum0 = __msa_fmadd_w(_sum0, _k41, _r41); in convdw5x5s1_pack4_msa() 481 v4f32 _k41 = (v4f32)__msa_ld_w(k0 + 4, 0); in convdw5x5s2_pack4_msa() local 488 _sum0 = __msa_fmadd_w(_sum0, _k41, _r41); in convdw5x5s2_pack4_msa()
|
H A D | convolution_7x7_pack1to4.h | 293 v4f32 _k41 = (v4f32)__msa_ld_w(kptr + 4, 0); in conv7x7s2_pack1to4_msa() local 324 _sum0 = __msa_fmadd_w(_sum0, _r41, _k41); in conv7x7s2_pack1to4_msa() 325 _sum1 = __msa_fmadd_w(_sum1, _r43, _k41); in conv7x7s2_pack1to4_msa() 326 _sum2 = __msa_fmadd_w(_sum2, _r45, _k41); in conv7x7s2_pack1to4_msa() 327 _sum3 = __msa_fmadd_w(_sum3, _r47, _k41); in conv7x7s2_pack1to4_msa() 567 v4f32 _k41 = (v4f32)__msa_ld_w(kptr + 4, 0); in conv7x7s2_pack1to4_msa() local 580 _sum0 = __msa_fmadd_w(_sum0, (v4f32)__msa_splati_w(_r4, 1), _k41); in conv7x7s2_pack1to4_msa()
|
/dports/graphics/vapoursynth-waifu2x-ncnn-vulkan/vapoursynth-waifu2x-ncnn-vulkan-r4/deps/ncnn/src/layer/x86/ |
H A D | convolutiondepthwise_5x5_pack8.h | 134 __m256 _k41 = _mm256_loadu_ps(k0 + 8); in convdw5x5s1_pack8_avx() local 141 _sum0 = _mm256_fmadd_ps(_k41, _r41, _sum0); in convdw5x5s1_pack8_avx() 288 __m256 _k41 = _mm256_loadu_ps(k0 + 8); in convdw5x5s2_pack8_avx() local 295 _sum0 = _mm256_fmadd_ps(_k41, _r41, _sum0); in convdw5x5s2_pack8_avx()
|
H A D | convolution_3x3_pack8.h | 227 __m256 _k41 = _mm256_loadu_ps(kptr + 8); in conv3x3s1_pack8_avx() local 238 _sum01 = _mm256_fmadd_ps(_r111, _k41, _sum01); in conv3x3s1_pack8_avx() 256 _sum11 = _mm256_fmadd_ps(_r121, _k41, _sum11); in conv3x3s1_pack8_avx() 567 __m256 _k41 = _mm256_loadu_ps(kptr + 8); in conv3x3s1_pack8_avx() local 578 _sum1 = _mm256_fmadd_ps(_r111, _k41, _sum1); in conv3x3s1_pack8_avx()
|
/dports/graphics/waifu2x-ncnn-vulkan/waifu2x-ncnn-vulkan-20210521/src/ncnn/src/layer/x86/ |
H A D | convolutiondepthwise_5x5_pack8.h | 134 __m256 _k41 = _mm256_loadu_ps(k0 + 8); in convdw5x5s1_pack8_avx() local 141 _sum0 = _mm256_fmadd_ps(_k41, _r41, _sum0); in convdw5x5s1_pack8_avx() 288 __m256 _k41 = _mm256_loadu_ps(k0 + 8); in convdw5x5s2_pack8_avx() local 295 _sum0 = _mm256_fmadd_ps(_k41, _r41, _sum0); in convdw5x5s2_pack8_avx()
|
H A D | convolution_3x3_pack8.h | 227 __m256 _k41 = _mm256_loadu_ps(kptr + 8); in conv3x3s1_pack8_avx() local 238 _sum01 = _mm256_fmadd_ps(_r111, _k41, _sum01); in conv3x3s1_pack8_avx() 256 _sum11 = _mm256_fmadd_ps(_r121, _k41, _sum11); in conv3x3s1_pack8_avx() 567 __m256 _k41 = _mm256_loadu_ps(kptr + 8); in conv3x3s1_pack8_avx() local 578 _sum1 = _mm256_fmadd_ps(_r111, _k41, _sum1); in conv3x3s1_pack8_avx()
|
/dports/misc/ncnn/ncnn-20211208/src/layer/x86/ |
