The processor is undoubtedly the most visible technology showcasing the pace of technological advancement in our society, we can proudly say our cellphones are faster than the building-sized ENIAC of the 1940s, and this is all due in part to the rapid development of the transistor and on a larger scale, the processor. These days we’ve got dual cores, quad cores and soon octo-cores coming down the pipe at speeds nearing 4GHz. I think people should take a look at how far we’ve really come, and if we really need to go much further in the current computing era.
First, a bit of recent background. Several years ago, Intel was focused on it’s blow dryer worthy Pentium 4. They thought the future was in higher clocks and bigger chips. AMD, on the other hand, had cooler chips at lower clock speeds (meaning the number of GHz or MHz) that were still outperforming Intel’s best chips. AMD thought to put two of their processing cores onto a single chip, effectively doubling the performance without increasing the clock speed. While in reality this didn’t offer perfect scaling (meaning it didn’t offer quite 200% improvement over a single core), it greatly increased AMD’s lead over Intel. Intel responded by putting two Pentium 4 cores into a single chip and naming it a Pentium D. The Pentium D was the apex of Intel’s volcano based processor technology (called the Netburst Architecture). Little did anyone suspect, Intel’s Israel team was working on a processor named the Pentium M, aimed specifically at mobile systems (laptops), that would later become the progenitor to their massively successful Core 2 Duo.
The Pentium M was so much more efficient than the Pentium 4 that Intel shifted their entire mainstream line from Pentium 4/D to a modified M architecture. This was the beginning of the Core architecture. By modifying the Pentium M for desktop use and using two cores, Intel was able to significantly lessen the divide between their processors and AMD’s X2 line. A little while later, the Core 2 models came out and Intel was back in the lead. They even tried putting 4 cores on a single chip, creating the first consumer quad core chips.
AMD has been in quite the bad spot for the past couple of years, never quite able to catch up with Intel in a timely manner. But that is beside the point. Both companies offer affordable processors that have four cores (AMDs Phenom X4 line and Intel’s Quad Core 2 line) and fairly high (above 2.4Ghz) clock speeds. While Intel still has a noticeable lead in performance, I’ve begun to wonder if processors have reached a point where leaps in performance are no longer necessary for most consumers. Most software cannot even take advantage of the hardware given to them (they are built to work with at most 2 cores, but they can’t actively work with more). Unless you are editing video or batch processing data, more speed isn’t all that necessary.
Consider the fact that the fastest Pentium 4 took 225 seconds to Render a 3D image, while a semi-current Intel Quad Core, the QX6850 can render the same image in 38 seconds. The newer processor is just about 600% faster in this case.
Most computers are bottlenecked by their Hard Drives and RAM (amount) far before the processor is a problem. When playing games, it is the GPU being stressed the hardest.
So! With Nehalem on the way for Intel, and Bulldozer and Fusion for AMD, processors will continue to become faster and more threaded (meaning many more cores). I do wonder though, how long will it be till the software catches up? When will going from 2 to 4 to 8 cores really show me faster and better computing outside of media manipulation? Is there anything that actually needs more speed in the first place? Something tells me we need more than just gaming to continue pushing processors forward.