Don't
get snowed by computing's 64-bit hype
by Alan Zisman (c) 2005 First
published in
Business
in Vancouver April
19-25, 2005; issue 808 High Tech Office column
The London Drugs flyer offers a pair of Compaq Presario notebooks.
Powered by AMD Athlon 64 processors, the screens of each claims "64-bit
computing" in big, bold black on yellow lettering.
Sixty-four bits? Should you care?
The current generation of computers processes data 32 bits at a time.
Each "bit" is a power of 2: 32 bits means 32 2s multiplied together. If
you start counting up the series: 2, 4, 8, 16, 32, 64 and on and on,
the numbers start getting large pretty quickly. Two to the 32nd power
is somewhat over four billion - four gigabytes in computer-speak. A
32-bit processor, according to the Wikipedia free encyclopedia, "can
address 4 GB of byte-addressable memory."
"Surely four GB of memory ought to be enough for anyone," I hear you
say.
I don't know anyone with that much computer memory installed on a home
or business computer. Still, some of us are starting to bump into
32-bit memory limitations. If you're working with very large files,
such as video files or large databases, you can start running into
those limitations. Software performs better when files are loaded into
memory for quick access. This can't be done with today's large files on
32-bit systems, and the result can bring even a fast modern computer to
its knees.
We've gone through such a shift in the past. The first generation
personal computers were eight-bit systems. IBM's original 1981 PC used
a 16-bit processor (crippled with an eight-bit memory bus) that could
access a then-expansive one-megabyte of memory. In 1984, Apple's
original Macintosh used a 32-bit processor. In 1985, Intel released the
386, its first 32-bit processor; 386-powered computers from Compaq, IBM
and others soon followed.
It was only with the release of Microsoft's Windows 95 that most
PC-users had an operating system and software that took advantage of
32-bit power. And that's the catch.
Hardware isn't usable without software; 32-bit hardware required an
operating system and applications designed for it. Until that was
available, 32-bit systems were just fast (for their times, at least)
and expensive computers running 16-bit software.
We're in a similar transition period now. Apple, for example,
advertises that its G5-powered models are "ready to run modern 64-bit
applications." Intel-competitor AMD notes that its Opteron and Athlon
64 processors offer "simultaneous 32-bit and 64-bit computing," and
points out that Microsoft has launched a pre-release beta version of
Windows XP 64-bit edition, with the release version rumoured for this
spring. (There is already 64-bit Linux support for AMD64 processors.)
That simultaneous 32-bit and 64-bit phrase is revealing; even with
cutting-edge 64-bit hardware, you'll need to be able to run your
current stock of 32-bit software. That was a limitation in Intel's
initial 64-bit Itanium processor. Designed to zoom through mostly
non-existent 64-bit software, it ran current-generation 32-bit software
slower than cheaper 32-bit hardware. Not surprisingly, sales were slow.
Intel is now pushing to catch up to AMD with the development of 64-bit
processors that will also do a good job running operating systems and
applications designed for 32-bit hardware. Intel has just released
64-bit versions of its Pentium 4 processors.
In 1988, a Compaq DeskPro-386 or IBM PS/2 Model 80 would have cost you
about $10,000 for a cutting-edge 32-bit system, but by the time Windows
95 finally brought 32-bit computing to the masses, they were
seven-year-old boat anchors. 64-bit Athlon 64, Intel Pentium 4 6xx, or
Apple G5-powered systems sell for quite a bit less. As a result, you
can now purchase a 64-bit PC or Mac for an affordable price.
Just don't get snowed in by that "64-bit" phrase. For most of us, it
offers no benefits running today's software. And by the time tomorrow's
64-bit software gets released, you're going to want to run it on
tomorrow's hardware.