| /dports/misc/ncnn/ncnn-20211208/src/layer/mips/ |
| H A D | convolutiondepthwise_5x5_pack4.h | 65 v4f32 _r00 = (v4f32)__msa_ld_w(r0, 0); in convdw5x5s1_pack4_msa()
71 v4f32 _k00 = (v4f32)__msa_ld_w(k0, 0); in convdw5x5s1_pack4_msa()
84 v4f32 _r10 = (v4f32)__msa_ld_w(r1, 0); in convdw5x5s1_pack4_msa()
96 v4f32 _k10 = (v4f32)__msa_ld_w(k0, 0); in convdw5x5s1_pack4_msa()
109 v4f32 _r20 = (v4f32)__msa_ld_w(r2, 0); in convdw5x5s1_pack4_msa()
121 v4f32 _k20 = (v4f32)__msa_ld_w(k0, 0); in convdw5x5s1_pack4_msa()
134 v4f32 _r30 = (v4f32)__msa_ld_w(r3, 0); in convdw5x5s1_pack4_msa()
146 v4f32 _k30 = (v4f32)__msa_ld_w(k0, 0); in convdw5x5s1_pack4_msa()
159 v4f32 _r40 = (v4f32)__msa_ld_w(r4, 0); in convdw5x5s1_pack4_msa()
171 v4f32 _k40 = (v4f32)__msa_ld_w(k0, 0); in convdw5x5s1_pack4_msa()
[all …]
|
| H A D | convolutiondepthwise_3x3_pack4.h | 45 v4f32 _k00 = (v4f32)__msa_ld_w(k0, 0); in convdw3x3s1_pack4_msa()
46 v4f32 _k01 = (v4f32)__msa_ld_w(k0 + 4, 0); in convdw3x3s1_pack4_msa()
47 v4f32 _k02 = (v4f32)__msa_ld_w(k0 + 4 * 2, 0); in convdw3x3s1_pack4_msa()
48 v4f32 _k10 = (v4f32)__msa_ld_w(k0 + 4 * 3, 0); in convdw3x3s1_pack4_msa()
49 v4f32 _k11 = (v4f32)__msa_ld_w(k0 + 4 * 4, 0); in convdw3x3s1_pack4_msa()
50 v4f32 _k12 = (v4f32)__msa_ld_w(k0 + 4 * 5, 0); in convdw3x3s1_pack4_msa()
51 v4f32 _k20 = (v4f32)__msa_ld_w(k0 + 4 * 6, 0); in convdw3x3s1_pack4_msa()
52 v4f32 _k21 = (v4f32)__msa_ld_w(k0 + 4 * 7, 0); in convdw3x3s1_pack4_msa()
53 v4f32 _k22 = (v4f32)__msa_ld_w(k0 + 4 * 8, 0); in convdw3x3s1_pack4_msa()
347 v4f32 _k00 = (v4f32)__msa_ld_w(k0, 0); in convdw3x3s2_pack4_msa()
[all …]
|
| H A D | convolution_7x7_pack1to4.h | 74 v4i32 _r0 = __msa_ld_w(r0, 0); in conv7x7s2_pack1to4_msa()
131 v4i32 _r1 = __msa_ld_w(r1, 0); in conv7x7s2_pack1to4_msa()
188 v4i32 _r2 = __msa_ld_w(r2, 0); in conv7x7s2_pack1to4_msa()
245 v4i32 _r3 = __msa_ld_w(r3, 0); in conv7x7s2_pack1to4_msa()
302 v4i32 _r4 = __msa_ld_w(r4, 0); in conv7x7s2_pack1to4_msa()
359 v4i32 _r5 = __msa_ld_w(r5, 0); in conv7x7s2_pack1to4_msa()
416 v4i32 _r6 = __msa_ld_w(r6, 0); in conv7x7s2_pack1to4_msa()
492 v4i32 _r0 = __msa_ld_w(r0, 0); in conv7x7s2_pack1to4_msa()
513 v4i32 _r1 = __msa_ld_w(r1, 0); in conv7x7s2_pack1to4_msa()
534 v4i32 _r2 = __msa_ld_w(r2, 0); in conv7x7s2_pack1to4_msa()
[all …]
|
| H A D | interp_bicubic_pack4.h | 62 v4f32 _S30 = (v4f32)__msa_ld_w(S3p - 4, 0); in resize_bicubic_image_pack4()
63 v4f32 _S31 = (v4f32)__msa_ld_w(S3p + 0, 0); in resize_bicubic_image_pack4()
64 v4f32 _S32 = (v4f32)__msa_ld_w(S3p + 4, 0); in resize_bicubic_image_pack4()
65 v4f32 _S33 = (v4f32)__msa_ld_w(S3p + 8, 0); in resize_bicubic_image_pack4()
101 v4f32 _S20 = (v4f32)__msa_ld_w(S2p - 4, 0); in resize_bicubic_image_pack4()
