229   int16x8_t temp2[32];  // Hold stages: 3, 6  in fdct32_vsx()  local
292   temp2[0] = vec_add(temp1[0], temp1[7]);  in fdct32_vsx()
293   temp2[1] = vec_add(temp1[1], temp1[6]);  in fdct32_vsx()
294   temp2[2] = vec_add(temp1[2], temp1[5]);  in fdct32_vsx()
295   temp2[3] = vec_add(temp1[3], temp1[4]);  in fdct32_vsx()
296   temp2[5] = vec_sub(temp1[2], temp1[5]);  in fdct32_vsx()
297   temp2[6] = vec_sub(temp1[1], temp1[6]);  in fdct32_vsx()
298   temp2[8] = temp1[8];  in fdct32_vsx()
299   temp2[9] = temp1[9];  in fdct32_vsx()
301   single_butterfly(temp1[13], temp1[10], &temp2[13], &temp2[10]);  in fdct32_vsx()
302   single_butterfly(temp1[12], temp1[11], &temp2[12], &temp2[11]);  in fdct32_vsx()
303   temp2[14] = temp1[14];  in fdct32_vsx()
304   temp2[15] = temp1[15];  in fdct32_vsx()
306   temp2[18] = vec_add(temp1[18], temp1[21]);  in fdct32_vsx()
307   temp2[19] = vec_add(temp1[19], temp1[20]);  in fdct32_vsx()
309   temp2[20] = vec_sub(temp1[19], temp1[20]);  in fdct32_vsx()
310   temp2[21] = vec_sub(temp1[18], temp1[21]);  in fdct32_vsx()
312   temp2[26] = vec_sub(temp1[29], temp1[26]);  in fdct32_vsx()
313   temp2[27] = vec_sub(temp1[28], temp1[27]);  in fdct32_vsx()
315   temp2[28] = vec_add(temp1[28], temp1[27]);  in fdct32_vsx()
316   temp2[29] = vec_add(temp1[29], temp1[26]);  in fdct32_vsx()
331   temp0[0] = vec_add(temp2[0], temp2[3]);  in fdct32_vsx()
332   temp0[1] = vec_add(temp2[1], temp2[2]);  in fdct32_vsx()
333   temp0[2] = vec_sub(temp2[1], temp2[2]);  in fdct32_vsx()
334   temp0[3] = vec_sub(temp2[0], temp2[3]);  in fdct32_vsx()
335   single_butterfly(temp2[6], temp2[5], &temp0[6], &temp0[5]);  in fdct32_vsx()
337   temp0[9] = vec_add(temp2[9], temp2[10]);  in fdct32_vsx()
338   temp0[10] = vec_sub(temp2[9], temp2[10]);  in fdct32_vsx()
339   temp0[13] = vec_sub(temp2[14], temp2[13]);  in fdct32_vsx()
340   temp0[14] = vec_add(temp2[14], temp2[13]);  in fdct32_vsx()
342   double_butterfly(temp2[29], cospi8_v, temp2[18], cospi24_v, &temp0[29],  in fdct32_vsx()
344   double_butterfly(temp2[28], cospi8_v, temp2[19], cospi24_v, &temp0[28],  in fdct32_vsx()
346   double_butterfly(temp2[27], cospi24_v, temp2[20], cospi8m_v, &temp0[27],  in fdct32_vsx()
348   double_butterfly(temp2[26], cospi24_v, temp2[21], cospi8m_v, &temp0[26],  in fdct32_vsx()
352   temp1[8] = vec_add(temp2[8], temp2[11]);  in fdct32_vsx()
353   temp1[11] = vec_sub(temp2[8], temp2[11]);  in fdct32_vsx()
354   temp1[12] = vec_sub(temp2[15], temp2[12]);  in fdct32_vsx()
355   temp1[15] = vec_add(temp2[15], temp2[12]);  in fdct32_vsx()
387   temp2[16] = vec_add(temp0[16], temp0[19]);  in fdct32_vsx()
388   temp2[19] = vec_sub(temp0[16], temp0[19]);  in fdct32_vsx()
389   temp2[20] = vec_sub(temp0[23], temp0[20]);  in fdct32_vsx()
390   temp2[23] = vec_add(temp0[23], temp0[20]);  in fdct32_vsx()
391   temp2[24] = vec_add(temp0[24], temp0[27]);  in fdct32_vsx()
392   temp2[27] = vec_sub(temp0[24], temp0[27]);  in fdct32_vsx()
393   temp2[28] = vec_sub(temp0[31], temp0[28]);  in fdct32_vsx()
