Posted by Wade Walker on Dec 20, 2013; 2:45am
URL: https://forum.jogamp.org/Hardcoded-float-not-working-tp4030974p4030985.html

That is very strange. Are you looking at any error output from the kernel compilation, to make sure the float literal isn't just failing to compile and doing nothing? Not sure why 1/(float)2 would compile and not 0.5f, though.