Hello
The file {Install}\lib\HCI\readme.txt explains it a little bit.
ansisi.lib is used when you specify you want no floating point support. (option -D__NO_FLOAT__ )
ansisf.lib is used when you want to get float represented as 32-bit IEEE and double as 64-bit IEEE
(option -Cf)
ansis.lib is used when you want to have both float and double represented as 32 bit IEEE.
As far as I can tell the 3 choices are available in the wizard for floating point support.
I hope this makes it a little bit clearer.
CrasyCat