IBM 88553RX, Installation Manual

Page: 167 / 232

Chapter 4. Installation
153
Keep these comparisons in mind while programming or porting applications to
any 64-bit operating system, not only Linux. It is easy to assume that every data
type is the same size, and that is why the LP Data Model standard was created.
Byte order
The IA-64 kernel of Linux also uses standard byte order at a byte level. The
native byte order is little-endian, while big-endian processes are still possible.
All Intel processors are little-endian based. That means that the bytes at lower
addresses have lower significance given a 16-bit or 32-bit word (the word is
stored “little-end-first”). In big-endian architectures, the bytes addresses have
higher significance (the word is stored “big-end-first”). The IA-64 Linux kernel
uses little-endian by default, but allows for the possibility of using big-endian byte
order.
4.4.2 Choosing a Linux distribution
Unlike other operating systems, when we talk about Linux, we are not talking
about one company, one version, or one distributor. Linux is the kernel, the base
of the system, and there are many companies that develop Linux distributions
built around the kernel.
In this section we discuss the major Linux distributors that work with IBM, and
their different products available for the Itanium 2 platform: Red Hat and
UnitedLinux.
Note that the IBM Linux support page is:
http://www.pc.ibm.com/qtechinfo/MIGR-48NT8D.html
Here you can find information about the Linux distributions supported by IBM,
and documents, information, and drivers for IBM products.
long double 128 128
Note: The LP64 Data Model example discussed in this publication is not the
only 64-bit Data Model defined by The Open Group. There are other models,
such as LP64, ILP64, or LLP64 for 64-bit; and ILP32 or LP32 for 32-bit. For
more information about the data models defined by The Open Group and also
about their standardization activities, go to:
http://www.opengroup.org
Data type in C LP32 LP64