1091       stbi_uc *row1 = bytes + (h - row - 1)*bytes_per_row;  in stbi__vertical_flip()  local
1097          memcpy(row0, row1, bytes_copy);  in stbi__vertical_flip()
1098          memcpy(row1, temp, bytes_copy);  in stbi__vertical_flip()
1100          row1 += bytes_copy;  in stbi__vertical_flip()
2337    __m128i row0, row1, row2, row3, row4, row5, row6, row7;  in stbi__idct_simd()  local
2405          dct_rot(y1o,y3o, row5,row1, rot3_0,rot3_1); \  in stbi__idct_simd()
2406          __m128i sum17 = _mm_add_epi16(row1, row7); \  in stbi__idct_simd()
2414          dct_bfly32o(row1,row6, x1,x6,bias,shift); \  in stbi__idct_simd()
2434    row1 = _mm_load_si128((const __m128i *) (data + 1*8));  in stbi__idct_simd()
2448       dct_interleave16(row1, row5);  in stbi__idct_simd()
2454       dct_interleave16(row1, row3);  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
2579       int16x8_t sum15 = vaddq_s16(row1, row5); \  in stbi__idct_simd()
2580       int16x8_t sum17 = vaddq_s16(row1, row7); \  in stbi__idct_simd()
2596       dct_long_mac(x7, sump14o, row1, rot3_3); \  in stbi__idct_simd()
2598       dct_bfly32o(row1,row6, x1,x6,shiftop,shift); \  in stbi__idct_simd()
2605    row1 = vld1q_s16(data + 1*8);  in stbi__idct_simd()
2628       dct_trn16(row0, row1); // a0b0a2b2a4b4a6b6  in stbi__idct_simd()
2635       dct_trn32(row1, row3);  in stbi__idct_simd()
2641       dct_trn64(row1, row5);  in stbi__idct_simd()
2659       uint8x8_t p1 = vqrshrun_n_s16(row1, 1);  in stbi__idct_simd()