/dports/math/reduce/Reduce-svn5758-src/packages/redlog/pasf/ |
H A D | pasf.rlg | 128 (b <= 0 and d >= 0) or bex _k23 [ - abs(a) <= _k23 <= abs(a)] ( 130 ((_k23 <= 0 and a < 0) or (_k23 >= 0 and a > 0)) and ( 132 (_k23*c - a*d + b*c <= 0 and a > 0) or (_k23*c - a*d + b*c >= 0 and a < 0)) 134 and a <> 0 and _k23 + b ~a~ 0) or bex _k23 [ - abs(c) <= _k23 <= abs(c)] ( 136 ((_k23 <= 0 and c > 0) or (_k23 >= 0 and c < 0)) and ( 138 (_k23*a + a*d - b*c <= 0 and c < 0) or (_k23*a + a*d - b*c >= 0 and c > 0)) 140 and c <> 0 and _k23 + d ~c~ 0)
|
/dports/graphics/vapoursynth-waifu2x-ncnn-vulkan/vapoursynth-waifu2x-ncnn-vulkan-r4/deps/ncnn/src/layer/arm/ |
H A D | convolutiondepthwise_5x5_pack4.h | 175 float32x4_t _k23 = vld1q_f32(k0 + 12); in convdw5x5s1_pack4_neon() local 203 _sum00 = vmlaq_f32(_sum00, _k23, _r23); in convdw5x5s1_pack4_neon() 208 _sum01 = vmlaq_f32(_sum01, _k23, _r24); in convdw5x5s1_pack4_neon() 213 _sum02 = vmlaq_f32(_sum02, _k23, _r25); in convdw5x5s1_pack4_neon() 218 _sum03 = vmlaq_f32(_sum03, _k23, _r26); in convdw5x5s1_pack4_neon() 240 _sum10 = vmlaq_f32(_sum10, _k23, _r33); in convdw5x5s1_pack4_neon() 245 _sum11 = vmlaq_f32(_sum11, _k23, _r34); in convdw5x5s1_pack4_neon() 250 _sum12 = vmlaq_f32(_sum12, _k23, _r35); in convdw5x5s1_pack4_neon() 664 _sum0 = vmlaq_f32(_sum0, _k23, _r23); in convdw5x5s1_pack4_neon() 683 _sum1 = vmlaq_f32(_sum1, _k23, _r33); in convdw5x5s1_pack4_neon() [all …]
|
H A D | convolutiondepthwise_5x5_pack8_fp16s.h | 1360 float16x8_t _k23 = vld1q_f16(k0 + 24); in convdw5x5s2_pack8_fp16sa_neon() local 1367 _sum0 = vfmaq_f16(_sum0, _k23, _r23); 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 | 175 float32x4_t _k23 = vld1q_f32(k0 + 12); in convdw5x5s1_pack4_neon() local 203 _sum00 = vmlaq_f32(_sum00, _k23, _r23); in convdw5x5s1_pack4_neon() 208 _sum01 = vmlaq_f32(_sum01, _k23, _r24); in convdw5x5s1_pack4_neon() 213 _sum02 = vmlaq_f32(_sum02, _k23, _r25); in convdw5x5s1_pack4_neon() 218 _sum03 = vmlaq_f32(_sum03, _k23, _r26); in convdw5x5s1_pack4_neon() 240 _sum10 = vmlaq_f32(_sum10, _k23, _r33); in convdw5x5s1_pack4_neon() 245 _sum11 = vmlaq_f32(_sum11, _k23, _r34); in convdw5x5s1_pack4_neon() 250 _sum12 = vmlaq_f32(_sum12, _k23, _r35); in convdw5x5s1_pack4_neon() 664 _sum0 = vmlaq_f32(_sum0, _k23, _r23); in convdw5x5s1_pack4_neon() 683 _sum1 = vmlaq_f32(_sum1, _k23, _r33); in convdw5x5s1_pack4_neon() [all …]
|
H A D | convolutiondepthwise_5x5_pack8_fp16s.h | 1360 float16x8_t _k23 = vld1q_f16(k0 + 24); in convdw5x5s2_pack8_fp16sa_neon() local 1367 _sum0 = vfmaq_f16(_sum0, _k23, _r23); in convdw5x5s2_pack8_fp16sa_neon()
|
