Why we need a double precision functions?

[if001]

For example, there is a same result

REAL(8) :: log_double = 5.D0

PRINT*,"LOG ( DOUBLE ) =", LOG(log_double)

PRINT*,"DLOG( DOUBLE ) =", DLOG(log_double)

LOG ( DOUBLE ) = 1.60943791243410

DLOG( DOUBLE ) = 1.60943791243410

Thanks for answers,

Pete.

[mecej4]

In Fortran 77 and earlier, there was ALOG for single precision arguments and DLOG for double precision arguments. Fortran 90 introduced the generic function LOG to make it easier to write code without having to remember the specific names of all the intrinsic functions. LOG() is equivalent to ALOG and LOG() is equivalent to DLOG.

The specific names are still retained for the sake of backwards compatibility.

[if001]

Thank you. I hope that a same way was made for SIN, COS and etc.....

[TimP]

All of the intrinsic functions mentioned in this thread were made generic in the f77 standard, making the f66 double precision versions de facto obsolescent. Internally, the support library may have separate entry points for various data types (not to mention scalar vs. vector, and for various instruction sets and, in ifort 12, arch-compatibility vs. specific sub-architectures), all accessed via the standard intrinsic names.

[if001]

Oh, the last stuppid question please!
About entry points.
Will be my prog run "faster" if old fashion DSIN(double) vs SIN(double)is using?

[Steven_L_Intel1]

No. Exactly the same run-time code is used in either case. However, use of specific functions such as DSIN can cause maintenance problems later so I recommend using the generic functions where possible.

[if001]

Thank you Steve,

This is sad news for all my previuos program code. :)
DEC is in our hearts. ;)