[net.unix-wizards] Does C support G_ and H_ floating point arithmetic?

tho%carcvax.uconn.csnet@csnet-relay.arpa (Tho Tran) (04/09/86)

Is is true that the C compiler on the VAX does not support all floating
point data types provided by the VAX macro assembler?  

If you look at what is available on the VAX macro assembler, there are four 
groups of floating point instructions, one for each floating point data 
type: F_floating, D_floating, G_floating, and H_floating.  F_floating
is of float and D_floating of double.

For instance, when you have a program as listed below, variables a and b 
are treated as D_float; which is clearly shown if you compile with the -S
option that generates assembler code.

      main () {

         double a, b;
            
         a=1;
         b=1;
         
         a = a +b;
      
      } 

Can anyone tell me how to force the C compiler to use the G_float or H_float?
Please respond by E-mail to:
    tho%swevax.uconn.csnet@csnet-relay

thanks
-tho

rbj@icst-cmr.arpa (Root Boy Jim) (04/09/86)

	Is is true that the C compiler on the VAX does not support all floating
	point data types provided by the VAX macro assembler?  

G & H format are relatively new additions to the VAX. There is a package
available by FTP'ing to SEISMO.ARPA. Its in something like
/mod.sources/Volume3/G.format.[0123]. I can send them to you if need be.
If you install these, you probably won't be able to use D & F formats
unless it's controlled by a switch. And you'll probably have to recompile
all the floating point librarys, or keep two versions. 

	(Root Boy) Jim Cottrell		<rbj@cmr>
	Baseball in D.C. in `87!