/dports/benchmarks/vkpeak/vkpeak-20210430/ncnn/src/layer/arm/ |
H A D | convolutiondepthwise_5x5_pack4.h | 175 float32x4_t _k23 = vld1q_f32(k0 + 12); in convdw5x5s1_pack4_neon() local 203 _sum00 = vmlaq_f32(_sum00, _k23, _r23); in convdw5x5s1_pack4_neon() 208 _sum01 = vmlaq_f32(_sum01, _k23, _r24); in convdw5x5s1_pack4_neon() 213 _sum02 = vmlaq_f32(_sum02, _k23, _r25); in convdw5x5s1_pack4_neon() 218 _sum03 = vmlaq_f32(_sum03, _k23, _r26); in convdw5x5s1_pack4_neon() 240 _sum10 = vmlaq_f32(_sum10, _k23, _r33); in convdw5x5s1_pack4_neon() 245 _sum11 = vmlaq_f32(_sum11, _k23, _r34); in convdw5x5s1_pack4_neon() 250 _sum12 = vmlaq_f32(_sum12, _k23, _r35); in convdw5x5s1_pack4_neon() 664 _sum0 = vmlaq_f32(_sum0, _k23, _r23); in convdw5x5s1_pack4_neon() 683 _sum1 = vmlaq_f32(_sum1, _k23, _r33); in convdw5x5s1_pack4_neon() [all …]
|
H A D | convolutiondepthwise_5x5_pack8_fp16s.h | 1360 float16x8_t _k23 = vld1q_f16(k0 + 24); in convdw5x5s2_pack8_fp16sa_neon() local 1367 _sum0 = vfmaq_f16(_sum0, _k23, _r23); in convdw5x5s2_pack8_fp16sa_neon()
|
/dports/misc/ncnn/ncnn-20211208/src/layer/arm/ |
H A D | convolutiondepthwise_5x5_pack4.h | 175 float32x4_t _k23 = vld1q_f32(k0 + 12); in convdw5x5s1_pack4_neon() local 203 _sum00 = vmlaq_f32(_sum00, _k23, _r23); in convdw5x5s1_pack4_neon() 208 _sum01 = vmlaq_f32(_sum01, _k23, _r24); in convdw5x5s1_pack4_neon() 213 _sum02 = vmlaq_f32(_sum02, _k23, _r25); in convdw5x5s1_pack4_neon() 218 _sum03 = vmlaq_f32(_sum03, _k23, _r26); in convdw5x5s1_pack4_neon() 240 _sum10 = vmlaq_f32(_sum10, _k23, _r33); in convdw5x5s1_pack4_neon() 245 _sum11 = vmlaq_f32(_sum11, _k23, _r34); in convdw5x5s1_pack4_neon() 250 _sum12 = vmlaq_f32(_sum12, _k23, _r35); in convdw5x5s1_pack4_neon() 664 _sum0 = vmlaq_f32(_sum0, _k23, _r23); in convdw5x5s1_pack4_neon() 683 _sum1 = vmlaq_f32(_sum1, _k23, _r33); in convdw5x5s1_pack4_neon() [all …]
|
H A D | convolutiondepthwise_5x5_pack8_fp16s.h | 1360 float16x8_t _k23 = vld1q_f16(k0 + 24); in convdw5x5s2_pack8_fp16sa_neon() local 1367 _sum0 = vfmaq_f16(_sum0, _k23, _r23); 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 | 175 float32x4_t _k23 = vld1q_f32(k0 + 12); in convdw5x5s1_pack4_neon() local 203 _sum00 = vmlaq_f32(_sum00, _k23, _r23); in convdw5x5s1_pack4_neon() 208 _sum01 = vmlaq_f32(_sum01, _k23, _r24); in convdw5x5s1_pack4_neon() 213 _sum02 = vmlaq_f32(_sum02, _k23, _r25); in convdw5x5s1_pack4_neon() 218 _sum03 = vmlaq_f32(_sum03, _k23, _r26); in convdw5x5s1_pack4_neon() 240 _sum10 = vmlaq_f32(_sum10, _k23, _r33); in convdw5x5s1_pack4_neon() 245 _sum11 = vmlaq_f32(_sum11, _k23, _r34); in convdw5x5s1_pack4_neon() 250 _sum12 = vmlaq_f32(_sum12, _k23, _r35); in convdw5x5s1_pack4_neon() 664 _sum0 = vmlaq_f32(_sum0, _k23, _r23); in convdw5x5s1_pack4_neon() 683 _sum1 = vmlaq_f32(_sum1, _k23, _r33); in convdw5x5s1_pack4_neon() [all …]
