* ssa_2 = fadd(ssa_1.x, ssa_1.y)
*
* While this is "worse" because it adds a bunch of unneeded dependencies, it
- * actually makes it much easier for vec4-based backends to coalesce the MOV's
+ * actually makes it much easier for vec4-based backends to coalesce the MOVs
* that result from the vec4 operation because it doesn't have to worry about
* quite as many reads.
*/