HP-UX Floating-Point Guide

178 Chapter 7

Performance Tuning

Inefﬁcient Code

Producing Position-Independent Code

By default, in 32-bit mode, compilers produce absolute code for HP 9000

systems. You can produce position-independent code (PIC) for use in

building shared libraries. In general, absolute code is faster than PIC

because addressing calculations are simpler and shorter. Consult the

HP-UX Linker and Libraries Online User Guide for more information

about absolute and position-independent code. See “Shared Libraries

versus Archive Libraries” on page 181 for more information on the

performance impact of shared libraries.

The +DA2.0W option, which generates 64-bit code, always generates PIC

code. This is one of several reasons why 64-bit code may run more slowly

than 32-bit code. See “Specifying the Architecture Type” on page 176 for

more information.

Using Proﬁle-Based Optimization

HP C, HP C++, HP ANSI C++, HP FORTRAN/9000, and HP Pascal

support proﬁle-based optimization (PBO) on HP 9000 systems. PBO can

improve the performance of programs that are branch-intensive and that

exhibit poor instruction memory locality. Although these tend not to be

issues in ﬂoating-point-intensive applications, if you suspect that they

may be degrading the performance of your program, you can use PBO to

minimize their impact on your program. Under PBO, the compiler and

linker work to optimize the executable ﬁle, using proﬁle data for a typical

data set to produce an executable ﬁle that will result in fewer instruction

cache misses, Translation Lookaside Buffer (TLB) misses, and

memory page faults. For information about PBO, see the HP-UX Linker

and Libraries Online User Guide and the appropriate compiler

documentation.

Creating and Zeroing Static Data (Fortran

only)

HP Fortran 90 provides an option, +save, that forces static storage for

all local variables and that forces the compiler to initialize all

uninitialized static variables to zero. HP FORTRAN/9000 provides an

equivalent option, -K; the +e option also automatically saves all local

variables, if possible.