|
H A D | convolutiondepthwise_5x5_pack8_fp16s.h | 1360 float16x8_t _k23 = vld1q_f16(k0 + 24); in convdw5x5s2_pack8_fp16sa_neon() local 1367 _sum0 = vfmaq_f16(_sum0, _k23, _r23); in convdw5x5s2_pack8_fp16sa_neon()
|
/dports/misc/ncnn/ncnn-20211208/src/layer/mips/ |
H A D | convolutiondepthwise_5x5_pack4.h | 124 v4f32 _k23 = (v4f32)__msa_ld_w(k0 + 4 * 3, 0); in convdw5x5s1_pack4_msa() local 131 _sum0 = __msa_fmadd_w(_sum0, _k23, _r23); in convdw5x5s1_pack4_msa() 143 _sum1 = __msa_fmadd_w(_sum1, _k23, _r33); in convdw5x5s1_pack4_msa() 282 v4f32 _k23 = (v4f32)__msa_ld_w(k0 + 4 * 3, 0); in convdw5x5s1_pack4_msa() local 289 _sum0 = __msa_fmadd_w(_sum0, _k23, _r23); in convdw5x5s1_pack4_msa() 445 v4f32 _k23 = (v4f32)__msa_ld_w(k0 + 4 * 3, 0); in convdw5x5s2_pack4_msa() local 452 _sum0 = __msa_fmadd_w(_sum0, _k23, _r23); in convdw5x5s2_pack4_msa()
|
H A D | convolution_7x7_pack1to4.h | 181 v4f32 _k23 = (v4f32)__msa_ld_w(kptr + 4 * 3, 0); in conv7x7s2_pack1to4_msa() local 218 _sum0 = __msa_fmadd_w(_sum0, _r23, _k23); in conv7x7s2_pack1to4_msa() 219 _sum1 = __msa_fmadd_w(_sum1, _r25, _k23); in conv7x7s2_pack1to4_msa() 220 _sum2 = __msa_fmadd_w(_sum2, _r27, _k23); in conv7x7s2_pack1to4_msa() 221 _sum3 = __msa_fmadd_w(_sum3, _r29, _k23); in conv7x7s2_pack1to4_msa() 527 v4f32 _k23 = (v4f32)__msa_ld_w(kptr + 4 * 3, 0); in conv7x7s2_pack1to4_msa() local 540 _sum0 = __msa_fmadd_w(_sum0, (v4f32)__msa_splati_w(_r2, 3), _k23); 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 | 98 __m256 _k23 = _mm256_loadu_ps(k0 + 24); in convdw5x5s1_pack8_avx() local 105 _sum0 = _mm256_fmadd_ps(_k23, _r23, _sum0); in convdw5x5s1_pack8_avx() 252 __m256 _k23 = _mm256_loadu_ps(k0 + 24); in convdw5x5s2_pack8_avx() local 259 _sum0 = _mm256_fmadd_ps(_k23, _r23, _sum0); in convdw5x5s2_pack8_avx()
|
H A D | convolution_3x3_pack8.h | 144 __m256 _k23 = _mm256_loadu_ps(kptr + 24); in conv3x3s1_pack8_avx() local 155 _sum01 = _mm256_fmadd_ps(_r023, _k23, _sum01); in conv3x3s1_pack8_avx() 173 _sum11 = _mm256_fmadd_ps(_r033, _k23, _sum11); in conv3x3s1_pack8_avx() 511 __m256 _k23 = _mm256_loadu_ps(kptr + 24); in conv3x3s1_pack8_avx() local 522 _sum1 = _mm256_fmadd_ps(_r023, _k23, _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 | 98 __m256 _k23 = _mm256_loadu_ps(k0 + 24); in convdw5x5s1_pack8_avx() local 105 _sum0 = _mm256_fmadd_ps(_k23, _r23, _sum0); in convdw5x5s1_pack8_avx() 252 __m256 _k23 = _mm256_loadu_ps(k0 + 24); in convdw5x5s2_pack8_avx() local 259 _sum0 = _mm256_fmadd_ps(_k23, _r23, _sum0); in convdw5x5s2_pack8_avx()