102 v4f32 _S21 = (v4f32)__msa_ld_w(S2p + 0, 0); in resize_bicubic_image_pack4()
103 v4f32 _S22 = (v4f32)__msa_ld_w(S2p + 4, 0); in resize_bicubic_image_pack4()
267 v4f32 _rows0 = (v4f32)__msa_ld_w(rows0p, 0); in resize_bicubic_image_pack4()
268 v4f32 _rows1 = (v4f32)__msa_ld_w(rows1p, 0); in resize_bicubic_image_pack4()
269 v4f32 _rows2 = (v4f32)__msa_ld_w(rows2p, 0); in resize_bicubic_image_pack4()
[all …]
|
| H A D | convolution_3x3_pack1to4.h | 45 v4f32 _k00 = (v4f32)__msa_ld_w(k0, 0); in conv3x3s1_pack1to4_msa()
46 v4f32 _k01 = (v4f32)__msa_ld_w(k0 + 4, 0); in conv3x3s1_pack1to4_msa()
70 v4i32 _r0 = __msa_ld_w(r0, 0); in conv3x3s1_pack1to4_msa()
110 v4i32 _r1 = __msa_ld_w(r1, 0); in conv3x3s1_pack1to4_msa()
150 v4i32 _r2 = __msa_ld_w(r2, 0); in conv3x3s1_pack1to4_msa()
212 v4i32 _r0 = __msa_ld_w(r0, 0); in conv3x3s1_pack1to4_msa()
235 v4i32 _r1 = __msa_ld_w(r1, 0); in conv3x3s1_pack1to4_msa()
258 v4i32 _r2 = __msa_ld_w(r2, 0); in conv3x3s1_pack1to4_msa()
297 v4i32 _r0 = __msa_ld_w(r0, 0); in conv3x3s1_pack1to4_msa()
310 v4i32 _r1 = __msa_ld_w(r1, 0); in conv3x3s1_pack1to4_msa()
[all …]
|
| H A D | convolution_3x3_pack4.h | 333 v4f32 _r0 = (v4f32)__msa_ld_w(r0, 0); in conv3x3s1_winograd64_pack4_msa()
399 v4f32 _r0 = (v4f32)__msa_ld_w(r0, 0); in conv3x3s1_winograd64_pack4_msa()
449 v4f32 _r0 = (v4f32)__msa_ld_w(r0, 0); in conv3x3s1_winograd64_pack4_msa()
483 v4f32 _r0 = (v4f32)__msa_ld_w(r0, 0); in conv3x3s1_winograd64_pack4_msa()
506 v4f32 _val = (v4f32)__msa_ld_w(r0, 0); in conv3x3s1_winograd64_pack4_msa()
556 v4i32 _val0123 = __msa_ld_w(r0, 0); in conv3x3s1_winograd64_pack4_msa()
612 v4i32 _val0123 = __msa_ld_w(r0, 0); in conv3x3s1_winograd64_pack4_msa()
1147 v4f32 _r0 = (v4f32)__msa_ld_w(r0, 0); in conv3x3s1_winograd42_pack4_msa()
1213 v4f32 _r0 = (v4f32)__msa_ld_w(r0, 0); in conv3x3s1_winograd42_pack4_msa()
1263 v4f32 _r0 = (v4f32)__msa_ld_w(r0, 0); in conv3x3s1_winograd42_pack4_msa()
[all …]
|
| H A D | eltwise_mips.cpp | 62 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward()
63 v4f32 _p1 = (v4f32)__msa_ld_w(ptr1, 0); in forward()
207 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward()
208 v4f32 _p1 = (v4f32)__msa_ld_w(ptr1, 0); in forward()
260 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward()
261 v4f32 _p1 = (v4f32)__msa_ld_w(ptr1, 0); in forward()
294 v4f32 _p1 = (v4f32)__msa_ld_w(ptr, 0); in forward()
329 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward()
402 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward()
475 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward()
[all …]
|
| H A D | padding_pack4.h | 39 __msa_st_w(__msa_ld_w(ptr, 0), outptr, 0); in padding_constant_pack4_msa()
