|
|
If you like this site, please support us by spreading
the word.
We have made it easy to let your friends know about
us. Just click on the Recommend-It icon below.
|
|
 |
|
| AMD, set to take over the world? |
|
|
In the beginning
Intel and AMD once worked hand in hand when manufacturing CPUs. It was not uncommon to see both of the brands on
one chip. However, somewhere along their course, the two companies decided to split up for a number of reasons.
Anyway, the situation has been the same ever since and the chip manufacturers became head to head competitors.
In general, Intel and AMD produced similar CPUs in terms of normal Microsoft Windows performance, but Intel has
always maintained a higher position. |
|
Only Intel inside will do
From day one, Intel always had the performance edge. It produced faster CPUs and had major advertising campaigns
beckoning the consumer with "Intel inside" logos. Intel really started its advertising campaigns in the
486 era, where it told the consumer that "Intel inside" computers were the best, and the consumer didn't
know what else to choose. AMD and Cyrix (the other big players) weren't advertising like Intel, why would anyone
want them?
However, Intel was right, they did make the best CPUs. The 486dx had an FPU (Floating Point Unit); the competition
at the time did not. So what does FPU mean? Well, essentially, speed. At this time, Intel offered products with
more features than the rest, and people had good reason to spend money on Intel CPUs. The FPU was the element of
raw power; many office applications were beginning to rely on them, and often operated at unsatisfactory speeds
without an FPU. The FPU is basically a powerful number crunching part of the CPU which can significantly speed
up programs which use it. At the time, even games were emerging which required FPUs - most notably Doom.
Gaming
Doom wouldn't run without an FPU. Doom also happened to be the most advanced game on the market. Gaming consumers
were Doom-mad; "3D" gaming was beginning to emerge. However, 3D gaming required a lot of CPU power, and
gamers knew that Intel made the best CPUs for these games. Time rolled on, and Doom was not as CPU intensive as
it once was - it was not necessary to have the most powerful CPU to play it, and you could play it on systems with
sub-optimal components.
Then Quake appeared. Quake has been a rather influential factor on the development of computer hardware. It is
thought to be a major contributor to the success and advance of 3D accelerator cards but in the CPU world, it made
one thing clear - if you wanted the best gaming performance, you needed an Intel CPU. Quake was a notorious FPU
thrasher; it needed immense amounts of power to run properly, so you needed a chip with a good FPU to run it well.
If you didn't have a chip with a good FPU, then Quake would not run to your satisfaction. Why was this? It was
because gamers had now become obsessed with FPS, or Frames Per Second. This was the phrase along which every quake
gamer has used to judge the speed of hardware they could use. If one CPU could offer them better FPS over another,
then the faster CPU would be better. And why was this? Because if the game ran faster, the gamer had a chance of
playing better, and therefore winning. Quake was a network and internet God, and if you were good at Quake, you
too were a God, so by default, you wanted a fast CPU.
In the Pentium era, FPUs were becoming important. It just so happened that Intel Pentiums had a very strong FPU,
and the Pentium advertised itself amongst the Gaming world as being the chip to buy. Forget the MMX adverts
and "Intel inside" logos, the Pentium was fast, and people were buying it for its merits.
AMD and Cyrix were happily skipping behind Intel, but they could not keep up. Their principle problem was their
"P" ratings. The P rating essentially advertised the theoretical performance of their CPUs. Cyrix in
particular were advertising P166 chips which actually ran at 150MHz. The consumer knew that the CPUs ran at 150MHz,
and could not understand how a 150MHz could be the equivalent of a 166MHz Intel Pentium. However, as the P rating
kept the average consumer away, FPU power kept the gamer away. The AMD K6 was a fast CPU in terms of Windows performance,
it actually ran normal programs faster than the Intel Pentium at equivalent speeds. However, on both the K6 and
Cyrix 6x86, FPU performance was appalling. Gamers wouldn't touch anything other than an Intel.
Tough competition
The story continued with the Pentium 2. The Pentium 2 was a the peak of Intel's dream. Intel once again produced
a far superior product, with AMD's K6-2 still lagging behind, despite having introduced 3DNOW! - AMD's equivalent
to Intel's MMX, which was designed to increase FPU performance in games. The Pentium 2 was the clear market leader
when it came to performance. AMD couldn't keep up, and Cyrix didn't stand a chance. By now Cyrix was really having
problems, stuck in the middle of the Intel and AMD; competition was tough, and if your products weren't at the
cutting edge, then you were going to have problems. Cyrix were close to being thrown out of the market, and the
final decision to leave followed shortly.
The P3 hit the shelves early in 1999 with the K6-3 appearing not long afterwards. Intel were blazing in glory,
the P3's entrance was unhindered by any form of competition. The K6-2 was now far off, and although the K6-3 was
based on much better technology than the K6-2, it was still operating on an ancient architecture- the Socket 7.
Intel left the socket 7 platform with the Pentium Pro, and the more mainstream P2; the P3 was also originally based
on the Slot format. This gave Intel a performance advantage especially with memory performance, showing the P3
to have a clear advantage over the K6-3.
However, why was AMD still continuing to release chips on a platform that they knew was dying? The answer was that
they weren't, the Athlon was almost ready to roll, and was the factor that could make or break AMD.
Before the Athlon, AMD had one major flaw, they produced budget chips. This is fine if you wanted a cheap PC with
mediocre performance, but in general, people don't. They want a PC with all the bells and whistles which has a
remote possibility of being future proof - second best just isn't good enough. This meant that OEMs were selling
Intel PCs rather than AMD PCs. To get a larger market share, AMD had to venture into making processors that performed.
The Athlon was a risky project for AMD. If it didn't sell, then major losses would be AMD's punishment, which is
why the K6-3 was launched prior to the Athlon. It was essentially a fallback, buying AMD time to re-think their
strategy if Athlon failed.
Enter the Athlon
The Athlon was a new stance for AMD. Launching into the performance market was very risky. Could they compete with
the brand image of Intel's new P3? Well, the answer was yes, the Athlon certainly could. Boasting a superior FPU
unit, and a more competitive price, the Athlon was designed to get Intel's attention. The Athlon was AMD's first
real performance CPU aimed directly at the top end P3s. As we all know, the Athlon was a success and loved by gamers.
The K6-3 project was faced with an early death and the Athlon became mainstream.
However, this created a new problem. AMD now had the favour of the performance demanding consumer but where was
it's budget chip now? The K6-2 was dragged back out of the closet with 500 and 550MHz flavours. These chips were
immensely cheap, but compared to the performance of Intel's fruity celeron, they simply weren't fast enough, even
for the budget market.
A year passed, with the Athlon still going strong, beating Intel past the ellusive 1GHz barrier. Then the issue
of the budget chip was addressed: the AMD Duron was released. The Duron, like the K6-2 series, was cheap. However,
whereas the Celeron ran circles around the K6-2, the Duron ate the Celeron. The Duron was cheaper and faster than
the Celeron. Intel had serious competition.
With the release of the new style Athlon, the Thunderbird, Intel was beginning to find itself swimming in deep
water. The T-bird was as fast, or faster, than similarly clocked P3s. However, it was also significantly cheaper.
AMD was running all over Intel, so much that Intel, in a mad panic, began doing things never seen before. |
| |
|
|
Continued . . .
|
| |
|
|
Back to Top
|
 |
|
|
