Another thought regarding memory is the testing which will be so important to these modules. Testing will continue to be very important as memory speeds increase. One of the problems with all the rock-bottom memory prices has been quality of memory. Because the prices have been so low we continue to believe that many companies just can't afford to correctly test memory.

Of course, certifying and testing memory modules leads to added stability, and as many of you know, these properly tested modules retail for much higher than some of the bargain-basement memory that many people are using. We can't even tell you the amount of email that we receive on a regular basis where you can almost tell that the root of the problem is that users are trying to push their systems with cheap memory.

The stability of the system will be affected by modules that do not perform to specifications or have not been tested to meet established industry standards. No matter if it is DDR or RDRAM, testing will make the difference on module integrity. To this end we asked one memory supplier, Kingston, if they could provide some insight into this important issue.

Dr. Ramon Co, VP of Technology & Test Engineering, Kingston, explains, "Kingston's test engineers perform a worst-case analysis for each memory technology (regardless of the merits of each memory technology) in terms of loading, reflections, timing, signal integrity, and other factors. Then, we ensure that our production testing accounts for this variability for the best compatibility with end-customer systems."

Based on this one must realize that with these advanced technologies, no matter if it is DDR or RDRAM, the days of cheap memory are over. Make no mistake, with the continued consolidation of memory manufacturers, coupled with the advanced technology, the price of memory is on the upswing. ...memory is a "what you get is what you pay for" deal as well.

Progress, Progress...All We Hear About Is Progress...

As we have pointed out in this article, progress will be made and perhaps it will be a strange combination of both new and old technology. Maybe we were too hard on Rambus from the beginning or perhaps we were not ready for the evolution that it offered us. Better yet it might be said that processor technology was not yet ready to fully exploit the potiential of Rambus. Or perhaps many of us were just confused about the fact that Rambus was a serial technology, rather than the parallel technology that we all have had more direct experience with.

At least for the time being, if you still have some of those PC800 RDRAM modules, you can certainly put them to good use with some of the current crop of 850 motherboards. This will result in a performance gain over the Intel DDR boards.

A last observation: In recent years, Intel may have made more technological progress than the market could handle in such a short period of time. Maybe that progress was the ultimate problem.

We ultimately believe that in hitching its processor roadmap to Rambus, Intel is ensuring its ability to push CPUs into higher speeds. Rambus has a clear cut roadmap that Intel can target, and that's been the case since the beginning.

The fruits of that strategy are just beginning to ripen. It's all about timing in the memory game, timing.