1090       stbi_uc *row0 = bytes + row*bytes_per_row;  in stbi__vertical_flip()  local
1096          memcpy(temp, row0, bytes_copy);  in stbi__vertical_flip()
1097          memcpy(row0, row1, bytes_copy);  in stbi__vertical_flip()
1099          row0 += bytes_copy;  in stbi__vertical_flip()
2337    __m128i row0, row1, row2, row3, row4, row5, row6, row7;  in stbi__idct_simd()  local
2395          __m128i sum04 = _mm_add_epi16(row0, row4); \  in stbi__idct_simd()
2396          __m128i dif04 = _mm_sub_epi16(row0, row4); \  in stbi__idct_simd()
2413          dct_bfly32o(row0,row7, x0,x7,bias,shift); \  in stbi__idct_simd()
2433    row0 = _mm_load_si128((const __m128i *) (data + 0*8));  in stbi__idct_simd()
2447       dct_interleave16(row0, row4);  in stbi__idct_simd()
2453       dct_interleave16(row0, row2);  in stbi__idct_simd()
2459       dct_interleave16(row0, row1);  in stbi__idct_simd()
2470       __m128i p0 = _mm_packus_epi16(row0, row1); // a0a1a2a3...a7b0b1b2b3...b7  in stbi__idct_simd()
2517    int16x8_t row0, row1, row2, row3, row4, row5, row6, row7;  in stbi__idct_simd()  local
2570       int16x8_t sum04 = vaddq_s16(row0, row4); \  in stbi__idct_simd()
2571       int16x8_t dif04 = vsubq_s16(row0, row4); \  in stbi__idct_simd()
2597       dct_bfly32o(row0,row7, x0,x7,shiftop,shift); \  in stbi__idct_simd()
2604    row0 = vld1q_s16(data + 0*8);  in stbi__idct_simd()
2614    row0 = vaddq_s16(row0, vsetq_lane_s16(1024, vdupq_n_s16(0), 0));  in stbi__idct_simd()
2628       dct_trn16(row0, row1); // a0b0a2b2a4b4a6b6  in stbi__idct_simd()
2634       dct_trn32(row0, row2); // a0b0c0d0a4b4c4d4  in stbi__idct_simd()
2640       dct_trn64(row0, row4); // a0b0c0d0e0f0g0h0  in stbi__idct_simd()
2658       uint8x8_t p0 = vqrshrun_n_s16(row0, 1);  in stbi__idct_simd()