During the 1980s and 1990s, accelerating a processor with specialized hardware was a fairly common goal. Essentially, the aim was to receive the benefits (total or partial) of a new generation of CPUs while keeping the original platform. Intel explored this concept for several years with its OverDrive units, aimed at 486, Pentium, and Pentium Pro computers. The success of the OverDrive family was very limited, but these days we could learn a thing or two from it…
One detail that old computers had in common was their price. You had to think it over very well before investing hundreds, or why not thousands of dollars in a computer that would become (for better or worse) another member of the family. Every single component was expensive, and even upgrades required careful analysis. In those years, users sought to squeeze the most out of their configurations and make them last over time, something diametrically opposed to the planned obsolescence that consumes us today.
The OverDrive Concept
That need gave rise to hardware like the so-called "CPU accelerators", which were generally developed by third parties, and adopted different formats depending on the computer that would receive them. With the spectacular advance of the 80x86 platforms and the introduction of the famous "sockets", Intel also created its own accelerators. The adventure began with RapidCAD chips that enabled access to an Intel 486DX processor on a 386 motherboard, but the whole concept became more exposed to the public thanks to the OverDrive brand. Depending on the model, an OverDrive processor could take on the role of direct replacement, or occupy a special socket alongside the original chip that was soldered to the PCB, and therefore impossible to remove. OverDrive processors came with an integrated heatsink, along with several additional voltage regulators. Intel launched more than a dozen variants ready to "improve" 486 computers, but the OverDrive name extended to Pentium, Pentium MMX, and even Pentium II in Socket 8 format to replace Pentium Pro.
Why OverDrive Didn't Succeed
Now, if the benefits were there, why didn't they succeed? The main reasons were three: First, their cost-benefit ratio was very poor. For example, the Pentium Overdrive 83 MHz debuted with a price of $299 in 1995, when at that time a user had the possibility of acquiring a competitor processor and obtaining superior performance (except in those cases with a special emphasis on FPU) for a fraction of that value. Second, compatibility conflicts were the order of the day, and beyond the efforts made by Intel, many motherboards were incapable of using these chips to their full potential, or lost functions in the attempt (such as external L2 cache). And third, access to new processors on old boards affected sales of new PCs. Personally, I think many users would love to have some variant of modern processors that can be used in retired sockets, like LGA775 (Intel) or AM2 (AMD), but let's say the manufacturers themselves took care of burying that...
