does anyone happen to know the overhead involved in parameter passing via C calling convention?
By the disassembly of this test file I can see there are some additional instructions compared to the implicit-use-of-global-data version but I'm not familiar enough with x86 to be able to make a determination on whether these are "free" or not and in what cases they might or might not be
yes I know this "probably doesn't matter" and is "microoptimizing and rabbitholing" but I want to know the answer anyway