Using AitCS is not as fast as using straight Tcl or Jim.
The decision to use AitCS is a question of convenience vs speed.
See the README for more info about speed and test targets.
AitCS is command-based and so will always be slower than
any variable a byte-compiler can touch directly.
These are the results of the "vs" tests (also in the doc dir)
run on my machine. AitCS is disappointingly slower than Tcl
but not that much slower than Jim.
This may change with future versions of Tcl or Jim.
2022-08-22 Running vs tests with Tcl 8.6.12
LD_LIBRARY_PATH=".:" TCLLIBPATH="." TCL8_5_TM_PATH='.' TCL8_6_TM_PATH='.' tclsh8.6 ../tests/vs.tcl
Operation AitCS Tcl Change
exists z 3.404100 2.022600 168.30%
set z 1 4.828000 2.080700 232.04%
exists z 3.686600 2.049600 179.87%
get z 4.068900 2.018800 201.55%
set l {} 4.829400 2.067500 233.59%
lappend l a b c 8.843100 6.530700 135.41%
unset l 5.526700 2.573500 214.75%
incr z 5.139800 2.445300 210.19%
mkexist 6.296000 1.894800 332.28%
gwd z 4.017400 2.381600 168.68%
gwd q 3 3.805900 2.158400 176.33%
setto a b c d e 0 8.348600 4.626400 180.46%
eref z 3.269800 2.650900 123.35%
aref 2.588700 2.156700 120.03%
cref 2.464000 5.420700 45.46%
2022-08-22 Running vs tests with Jim 0.80
JIMLIB="." jimsh ../tests/vs.tcl
Operation AitCS Jim Change
exists z 8.6 8.4 103.29%
set z 1 8.4 6.2 134.28%
exists z 6.0 5.1 117.08%
get z 5.6 4.3 131.69%
set l {} 8.3 6.3 132.53%
lappend l a b c 7.1 7.2 97.98%
unset l 6.1 6.1 100.34%
incr z 6.5 5.3 122.67%
mkexist 2.5 2.5 99.08%
gwd z 9.3 7.3 127.43%
gwd q 3 12.0 10.0 120.00%
setto a b c d e 0 19.0 15.0 126.67%
eref z 3.9 3.4 114.62%
aref 2.2 2.1 102.86%
cref 0.6 13.0 4.97%