The compiler allows you to use all the standard run-time libraries. By default, the compiler automatically expands a number of standard C, C++, and math library functions. For more information, see Inline Expansion of Library Functions.
The following libraries are supplied.
| Library | Description |
|---|---|
|
libc.a |
GNU* C library (included with Red Hat* Linux*) |
|
libguide.a |
for OpenMP* implementation |
|
libsvml.a |
short vector math library |
|
libirc.a |
Intel support library for PGO and CPU dispatch |
|
libimf.a |
Intel math library |
|
libcprts.a |
Dinkumware C++ Library |
|
libcxa.a |
Intel support library for EH and RTTI |
If you want to link your program with alternate or additional libraries, specify them at the end of the command line. For example, to compile and link hello.cpp with mylib.a, use the following command:
IA-32 systems: prompt>icc -ohello hello.cpp mylib.a
Itanium(TM)-based systems: prompt>ecc -ohello hello.cpp mylib.a
The mylib.a library appears prior to the libimf.a library in the command line for the LINK linker.
In the compiler package, you received the Intel math library, libimf.a, which contains optimized versions of the math functions in the standard C run-time library. The functions in the library are optimized for program execution speed on the PentiumŪ processor.
To enable the optimized math library, the installation creates a directory for libimf.a and adds the new directory path to the LIB variable. Intel recommends you keep libimf.a in the first directory specified in the path.
The -fdiv_check option enables a software patch on IA-32 for the floating-point division flaw that exists on some steppings of the Pentium processor. This patch ensures correct precision of your floating-point division calculations.
Note
The -fdiv_check option is off by default when you specify -tpp5.
When you enable -fdiv_check, the compiler links your programs with libm_chk.a instead of libimf.a. As a result, you enable the support routines to fix the floating-point division flaw for the affected functions.
Use-fdiv_check- to disable the software patch for the floating-point division flaw regardless of whatever other options are specified. When you specify -fdiv_check-, the compiler links with libimf.a and uses simple hardware instructions for floating-point division and affected intrinsics. Similarly, specify -fdiv_check- to disable the special version of the optimized math library (libm_chk.a). The -fdiv_check- option is the default.