66 v4f32 _p = (v4f32)__msa_ld_w(ptr0, 0); in padding_replicate_pack4_msa()
74 _p = (v4f32)__msa_ld_w(ptr0, 0); in padding_replicate_pack4_msa()
88 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in padding_replicate_pack4_msa()
96 _p = (v4f32)__msa_ld_w(ptr, 0); in padding_replicate_pack4_msa()
112 v4f32 _p = (v4f32)__msa_ld_w(ptr0, 0); in padding_replicate_pack4_msa()
120 _p = (v4f32)__msa_ld_w(ptr0, 0); in padding_replicate_pack4_msa()
151 v4f32 _p = (v4f32)__msa_ld_w(ptr0, 0); in padding_reflect_pack4_msa()
175 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in padding_reflect_pack4_msa()
182 v4f32 _p = (v4f32)__msa_ld_w(ptr - 8 - x * 4, 0); in padding_reflect_pack4_msa()
[all …]
|
| H A D | convolution_sgemm_pack4.h | 57 v4f32 _r0 = (v4f32)__msa_ld_w(img0, 0); in im2col_sgemm_pack4_msa()
131 v4f32 _r0 = (v4f32)__msa_ld_w(img0, 0); in im2col_sgemm_pack4_msa()
189 v4f32 _r0 = (v4f32)__msa_ld_w(img0, 0); in im2col_sgemm_pack4_msa()
299 v4i32 _val0123 = __msa_ld_w(tmpptr, 0); in im2col_sgemm_pack4_msa()
302 v4f32 _w0 = (v4f32)__msa_ld_w(kptr0, 0); in im2col_sgemm_pack4_msa()
355 v4i32 _val0123 = __msa_ld_w(tmpptr, 0); in im2col_sgemm_pack4_msa()
357 v4f32 _w0 = (v4f32)__msa_ld_w(kptr0, 0); in im2col_sgemm_pack4_msa()
398 v4i32 _val0123 = __msa_ld_w(tmpptr, 0); in im2col_sgemm_pack4_msa()
399 v4f32 _w0 = (v4f32)__msa_ld_w(kptr0, 0); in im2col_sgemm_pack4_msa()
432 v4f32 _w0 = (v4f32)__msa_ld_w(kptr0, 0); in im2col_sgemm_pack4_msa()
[all …]
|
| H A D | convolution_sgemm_pack4to1.h | 54 v4f32 _r0 = (v4f32)__msa_ld_w(img0, 0); in im2col_sgemm_pack4to1_msa()
128 v4f32 _r0 = (v4f32)__msa_ld_w(img0, 0); in im2col_sgemm_pack4to1_msa()
258 v4i32 _bias = __msa_ld_w(biasptr, 0); in im2col_sgemm_pack4to1_msa()
279 v4i32 _w0123 = __msa_ld_w(kptr0, 0); in im2col_sgemm_pack4to1_msa()
322 v4i32 _bias = __msa_ld_w(biasptr, 0); in im2col_sgemm_pack4to1_msa()
338 v4i32 _w0123 = __msa_ld_w(kptr0, 0); in im2col_sgemm_pack4to1_msa()
373 v4i32 _bias = __msa_ld_w(biasptr, 0); in im2col_sgemm_pack4to1_msa()
384 v4i32 _w0123 = __msa_ld_w(kptr0, 0); in im2col_sgemm_pack4to1_msa()
411 v4f32 _sum = (v4f32)__msa_ld_w(biasptr, 0); in im2col_sgemm_pack4to1_msa()
418 v4f32 _w0 = (v4f32)__msa_ld_w(kptr0, 0); in im2col_sgemm_pack4to1_msa()
[all …]
|
| H A D | binaryop_mips.cpp | 73 v4f32 _b0 = (v4f32)__msa_ld_w(b0, 0); in binary_op_pack4()
77 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in binary_op_pack4()
130 v4f32 _a0 = (v4f32)__msa_ld_w(a0, 0); in binary_op_pack4()
323 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in binary_op_pack4()
468 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in binary_op_pack4()
