Category: Power Supply
Manufacturer: Cooler Master
Product: Real Power Pro 1000
Gallery: Click Here
Price: $349.99

There are two sides to every story. What one person thinks is right, another will think otherwise. It is a fact of life that we deal with everyday in argument and discussion. The hardware world is the setting for a lot of these conversations; what someone considers progression, another might consider regression. When the term Moore’s Law was first uttered in 1965, then coined in 1970, it became a driving force behind progression in the computing realm.

“The complexity for minimum component costs has increased at a rate of roughly a factor of two per year ... Certainly over the short term this rate can be expected to continue, if not to increase. Over the longer term, the rate of increase is a bit more uncertain, although there is no reason to believe it will not remain nearly constant for at least 10 years. That means by 1975, the number of components per integrated circuit for minimum cost will be 65,000. I believe that such a large circuit can be built on a single wafer.”

For those who are unfamiliar with the message relayed in the above quote, Moore’s Law simply states that the optimum transistor count for processor die (this is a generalization for the purposes of our discussion) will double every two years for the foreseeable future. Over the years, Moore’s Law has developed beyond the context of transistor counts and has almost become a basis for technological growth in general. However if we take a look to see how it has held up over the past 36 years, since Intel released the world’s first microprocessor in the 4004, it is pretty clear that Moore’s prophecy has become self-fulfilling. Intel adopted Moore’s Law as a goal, and as a result has been able to keep it true. The Intel 4004 was built on a circuit of 2,250 transistors. If we multiply 2,250 by 2^18 (doubling every 2 years over 36 years), we end up with 589,824,000 – or 590 million transistors. Intel’s Kentsfield processor, released last year, has 582 million transistors. By the end of this year, Intel’s Penryn processors will most assuredly have surpassed the 590 million mark.

This increase in transistor count yields a tremendous increase in computing power. Processors are operating at higher frequencies than ever before and they are consuming more power than ever before. The vehicle for increasing the amount of transistors you can fit on a die is the manufacturing process used to create the die. When the transistor density increases on a processor die, the power density is also increases. This typically allows for increased efficiency. However, when engineers design new processors to be built on a smaller manufacturing process, they have a few choices to make. Will they opt for lower power consumption, due to the decreased feature sizes on the chip, or will they chose to add functionality (and thereby size) to the chip and utilize the extra transistors that are afforded due to the smaller manufacturing process. History tells us that far more often than not, engineers go for the second option. This, combined with steady frequency increases over the past five or so years, has led to processors consuming more and more power. AMD’s current most powerful Quad FX platform consumes around 500 Watts under full load. Intel’s Kentsfield consumes well over 300 Watts when overclocked and under load.

Using nothing more than my own personal foresight, I have to say that the processor market is starting to head in the wrong direction. Not only are CPU’s using more power, video cards and cooling solutions are too. The result is a complete computer that is capable of consuming more than 1000 Watts of power. 1000 Watts is probably enough to power every lightbulb in your home. Those readers hailing from areas of infamous energy shortages, such as Southern California, will know that turning on and off lightbulbs in your home over the summer can save quite a lot of money. Now imagine that effect being negated because you have a super powerful computer running some CS 1.6 all day. The problems of cost and energy conservation are not what I base my opinion on, however. Perhaps the biggest problem we are facing with computers today is that basically none of the technologies that go into your computer are independent from one another. Without a capable motherboard, your super fast new RAM will not be able to run at its rated speed. Without a capable processor, that video card you just bought with your entire paycheck won’t be able to deliver quite as advertised. And let’s not forget the biggest dependence of them all: Every component in your system would not be able to do anything AT ALL if the power supply was insufficient. Power supply technology, while currently able to keep the pace, could very possibly become overwhelmed in the near future. Trying to produce a power supply to the ATX standard that can produce upwards of 1kW is like trying to shove an elephant into a Volkswagen bug. There are only so many components that can be shrunk in a power supply to reduce size. Transformers, inducers, heatsinks, and fans are constantly increasing in size with increased power output. More efficient MOSFETS might get us a few hundred more Watts, but that performance wall is fast approaching. Standards will have to be rewritten, products will have to be scrapped, and countless millions will need to be poured into R&D to develop new power supplies that can handle the beastly computers that we will see in the next few years, assuming current trends continue.