89           for (int cell = 0; cell < kCells; cell += 2) {  in Pack()  local
92                     src_lines_intertwined_4x[2 * (cell + 1) + outer][inner]),  in Pack()
94                     src_lines_intertwined_4x[2 * cell + outer][inner]));  in Pack()
98           for (int cell = 0; cell < kCells; cell += 2) {  in Pack()  local
101                     src_lines_intertwined_4x[2 * (cell + 1) + outer][inner]),  in Pack()
103                     src_lines_intertwined_4x[2 * cell + outer][inner]));  in Pack()
109           for (int cell = 0; cell < kCells - 1; cell += 2) {  in Pack()  local
112                     src_lines_intertwined_4x[2 * (cell + 1) + outer][inner]),  in Pack()
114                     src_lines_intertwined_4x[2 * cell + outer][inner]));  in Pack()
128           for (int cell = 1; cell < kCells; cell += 2) {  in Pack()  local
131                     src_lines_intertwined_4x[2 * (cell + 1) + outer][inner]),  in Pack()
133                     src_lines_intertwined_4x[2 * cell + outer][inner]));  in Pack()
146         for (int cell = 0; cell < kCells; cell++) {  in Pack()  local
148               zeroes, src_lines_intertwined_4x[2 * cell + outer][inner]));  in Pack()
150               zeroes, src_lines_intertwined_4x[2 * cell + outer][inner]));  in Pack()
151           sums_of_2_cells[cell][i] = __builtin_msa_addv_h(tmp0, tmp1);  in Pack()
157       for (int cell = 0; cell < kCells; cell++) {  in Pack()  local
159             reinterpret_cast<v8i16>(zeroes), sums_of_2_cells[cell][i]));  in Pack()
161             reinterpret_cast<v8i16>(zeroes), sums_of_2_cells[cell][i]));  in Pack()
162         sums_of_4_cells[cell][i] = __builtin_msa_addv_w(tmp0, tmp1);  in Pack()
166     for (int cell = 0; cell < kCells; cell++) {  in Pack()  local
167       v4i32 s01 = __builtin_msa_addv_w(sums_of_4_cells[cell][0],  in Pack()
168                                        sums_of_4_cells[cell][1]);  in Pack()
169       v4i32 s23 = __builtin_msa_addv_w(sums_of_4_cells[cell][2],  in Pack()
170                                        sums_of_4_cells[cell][3]);  in Pack()
173           dst->sums_of_each_slice() + start_width + 4 * cell;  in Pack()
252           for (int cell = 0; cell < kCells; cell += 2) {  in Pack()  local
255                     src_lines_intertwined_4x[2 * (cell + 1) + outer][inner]),  in Pack()
257                     src_lines_intertwined_4x[2 * cell + outer][inner]));  in Pack()
261           for (int cell = 0; cell < kCells; cell += 2) {  in Pack()  local
264                     src_lines_intertwined_4x[2 * (cell + 1) + outer][inner]),  in Pack()
266                     src_lines_intertwined_4x[2 * cell + outer][inner]));  in Pack()
272           for (int cell = 0; cell < kCells - 1; cell += 2) {  in Pack()  local
275                     src_lines_intertwined_4x[2 * (cell + 1) + outer][inner]),  in Pack()
277                     src_lines_intertwined_4x[2 * cell + outer][inner]));  in Pack()
291           for (int cell = 1; cell < kCells; cell += 2) {  in Pack()  local
294                     src_lines_intertwined_4x[2 * (cell + 1) + outer][inner]),  in Pack()
296                     src_lines_intertwined_4x[2 * cell + outer][inner]));  in Pack()
308         for (int cell = 0; cell < kCells; cell++) {  in Pack()  local
309           sums_of_2[cell][i] = reinterpret_cast<v8i16>(__builtin_msa_hadd_u_h(  in Pack()
311                   src_lines_intertwined_4x[2 * cell + outer][inner]),  in Pack()
313                   src_lines_intertwined_4x[2 * cell + outer][inner])));  in Pack()
319       for (int cell = 0; cell < kCells; cell++) {  in Pack()  local
320         sums_of_4[cell][i] = __builtin_msa_addv_h(sums_of_2[cell][2 * i],  in Pack()
321                                                   sums_of_2[cell][2 * i + 1]);  in Pack()
325     for (int cell = 0; cell < kCells; cell++) {  in Pack()  local
326       sums_of_8[cell] =  in Pack()
327           __builtin_msa_addv_h(sums_of_4[cell][0], sums_of_4[cell][1]);  in Pack()
332     for (int cell = 0; cell < kCells; cell++) {  in Pack()  local
333       sums_of_16[cell] = reinterpret_cast<v4i32>(  in Pack()
334           __builtin_msa_ilvr_h(zeroes, sums_of_8[cell]));  in Pack()
335       v8i16 tmp = __builtin_msa_ilvl_h(zeroes, sums_of_8[cell]);  in Pack()
336       sums_of_16[cell] =  in Pack()
337           __builtin_msa_addv_w(sums_of_16[cell], reinterpret_cast<v4i32>(tmp));  in Pack()
340     for (int cell = 0; cell < kCells; cell++) {  in Pack()  local
342           dst->sums_of_each_slice() + start_width + 4 * cell;  in Pack()
344       tmp = __builtin_msa_addv_w(tmp, sums_of_16[cell]);  in Pack()