Benchmark: timing 10000 iterations of scalarNotSaved, scalarSavedGlobally, scalarSavedLocally...
scalarNotSaved:  9 wallclock secs ( 9.14 usr +  0.00 sys =  9.14 CPU) @ 1093.85/s (n=10000)
scalarSavedGlobally:  6 wallclock secs ( 5.63 usr +  0.00 sys =  5.63 CPU) @ 1775.57/s (n=10000)
scalarSavedLocally:  6 wallclock secs ( 5.68 usr +  0.00 sys =  5.68 CPU) @ 1761.18/s (n=10000)