469 v4f32 _p1 = (v4f32)__msa_ld_w(ptr1, 0); in binary_op_pack4()
512 v4f32 _b0 = (v4f32)__msa_ld_w(ptr1, 0); in binary_op_pack4()
649 v4f32 _a0 = (v4f32)__msa_ld_w(ptr, 0); in binary_op_pack4()
699 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in binary_op_pack4()
700 v4f32 _p1 = (v4f32)__msa_ld_w(ptr1, 0); in binary_op_pack4()
[all …]
|
| H A D | interp_bilinear_pack4.h | 55 v4f32 _S10 = (v4f32)__msa_ld_w(S1p, 0); in resize_bilinear_image_pack4()
56 v4f32 _S11 = (v4f32)__msa_ld_w(S1p + 4, 0); in resize_bilinear_image_pack4()
83 v4f32 _S00 = (v4f32)__msa_ld_w(S0p, 0); in resize_bilinear_image_pack4()
84 v4f32 _S01 = (v4f32)__msa_ld_w(S0p + 4, 0); in resize_bilinear_image_pack4()
85 v4f32 _S10 = (v4f32)__msa_ld_w(S1p, 0); in resize_bilinear_image_pack4()
86 v4f32 _S11 = (v4f32)__msa_ld_w(S1p + 4, 0); in resize_bilinear_image_pack4()
110 v4f32 _rows0 = (v4f32)__msa_ld_w(rows0p, 0); in resize_bilinear_image_pack4()
111 v4f32 _rows1 = (v4f32)__msa_ld_w(rows1p, 0); in resize_bilinear_image_pack4()
|
| H A D | unaryop_mips.cpp | 52 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in unary_op_inplace_pack4()
89 return (v4f32)__msa_ld_w(tmp, 0); in operator ()()
109 return (v4f32)__msa_ld_w(tmp, 0); in operator ()()
169 return (v4f32)__msa_ld_w(tmp, 0); in operator ()()
184 return (v4f32)__msa_ld_w(tmp, 0); in operator ()()
199 return (v4f32)__msa_ld_w(tmp, 0); in operator ()()
214 return (v4f32)__msa_ld_w(tmp, 0); in operator ()()
229 return (v4f32)__msa_ld_w(tmp, 0); in operator ()()
244 return (v4f32)__msa_ld_w(tmp, 0); in operator ()()
|
| H A D | prelu_mips.cpp | 55 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward_inplace()
56 v4f32 _slope = (v4f32)__msa_ld_w(slope + i * 4, 0); in forward_inplace()
72 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward_inplace()
90 …v4f32 _slope = num_slope > 1 ? (v4f32)__msa_ld_w((const float*)slope_data + i * 4, 0) : (v4f32)__m… in forward_inplace()
95 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward_inplace()
117 …v4f32 _slope = num_slope > 1 ? (v4f32)__msa_ld_w((const float*)slope_data + q * 4, 0) : (v4f32)__m… in forward_inplace()
122 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward_inplace()
189 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward_inplace()
232 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward_inplace()
|
| H A D | softmax_mips.cpp | 77 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward_inplace()
78 v4f32 _max = (v4f32)__msa_ld_w(maxptr, 0); in forward_inplace()
118 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward_inplace()
119 v4f32 _sum = (v4f32)__msa_ld_w(sumptr, 0); in forward_inplace()
153 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward_inplace()