H A D | convolutiondepthwise_5x5_pack8.h | 134 __m256 _k41 = _mm256_loadu_ps(k0 + 8); in convdw5x5s1_pack8_avx() local 141 _sum0 = _mm256_comp_fmadd_ps(_k41, _r41, _sum0); in convdw5x5s1_pack8_avx() 288 __m256 _k41 = _mm256_loadu_ps(k0 + 8); in convdw5x5s2_pack8_avx() local 295 _sum0 = _mm256_comp_fmadd_ps(_k41, _r41, _sum0); in convdw5x5s2_pack8_avx()
|
H A D | convolution_3x3_pack8.h | 227 __m256 _k41 = _mm256_loadu_ps(kptr + 8); in conv3x3s1_pack8_avx() local 238 _sum01 = _mm256_comp_fmadd_ps(_r111, _k41, _sum01); in conv3x3s1_pack8_avx() 256 _sum11 = _mm256_comp_fmadd_ps(_r121, _k41, _sum11); in conv3x3s1_pack8_avx() 567 __m256 _k41 = _mm256_loadu_ps(kptr + 8); in conv3x3s1_pack8_avx() local 578 _sum1 = _mm256_comp_fmadd_ps(_r111, _k41, _sum1); in conv3x3s1_pack8_avx()
|
/dports/benchmarks/vkpeak/vkpeak-20210430/ncnn/src/layer/x86/ |
H A D | convolutiondepthwise_5x5_pack8.h | 134 __m256 _k41 = _mm256_loadu_ps(k0 + 8); in convdw5x5s1_pack8_avx() local 141 _sum0 = _mm256_fmadd_ps(_k41, _r41, _sum0); in convdw5x5s1_pack8_avx() 288 __m256 _k41 = _mm256_loadu_ps(k0 + 8); in convdw5x5s2_pack8_avx() local 295 _sum0 = _mm256_fmadd_ps(_k41, _r41, _sum0); in convdw5x5s2_pack8_avx()
|
H A D | convolution_3x3_pack8.h | 227 __m256 _k41 = _mm256_loadu_ps(kptr + 8); in conv3x3s1_pack8_avx() local 238 _sum01 = _mm256_fmadd_ps(_r111, _k41, _sum01); in conv3x3s1_pack8_avx() 256 _sum11 = _mm256_fmadd_ps(_r121, _k41, _sum11); in conv3x3s1_pack8_avx() 567 __m256 _k41 = _mm256_loadu_ps(kptr + 8); in conv3x3s1_pack8_avx() local 578 _sum1 = _mm256_fmadd_ps(_r111, _k41, _sum1); in conv3x3s1_pack8_avx()
|
/dports/graphics/realsr-ncnn-vulkan/realsr-ncnn-vulkan-20210210/src/ncnn/src/layer/x86/ |
H A D | convolutiondepthwise_5x5_pack8.h | 134 __m256 _k41 = _mm256_loadu_ps(k0 + 8); in convdw5x5s1_pack8_avx() local 141 _sum0 = _mm256_fmadd_ps(_k41, _r41, _sum0); in convdw5x5s1_pack8_avx() 288 __m256 _k41 = _mm256_loadu_ps(k0 + 8); in convdw5x5s2_pack8_avx() local 295 _sum0 = _mm256_fmadd_ps(_k41, _r41, _sum0); in convdw5x5s2_pack8_avx()
|
/dports/misc/ncnn/ncnn-20211208/src/layer/riscv/ |
H A D | convolutiondepthwise_5x5_packn.h | 166 vfloat32m1_t _k41 = vle32_v_f32m1(k0 + packn, vl); in convdw5x5s1_packn_rvv() local 173 _sum0 = vfmacc_vv_f32m1(_sum0, _k41, _r41, vl); in convdw5x5s1_packn_rvv() 185 _sum1 = vfmacc_vv_f32m1(_sum1, _k41, _r51, vl); in convdw5x5s1_packn_rvv() 304 vfloat32m1_t _k41 = vle32_v_f32m1(k0 + packn, vl); in convdw5x5s1_packn_rvv() local 311 _sum0 = vfmacc_vv_f32m1(_sum0, _k41, _r41, vl); in convdw5x5s1_packn_rvv() 462 vfloat32m1_t _k41 = vle32_v_f32m1(k0 + packn, vl); in convdw5x5s2_packn_rvv() local 469 _sum0 = vfmacc_vv_f32m1(_sum0, _k41, _r41, vl); in convdw5x5s2_packn_rvv()
