As of 459a1512f13ae1fdd404ff9281b02d225ac70a09, the code is unrolled to process two rows per iteration. The output cursor thus needs to be incremented by twice the stride, which is taken care of with SH1ADD. However the original ADD from the original implemetation was incorrectly left over.
Unified · Split
| @@ -52,7 +52,6 @@ func ff_h264_weight_pixels_simple_8_rvv, zve32x, b · 51 unmodified lines | |||
| 52 | vse8.v v8, (a0) | 52 | vse8.v v8, (a0) |
| 53 | vse8.v v9, (t0) | 53 | vse8.v v9, (t0) |
| 54 | sh1add a0, a1, a0 | 54 | sh1add a0, a1, a0 |
| 55 | add a0, a0, a1 | ||
| 56 | bnez a2, 1b | 55 | bnez a2, 1b |
| 57 | 56 | ||
| 58 | ret | 57 | ret |