154 v4f32 _sum = (v4f32)__msa_ld_w(sumptr, 0); in forward_inplace()
|
| H A D | convolution1d_mips.cpp | 136 _sum = (v4f32)__msa_ld_w((const float*)bias_data + p * 4, 0); in forward()
152 v4f32 _w0 = (v4f32)__msa_ld_w(kptr, 0); in forward()
153 v4f32 _w1 = (v4f32)__msa_ld_w(kptr + 4, 0); in forward()
154 v4f32 _w2 = (v4f32)__msa_ld_w(kptr + 8, 0); in forward()
155 v4f32 _w3 = (v4f32)__msa_ld_w(kptr + 12, 0); in forward()
190 _sum = (v4f32)__msa_ld_w((const float*)bias_data + p * 4, 0); in forward()
202 v4f32 _w = (v4f32)__msa_ld_w(kptr, 0); in forward()
246 v4f32 _val = (v4f32)__msa_ld_w(sptr, 0); in forward()
247 v4f32 _w = (v4f32)__msa_ld_w(kptr, 0); in forward()
|
| H A D | packing_mips.cpp | 125 v4f32 _r0 = (v4f32)__msa_ld_w(r0, 0); in forward()
126 v4f32 _r1 = (v4f32)__msa_ld_w(r1, 0); in forward()
127 v4f32 _r2 = (v4f32)__msa_ld_w(r2, 0); in forward()
128 v4f32 _r3 = (v4f32)__msa_ld_w(r3, 0); in forward()
179 v4f32 _r0 = (v4f32)__msa_ld_w(r0, 0); in forward()
180 v4f32 _r1 = (v4f32)__msa_ld_w(r0 + 4, 0); in forward()
250 v4f32 _r0 = (v4f32)__msa_ld_w(r0, 0); in forward()
251 v4f32 _r1 = (v4f32)__msa_ld_w(r1, 0); in forward()
252 v4f32 _r2 = (v4f32)__msa_ld_w(r2, 0); in forward()
253 v4f32 _r3 = (v4f32)__msa_ld_w(r3, 0); in forward()
[all …]
|
| H A D | convolution_pack4.h | 60 _sum = (v4f32)__msa_ld_w(bias_data_ptr + p * 4, 0); in convolution_pack4_msa()
80 v4f32 _w0 = (v4f32)__msa_ld_w(kptr, 0); in convolution_pack4_msa()
81 v4f32 _w1 = (v4f32)__msa_ld_w(kptr + 4, 0); in convolution_pack4_msa()
82 v4f32 _w2 = (v4f32)__msa_ld_w(kptr + 8, 0); in convolution_pack4_msa()
83 v4f32 _w3 = (v4f32)__msa_ld_w(kptr + 12, 0); in convolution_pack4_msa()
|
| H A D | deconvolution_pack4.h | 46 _sum = (v4f32)__msa_ld_w((const float*)bias_data_ptr + p * 4, 0); in deconvolution_pack4_msa()
84 v4f32 _w0 = (v4f32)__msa_ld_w(kptr + k, 0); in deconvolution_pack4_msa()
85 v4f32 _w1 = (v4f32)__msa_ld_w(kptr + k + 4, 0); in deconvolution_pack4_msa()
86 v4f32 _w2 = (v4f32)__msa_ld_w(kptr + k + 8, 0); in deconvolution_pack4_msa()
87 v4f32 _w3 = (v4f32)__msa_ld_w(kptr + k + 12, 0); in deconvolution_pack4_msa()
|
| H A D | pooling_mips.cpp | 86 v4f32 _max = (v4f32)__msa_ld_w(ptr, 0); in forward()
89 v4f32 _val = (v4f32)__msa_ld_w(ptr, 0); in forward()
108 v4f32 _val = (v4f32)__msa_ld_w(ptr, 0); in forward()
173 v4f32 _max = (v4f32)__msa_ld_w(sptr, 0); in forward()
177 v4f32 _val = (v4f32)__msa_ld_w(sptr + space_ofs[k] * 4, 0); in forward()
238 v4f32 _val = (v4f32)__msa_ld_w(m.row(sy) + sx * 4, 0); in forward()
272 v4f32 _val = (v4f32)__msa_ld_w(sptr + space_ofs[k] * 4, 0); in forward()
|
| H A D | flatten_mips.cpp | 104 v4f32 _r0 = (v4f32)__msa_ld_w(ptr, 0); in forward()
