[x265] Fwd: [PATCH] asm code for filterHorizontal_p_p 4 tap filter

[Jason Garrett-Glaser]

+_loop_col:
+    cmp         col,        width
+    jge         _end_col
+
+    FILTER_H4   x0, x1, x3
+    movh        [dst + col], x1
+    add         col,        8
+
+    jmp         _loop_col
+
+_end_col:
+    test        widthleft,  widthleft
+    jz          _next_row

Again, the loop condition should be at the end; this is extremely
overcomplicated and involves many redundant instructions and lines of code.

This should look more like:

+_loop_col:
+    FILTER_H4   x0, x1, x3
+    movh        [dst + col], x1
+    add         col,        8
+    cmp        col, width
+    jl          loop_col

Jason