.

.

ConclusionFor the ideone.com online compiler with UNKNOWN cache architecture it shows clearly that the SMART linked-list is just as good or actually better upto a range of about 6.000 elements. (data at tab data_ideone)

This is HIGHLY cache dependent as is obvious if you look below for the same run, but for the x64 computer (data at tab data_x64).

For x64 the cache architecture is L1,L2,L3 is likely more advanced than at ideone.com. Very quickly it shows that the cache can be utilized better for the Vector. Somewhere around 100 elements SMART LL is caught up with Vector. At 500 elements the Vector is already more efficient.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

ideone.com

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

x64data_x64!A4:D13

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.