Friday, 30 March 2012 21:34

DEC Helped To Make Both AMD And Intel What They Are Today, But Who Got More Out Of The Deal?

Written by

Reading time is around minutes.

800px-DEC_logo_svg_For those of you that are old enough to remember the birth of the AMD Athlon CPU then you might also remember that AMD picked up quite a bit of the technology from a company called DEC. DEC had a CPU that at the time was simply amazing. To make matters worse it was a 667MHz CPU that was able to run rings around both Intel’s and AMD’s 1GHZ CPUs. There were multiple reasons for this but the most under rated one was the efficient caching structure. The DEC Alpha processor was light years ahead of most CPUs on the market at the time, but due to it being a RISC (Reduced Instruction Set Computing) CPU it was not as flexible as even the slower X86 CPUs.

However, this did not stop DEC from being a very hot commodity; Compaq bought them and then sold them off to Intel after the launch of the Athlon. When this sale took place many AMD fans thought that it would be the end of the Athlon (and possibly AMD). Fortunately that was not the case as AMD had not licensed the technology for the EV6 and EV7 bus directly from DEC. AMD also ended up with a number of DEC’s top engineers right after the purchase. This allowed them to continue to build on the Athlon, but one thing they were never able to do was recreate the efficient caching structure that DEC had with the Alpha. I can vividly remember sitting in on AMD talks where they described how the Athlon microarchitecture would be able to scale up to 64MB of cache at full CPU speed. This has never happened and since the assimilation of DEC by Intel AMD has struggled with balancing the latency issues you get with an IMC and truly efficient caching.  

Meanwhile Intel has taken a lot of what they picked up from DEC and managed to build a very efficient caching structure. We first saw this with Conroe as this dual core CPU managed to exceed expectations and almost break the law of diminishing returns (Kentsfield certainly did). Since then Intel has improved their caching to the point where they were able to efficiently embed a DDR3 enabled Internal Memory controller without the normal loss of performance associated with the high latency of DDR3. This is a pretty impressive feat all on its own and, as anyone that has heard one of Francois Piednoel’s rants, is what makes all the difference between just performance and high-performance. Intel has continued to tweak and adjust the caching on their CPUs and are expected to have a new version when Ivy Bridge comes out. Meanwhile AMD seems to be losing memory and cahcing performance to the point that they have publicly stated they will no longer even try to compete with Intel (which is very sad to be honest)

So while AMD owes the foundation of their CPU line up to DEC, it was Intel in the end that was able to capitalize on the technology they had to offer. If AMD had made the move to purchase DEC back when they were looking into ATi things could be very different now. Instead we watch as AMD slowly moves to mobile CPUs and becomes more of a GPU maker than a CPU maker. At some point you have to wonder, who ended up buying who? We also often wonder where AMD would be today if they had not bought ATi and instead had put that money back into the company. It was a move that many predicted would lead to AMD’s failure and while they are still in existence as a company they are truly only a shell of what they once were.

Discuss this in our Forum

Read 3729 times Last modified on Friday, 30 March 2012 21:48

Leave a comment

Make sure you enter all the required information, indicated by an asterisk (*). HTML code is not allowed.