394   temp2[31] = vec_add(temp0[31], temp0[28]);  in fdct32_vsx()
402   temp2[8] = vec_add(temp1[8], temp1[9]);  in fdct32_vsx()
403   temp2[9] = vec_sub(temp1[8], temp1[9]);  in fdct32_vsx()
404   temp2[10] = vec_sub(temp1[11], temp1[10]);  in fdct32_vsx()
405   temp2[11] = vec_add(temp1[11], temp1[10]);  in fdct32_vsx()
406   temp2[12] = vec_add(temp1[12], temp1[13]);  in fdct32_vsx()
407   temp2[13] = vec_sub(temp1[12], temp1[13]);  in fdct32_vsx()
408   temp2[14] = vec_sub(temp1[15], temp1[14]);  in fdct32_vsx()
409   temp2[15] = vec_add(temp1[15], temp1[14]);  in fdct32_vsx()
411   double_butterfly(temp1[30], cospi4_v, temp1[17], cospi28_v, &temp2[30],  in fdct32_vsx()
412                    &temp2[17]);  in fdct32_vsx()
413   double_butterfly(temp1[29], cospi28_v, temp1[18], cospi4m_v, &temp2[29],  in fdct32_vsx()
414                    &temp2[18]);  in fdct32_vsx()
415   double_butterfly(temp1[26], cospi20_v, temp1[21], cospi12_v, &temp2[26],  in fdct32_vsx()
416                    &temp2[21]);  in fdct32_vsx()
417   double_butterfly(temp1[25], cospi12_v, temp1[22], cospi20m_v, &temp2[25],  in fdct32_vsx()
418                    &temp2[22]);  in fdct32_vsx()
421   double_butterfly(temp2[15], cospi2_v, temp2[8], cospi30_v, &out[2], &out[30]);  in fdct32_vsx()
422   double_butterfly(temp2[14], cospi18_v, temp2[9], cospi14_v, &out[18],  in fdct32_vsx()
424   double_butterfly(temp2[13], cospi10_v, temp2[10], cospi22_v, &out[10],  in fdct32_vsx()
426   double_butterfly(temp2[12], cospi26_v, temp2[11], cospi6_v, &out[26],  in fdct32_vsx()
429   temp0[16] = vec_add(temp2[16], temp2[17]);  in fdct32_vsx()
430   temp0[17] = vec_sub(temp2[16], temp2[17]);  in fdct32_vsx()
431   temp0[18] = vec_sub(temp2[19], temp2[18]);  in fdct32_vsx()
432   temp0[19] = vec_add(temp2[19], temp2[18]);  in fdct32_vsx()
433   temp0[20] = vec_add(temp2[20], temp2[21]);  in fdct32_vsx()
434   temp0[21] = vec_sub(temp2[20], temp2[21]);  in fdct32_vsx()
435   temp0[22] = vec_sub(temp2[23], temp2[22]);  in fdct32_vsx()
436   temp0[23] = vec_add(temp2[23], temp2[22]);  in fdct32_vsx()
437   temp0[24] = vec_add(temp2[24], temp2[25]);  in fdct32_vsx()
438   temp0[25] = vec_sub(temp2[24], temp2[25]);  in fdct32_vsx()
439   temp0[26] = vec_sub(temp2[27], temp2[26]);  in fdct32_vsx()
440   temp0[27] = vec_add(temp2[27], temp2[26]);  in fdct32_vsx()
441   temp0[28] = vec_add(temp2[28], temp2[29]);  in fdct32_vsx()
442   temp0[29] = vec_sub(temp2[28], temp2[29]);  in fdct32_vsx()
443   temp0[30] = vec_sub(temp2[31], temp2[30]);  in fdct32_vsx()
444   temp0[31] = vec_add(temp2[31], temp2[30]);  in fdct32_vsx()
474   int16x8_t temp2[32];  in vpx_fdct32x32_rd_vsx()  local
485   fdct32_vsx(temp0, temp2, 0);  in vpx_fdct32x32_rd_vsx()
496   transpose_8x8(&temp2[0], &temp0[8]);  in vpx_fdct32x32_rd_vsx()
511   transpose_8x8(&temp2[8], &temp0[8]);  in vpx_fdct32x32_rd_vsx()
526   transpose_8x8(&temp2[16], &temp0[8]);  in vpx_fdct32x32_rd_vsx()
541   transpose_8x8(&temp2[24], &temp0[8]);  in vpx_fdct32x32_rd_vsx()