|
H A D | convolutiondepthwise_5x5_packn_fp16s.h | 166 vfloat16m1_t _k41 = vle16_v_f16m1(k0 + packn, vl); in convdw5x5s1_packn_fp16sa_rvv() local 173 _sum0 = vfmacc_vv_f16m1(_sum0, _k41, _r41, vl); in convdw5x5s1_packn_fp16sa_rvv() 185 _sum1 = vfmacc_vv_f16m1(_sum1, _k41, _r51, vl); in convdw5x5s1_packn_fp16sa_rvv() 304 vfloat16m1_t _k41 = vle16_v_f16m1(k0 + packn, vl); in convdw5x5s1_packn_fp16sa_rvv() local 311 _sum0 = vfmacc_vv_f16m1(_sum0, _k41, _r41, vl); in convdw5x5s1_packn_fp16sa_rvv() 462 vfloat16m1_t _k41 = vle16_v_f16m1(k0 + packn, vl); in convdw5x5s2_packn_fp16sa_rvv() local 469 _sum0 = vfmacc_vv_f16m1(_sum0, _k41, _r41, vl); in convdw5x5s2_packn_fp16sa_rvv()
|
H A D | convolution_7x7_pack1ton.h | 357 _sum0 = vfmacc_vf_f32m1(_sum0, r4[1], _k41, vl); in conv7x7s2_pack1ton_rvv() 358 _sum1 = vfmacc_vf_f32m1(_sum1, r4[3], _k41, vl); in conv7x7s2_pack1ton_rvv() 359 _sum2 = vfmacc_vf_f32m1(_sum2, r4[5], _k41, vl); in conv7x7s2_pack1ton_rvv() 360 _sum3 = vfmacc_vf_f32m1(_sum3, r4[7], _k41, vl); in conv7x7s2_pack1ton_rvv() 361 _sum4 = vfmacc_vf_f32m1(_sum4, r4[9], _k41, vl); in conv7x7s2_pack1ton_rvv() 362 _sum5 = vfmacc_vf_f32m1(_sum5, r4[11], _k41, vl); in conv7x7s2_pack1ton_rvv() 363 _sum6 = vfmacc_vf_f32m1(_sum6, r4[13], _k41, vl); in conv7x7s2_pack1ton_rvv() 364 _sum7 = vfmacc_vf_f32m1(_sum7, r4[15], _k41, vl); in conv7x7s2_pack1ton_rvv() 736 _sum0 = vfmacc_vf_f32m1(_sum0, r4[1], _k41, vl); in conv7x7s2_pack1ton_rvv() 737 _sum1 = vfmacc_vf_f32m1(_sum1, r4[3], _k41, vl); in conv7x7s2_pack1ton_rvv() [all …]
|
H A D | convolution_7x7_pack1ton_fp16s.h | 357 _sum0 = vfmacc_vf_f16m1(_sum0, r4[1], _k41, vl); in conv7x7s2_pack1ton_fp16sa_rvv() 358 _sum1 = vfmacc_vf_f16m1(_sum1, r4[3], _k41, vl); in conv7x7s2_pack1ton_fp16sa_rvv() 359 _sum2 = vfmacc_vf_f16m1(_sum2, r4[5], _k41, vl); in conv7x7s2_pack1ton_fp16sa_rvv() 360 _sum3 = vfmacc_vf_f16m1(_sum3, r4[7], _k41, vl); in conv7x7s2_pack1ton_fp16sa_rvv() 361 _sum4 = vfmacc_vf_f16m1(_sum4, r4[9], _k41, vl); in conv7x7s2_pack1ton_fp16sa_rvv() 362 _sum5 = vfmacc_vf_f16m1(_sum5, r4[11], _k41, vl); in conv7x7s2_pack1ton_fp16sa_rvv() 363 _sum6 = vfmacc_vf_f16m1(_sum6, r4[13], _k41, vl); in conv7x7s2_pack1ton_fp16sa_rvv() 364 _sum7 = vfmacc_vf_f16m1(_sum7, r4[15], _k41, vl); in conv7x7s2_pack1ton_fp16sa_rvv() 736 _sum0 = vfmacc_vf_f16m1(_sum0, r4[1], _k41, vl); in conv7x7s2_pack1ton_fp16sa_rvv() 737 _sum1 = vfmacc_vf_f16m1(_sum1, r4[3], _k41, vl); in conv7x7s2_pack1ton_fp16sa_rvv() [all …]
|