Intel's Conroe has 67 errata (bugs)
How many are planned to be fixed? Not all of them.
BLURB
I'm not going to write an article as I did the last time for AMD64 and Core architecture errata. If you remember, it caused rather a stir across the Internet. However, it is worth pointing out that Intel's future architecture, its Core 2 x86 flagship architecture, already has 67 errata on file, with only 20 planned fixes. That means 47 errata will persist in future steppings. For the most part, errata will cause no damage to a desktop or server system. However, that they're there and have been identified would seem to be an indicator that Intel knows what's causing them and, therefore, they could be fixed. The reality remains that many companies will not invest resources in fixing problems that do not impact many people. This is true in software, hardware, manufacturing, and other industries as well. If a particular part is defective, it just won't be fixed, because the risk of introducing new problems through the fix, or the expense of testing the new correction (compared to the small expense of actually fixing it) to have proper validation before release is just prohibitive. I've wondered from time to time how a chip making company with as many years of experience as Intel in such a wide array of areas could still release chips that are buggy. I would think that each internal "system" would be nailed down, solidified in design, fully understood, and would behave only as they expected before being introduced into a production chip. And whereas there are always things that slip through the cracks, I would think a high enough value would be placed on fixing it--or making it right--for the end consumer that it would be done that way. Of course, there I go injecting morality or ethics into business. Bad RickGeek, bad... Read about all of
Intel's Core 2 errata PDF (as of August, 2006), and post your thoughts below about Intel's flagship architecture having 47 bugs with no planned fix.
