| Intel versus AMD CPU Technology | | Print | |
success in getting higher clock rates without proportional performance gains from
the Pentium architecture, as well as its Xeon server-optimized configurations.
In the battle between Intel and AMD, how much opportunity is there still for
hardware to analyze an increased flow of Pentium-style instructions? It is two
different answers that define the differences between the Itanium and Opteron as
the next big step for server computer systems.
Intel is banking on the idea that the hardware for on-chip instruction scheduling
hardware is nearing its limit. Intel is proposing to look at programs when they are
compiled into a form that is executable and encode concurrent operations ahead of
time.
This EPIC approach, as Intel calls it, is the main difference between the Itanium and
AMD’s x 86-64 in their continuing competition. There is a drawback to EPIC. It does
not offer an upward-compatible path for x 86 codes that already exists, and its speed
in running such code has been somewhat disappointing.
This drawback gives AMD an opportunity. They are betting that it will be
cheaper to perform a duplication of volume-produced microprocessors than
revamping the software base. The Opteron and the Athlon 64 use all of the AMD
available talent to running x86 instructions as fast as possible, while still introducing
64-bit hardware and instruction set extensions.
If vendors of industry infrastructure and software developers think AMD will do well,
IT buyers will see a big number of optimized driver software, middleware, and
applications for x86-64 (AMD) that may add tens of percentages to the performance
of Opteron as well Athlon 64 machines.
Right now, AMD is winning the race, but AMD’s and Intel’s strategies are really pretty
similar. For instance, it is a safe assumption that computing requirements will
expand in all economic segments for 64-processors like Opteron and Itanium. Both
work with growing data collections, have intrinsic computing speed, and both have
multi-way scalability.
The thing is, though, 64-big computing is not a new innovation. Many hardware
builders offer respected families of such systems that are doing well in a number of
applications, but they are failing to achieve the low mass-market prices that will be
needed if they are to apply to a broader base. The question then, is how many high
end processor architectures can this industry develop and build and yet still afford?
As Intel and AMD battle, each subsequent generation raises the stakes. However, in
the end, only a handful of processor families will be able to survive this.
What AMD and Intel will have to do is see all of the enterprise of the IT buyer. There
needs to be a readiness to acquire and incorporate new bases of operating systems,
enterprise middleware, and applications that are to take advantage of the Itanium
build’s new instruction set.
All of this may seem very confusing to the lay person or everyday computer industry
consumer. That is because it is a little bit. What all of this Intel versus AMD
competition and bickering really means is just good news.
Competition in any industry is a way of raising the bar and lowering the costs. As
long as Intel and AMD compete in this processor technology market, it just means
that notebooks, PDA’s, and even personal home computers will become faster,
better, and less expensive. Business, lay people, and techies alike will gain the
benefit of healthy open market competition. No matter what the final outcome of
who has the better processor technology, the technology will improve, and will do so
at a lower cost to consumers and businesses.
So don’t worry if it all doesn’t make sense in terms of Pentiums and Athlons and
x86’s. You don’t have to always know what it all means as long as those competing
and increasing the technology do. In the end, all you have to really do is watch the
prices of your laptops and computers fall as the speed of your PDA’s and processors
rise. Free market, open competition is good for the industry, for you, and for
business in general.
| Users' Comments (0) |
|
No comment posted
