The documentation for -fast says "
The -fast option acts like a macro expansion on the
command line. Therefore, you can override any of the
expanded options by following -fast with the desired
option."
However, I am unable to find a documented override to turn off the -fsingle option included by -fast. What am I missing?
-fprecision only works on x86. And there do not change the speed of the program.
The default on x86 is extended precision, and will probably give the poster what he wants by default.
So phargrov is right, there is no way to turn of -fsingle. But why he would like to I do not know. Where you loose precision on -fsingle is in expressions like:
[code]float a,b,c,d;
a= b*c+-*a;[/code]
Then with -fsingle b*c, d*a, and the subtraction is done in single precision, while
under K&R it is computed as if the expression was:
[code]a= (double)b*c-(double)d*a;[/code]
-fsingle is the default for all compilation modes except the -Xs and -Xt (K&R and K&R preferred).
So if you would need to cancel the -fsingle you must be compiling with -Xs or -Xt.
