Hi Seisuke,

1) Thank you for reminding about this shortcut and for sharing the associated workaround.

2) Agree.

In other words, it's a kind of performance optimization when using $increment

I was surprised when discovered that $increment with locals is not so well optimized as it can be and as it is for globals. A small proof:

USER> set top=1E7,^||a=0,t=$zh for i=1:1:top {set ^||a=$get(^||a)+1} w $fn($zh-t*1E6/top,"",3)
0.265
USER> set top=1E7,^||a=0,t=$zh for i=1:1:top {do $i(^||a)} w $fn($zh-t*1E6/top,"",3)
0.176
USER> set top=1E7,a=0,t=$zh for i=1:1:top {set a=$get(a)+1} w $fn($zh-t*1E6/top,"",3)
0.050
USER> set top=1E7,a=0,t=$zh for i=1:1:top {do $i(a)} w $fn($zh-t*1E6/top,"",3)
0.056

I intentionally inserted $get() call just to make samples functionality closer to each other.

It's rather difficult to get for what reason $i(a) is not just the same as `set a=$get(a)+1` as it's nothing to do with TP and other globals related stuff here?

Maybe nobody was interested in deeper optimization of $i() in this context...

Robert,
don't you think that traditional (row) storage w/o any indexing would be slow anyway?
If so, maybe the results DemoB are worth publishing as well.