SDRAM: Hype or the Future

First introduced to the desktop PC market over a year ago, Synchronous Dynamic Random Access Memory (SDRAM for short) has since then become a hideous fad most PC users have caught on to, however is SDRAM really necessary?  To answer that question you must first look at the migration from good old 72pin SIMMs to the 168pin monsters you see today called DIMMs.  It all started in a place you wouldn’t normally consider, inside Apple Computer's own Power Macintosh. What exactly was inside these Power MACs?

Limitation #1

Support for up to and in excess of 1GB of RAM. Apple knew that the type of users that would absolutely need a MAC would be professionals interested in DV editing as well as desktop publishing...in turn, they needed RAM. In order to fulfill the demanding needs of the user, Apple took a huge step forward from the normal means of memory expansion, which both PC and MAC owners had been using for years: 72pin SIMM (Single Inline Memory Modules). At the time, the 486 was still popular among average users, Cyrix had just introduced their 150MHz PR-200+ processor and the 32MB EDO SIMM finally made its way into the mainstream market. Realizing that installing a total of 1GB of RAM with 32MB SIMMs would put heavy limitations on various aspects of the motherboard's design (physical layout, chipset, etc...), Apple therefore decided to make an almost instantaneous migration to the newly announced RAM specification standard, the Dual In-line Memory Module, commonly known as the DIMM. Mind you, we are still not talking about SDRAM, just plain old EDO DIMMs here. But why on earth would we invent another way of packaging RAM with no advantages over the technology that was current at the time? We wouldn't...that's why Apple implemented a technology with its fair set of advantages over standard SIMMs. Using the technology available at the time, one could easily fit 128MB on a single module, therefore increasing the realistic possibility of surpassing the 128MB and 512MB limits of everyday desktop PCs. Apple began outfitting their Power Macs with up to 12 of these DIMM slots, featuring 168 pins (84 pins on each side) and measuring approximately 5.25" in length. Apple managed to eliminate one major problem a 1GB memory capacity posed, physically fitting memory slots on the motherboard. However, one more major limitation remained, the chipset.

Limitation #2

Ever wonder why you only see a maximum of 3 DIMM and 4 SIMM slots on a TX based motherboard? And have you ever wondered why you cannot occupy all 7 slots at once? The answer to this question lies in between the "lines." A major function of the chipset on a motherboard is to map out the expandability features of that particular board, part of this duty deals with the number of RAS (Row Access Strobe) lines the chipset provides for. The TX chipset, for example, has 6 allocable RAS lines, with each Double RAS SDRAM DIMM occupying 2 RAS lines, a maximum of 3 DIMMs can be installed at once without exceeding this limit. Likewise, if Apple chose to stick with the 72 pin SIMM standard that would provide them with one major problem allocating the required number of RAS lines from the chipset to correspond for the number of SIMMs installed. Imagine this, a double RAS 32MB EDO DIMM occupies 2 RAS lines, and you would need at least 32 modules...an AMAZING amount! However, with 168 pin DIMMs you can simply use a few 128MB sticks and your problems are solved.