105 v4f32 _r1 = (v4f32)__msa_ld_w(ptr + 4, 0); in forward()
106 v4f32 _r2 = (v4f32)__msa_ld_w(ptr + 4 * 2, 0); in forward()
107 v4f32 _r3 = (v4f32)__msa_ld_w(ptr + 4 * 3, 0); in forward()
161 v4f32 _r0 = (v4f32)__msa_ld_w(ptr, 0); in forward()
162 v4f32 _r1 = (v4f32)__msa_ld_w(ptr + 4, 0); in forward()
163 v4f32 _r2 = (v4f32)__msa_ld_w(ptr + 4 * 2, 0); in forward()
164 v4f32 _r3 = (v4f32)__msa_ld_w(ptr + 4 * 3, 0); in forward()
211 __msa_st_w(__msa_ld_w(ptr, 0), outptr, 0); in forward()
|
| /dports/graphics/waifu2x-ncnn-vulkan/waifu2x-ncnn-vulkan-20210521/src/ncnn/src/layer/mips/ |
| H A D | softmax_mips.cpp | 78 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward_inplace()
79 v4f32 _max = (v4f32)__msa_ld_w(maxptr, 0); in forward_inplace()
119 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward_inplace()
120 v4f32 _sum = (v4f32)__msa_ld_w(sumptr, 0); in forward_inplace()
154 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward_inplace()
155 v4f32 _sum = (v4f32)__msa_ld_w(sumptr, 0); in forward_inplace()
|
| /dports/graphics/vapoursynth-waifu2x-ncnn-vulkan/vapoursynth-waifu2x-ncnn-vulkan-r4/deps/ncnn/src/layer/mips/ |
| H A D | softmax_mips.cpp | 78 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward_inplace()
79 v4f32 _max = (v4f32)__msa_ld_w(maxptr, 0); in forward_inplace()
119 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward_inplace()
120 v4f32 _sum = (v4f32)__msa_ld_w(sumptr, 0); in forward_inplace()
154 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward_inplace()
155 v4f32 _sum = (v4f32)__msa_ld_w(sumptr, 0); in forward_inplace()
|
| /dports/benchmarks/vkpeak/vkpeak-20210430/ncnn/src/layer/mips/ |
| H A D | softmax_mips.cpp | 78 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward_inplace()
79 v4f32 _max = (v4f32)__msa_ld_w(maxptr, 0); in forward_inplace()
119 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward_inplace()
120 v4f32 _sum = (v4f32)__msa_ld_w(sumptr, 0); in forward_inplace()
154 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward_inplace()
155 v4f32 _sum = (v4f32)__msa_ld_w(sumptr, 0); in forward_inplace()
|
| /dports/graphics/realsr-ncnn-vulkan/realsr-ncnn-vulkan-20210210/src/ncnn/src/layer/mips/ |
| H A D | softmax_mips.cpp | 78 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward_inplace()
79 v4f32 _max = (v4f32)__msa_ld_w(maxptr, 0); in forward_inplace()
119 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward_inplace()
120 v4f32 _sum = (v4f32)__msa_ld_w(sumptr, 0); in forward_inplace()
154 v4f32 _p = (v4f32)__msa_ld_w(ptr, 0); in forward_inplace()
155 v4f32 _sum = (v4f32)__msa_ld_w(sumptr, 0); in forward_inplace()
|