|
H A D | convolution_3x3_pack8.h | 144 __m256 _k23 = _mm256_loadu_ps(kptr + 24); in conv3x3s1_pack8_avx() local 155 _sum01 = _mm256_fmadd_ps(_r023, _k23, _sum01); in conv3x3s1_pack8_avx() 173 _sum11 = _mm256_fmadd_ps(_r033, _k23, _sum11); in conv3x3s1_pack8_avx() 511 __m256 _k23 = _mm256_loadu_ps(kptr + 24); in conv3x3s1_pack8_avx() local 522 _sum1 = _mm256_fmadd_ps(_r023, _k23, _sum1); in conv3x3s1_pack8_avx()
|
/dports/misc/ncnn/ncnn-20211208/src/layer/x86/ |
H A D | convolutiondepthwise_5x5_pack8.h | 98 __m256 _k23 = _mm256_loadu_ps(k0 + 24); in convdw5x5s1_pack8_avx() local 105 _sum0 = _mm256_comp_fmadd_ps(_k23, _r23, _sum0); in convdw5x5s1_pack8_avx() 252 __m256 _k23 = _mm256_loadu_ps(k0 + 24); in convdw5x5s2_pack8_avx() local 259 _sum0 = _mm256_comp_fmadd_ps(_k23, _r23, _sum0); in convdw5x5s2_pack8_avx()
|
/dports/benchmarks/vkpeak/vkpeak-20210430/ncnn/src/layer/x86/ |
H A D | convolutiondepthwise_5x5_pack8.h | 98 __m256 _k23 = _mm256_loadu_ps(k0 + 24); in convdw5x5s1_pack8_avx() local 105 _sum0 = _mm256_fmadd_ps(_k23, _r23, _sum0); in convdw5x5s1_pack8_avx() 252 __m256 _k23 = _mm256_loadu_ps(k0 + 24); in convdw5x5s2_pack8_avx() local 259 _sum0 = _mm256_fmadd_ps(_k23, _r23, _sum0); in convdw5x5s2_pack8_avx()
|
H A D | convolution_3x3_pack8.h | 144 __m256 _k23 = _mm256_loadu_ps(kptr + 24); in conv3x3s1_pack8_avx() local 155 _sum01 = _mm256_fmadd_ps(_r023, _k23, _sum01); in conv3x3s1_pack8_avx() 173 _sum11 = _mm256_fmadd_ps(_r033, _k23, _sum11); in conv3x3s1_pack8_avx() 511 __m256 _k23 = _mm256_loadu_ps(kptr + 24); in conv3x3s1_pack8_avx() local 522 _sum1 = _mm256_fmadd_ps(_r023, _k23, _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 | 98 __m256 _k23 = _mm256_loadu_ps(k0 + 24); in convdw5x5s1_pack8_avx() local 105 _sum0 = _mm256_fmadd_ps(_k23, _r23, _sum0); in convdw5x5s1_pack8_avx() 252 __m256 _k23 = _mm256_loadu_ps(k0 + 24); in convdw5x5s2_pack8_avx() local 259 _sum0 = _mm256_fmadd_ps(_k23, _r23, _sum0); in convdw5x5s2_pack8_avx()
|
/dports/misc/ncnn/ncnn-20211208/src/layer/riscv/ |
H A D | convolutiondepthwise_5x5_packn.h | 118 vfloat32m1_t _k23 = vle32_v_f32m1(k0 + packn * 3, vl); in convdw5x5s1_packn_rvv() local 125 _sum0 = vfmacc_vv_f32m1(_sum0, _k23, _r23, vl); in convdw5x5s1_packn_rvv() 137 _sum1 = vfmacc_vv_f32m1(_sum1, _k23, _r33, vl); in convdw5x5s1_packn_rvv() 268 vfloat32m1_t _k23 = vle32_v_f32m1(k0 + packn * 3, vl); in convdw5x5s1_packn_rvv() local 275 _sum0 = vfmacc_vv_f32m1(_sum0, _k23, _r23, vl); in convdw5x5s1_packn_rvv() 426 vfloat32m1_t _k23 = vle32_v_f32m1(k0 + packn * 3, vl); in convdw5x5s2_packn_rvv() local 433 _sum0 = vfmacc_vv_f32m1(_sum0, _k23, _r23, vl); in convdw5x5s2_packn_rvv()
|
H A D | convolutiondepthwise_5x5_packn_fp16s.h | 118 vfloat16m1_t _k23 = vle16_v_f16m1(k0 + packn * 3, vl); in convdw5x5s1_packn_fp16sa_rvv() local 125 _sum0 = vfmacc_vv_f16m1(_sum0, _k23, _r23, vl); in convdw5x5s1_packn_fp16sa_rvv() 137 _sum1 = vfmacc_vv_f16m1(_sum1, _k23, _r33, vl); in convdw5x5s1_packn_fp16sa_rvv() 268 vfloat16m1_t _k23 = vle16_v_f16m1(k0 + packn * 3, vl); in convdw5x5s1_packn_fp16sa_rvv() local 275 _sum0 = vfmacc_vv_f16m1(_sum0, _k23, _r23, vl); in convdw5x5s1_packn_fp16sa_rvv() 426 vfloat16m1_t _k23 = vle16_v_f16m1(k0 + packn * 3, vl); in convdw5x5s2_packn_fp16sa_rvv() local 433 _sum0 = vfmacc_vv_f16m1(_sum0, _k23, _r23, vl); in convdw5x5s2_packn_fp16sa_rvv()
|
H A D | convolution_7x7_pack1ton.h | 239 _sum0 = vfmacc_vf_f32m1(_sum0, r2[3], _k23, vl); in conv7x7s2_pack1ton_rvv() 240 _sum1 = vfmacc_vf_f32m1(_sum1, r2[5], _k23, vl); in conv7x7s2_pack1ton_rvv() 241 _sum2 = vfmacc_vf_f32m1(_sum2, r2[7], _k23, vl); in conv7x7s2_pack1ton_rvv() 242 _sum3 = vfmacc_vf_f32m1(_sum3, r2[9], _k23, vl); in conv7x7s2_pack1ton_rvv() 243 _sum4 = vfmacc_vf_f32m1(_sum4, r2[11], _k23, vl); in conv7x7s2_pack1ton_rvv() 244 _sum5 = vfmacc_vf_f32m1(_sum5, r2[13], _k23, vl); in conv7x7s2_pack1ton_rvv() 245 _sum6 = vfmacc_vf_f32m1(_sum6, r2[15], _k23, vl); in conv7x7s2_pack1ton_rvv() 246 _sum7 = vfmacc_vf_f32m1(_sum7, r2[17], _k23, vl); in conv7x7s2_pack1ton_rvv() 666 _sum0 = vfmacc_vf_f32m1(_sum0, r2[3], _k23, vl); in conv7x7s2_pack1ton_rvv() 667 _sum1 = vfmacc_vf_f32m1(_sum1, r2[5], _k23, vl); in conv7x7s2_pack1ton_rvv() [all …]
|
H A D | convolution_7x7_pack1ton_fp16s.h | 239 _sum0 = vfmacc_vf_f16m1(_sum0, r2[3], _k23, vl); in conv7x7s2_pack1ton_fp16sa_rvv() 240 _sum1 = vfmacc_vf_f16m1(_sum1, r2[5], _k23, vl); in conv7x7s2_pack1ton_fp16sa_rvv() 241 _sum2 = vfmacc_vf_f16m1(_sum2, r2[7], _k23, vl); in conv7x7s2_pack1ton_fp16sa_rvv() 242 _sum3 = vfmacc_vf_f16m1(_sum3, r2[9], _k23, vl); in conv7x7s2_pack1ton_fp16sa_rvv() 243 _sum4 = vfmacc_vf_f16m1(_sum4, r2[11], _k23, vl); in conv7x7s2_pack1ton_fp16sa_rvv() 244 _sum5 = vfmacc_vf_f16m1(_sum5, r2[13], _k23, vl); in conv7x7s2_pack1ton_fp16sa_rvv() 245 _sum6 = vfmacc_vf_f16m1(_sum6, r2[15], _k23, vl); in conv7x7s2_pack1ton_fp16sa_rvv() 246 _sum7 = vfmacc_vf_f16m1(_sum7, r2[17], _k23, vl); in conv7x7s2_pack1ton_fp16sa_rvv() 666 _sum0 = vfmacc_vf_f16m1(_sum0, r2[3], _k23, vl); in conv7x7s2_pack1ton_fp16sa_rvv() 667 _sum1 = vfmacc_vf_f16m1(_sum1, r2[5], _k23, vl); in conv7x7s2_pack1ton_fp16sa_rvv() [all …]
|