Unit 5 Analysis 1: Pentium Flaw
The nature of the Pentium flaw was in the floating – point math subsection. In certain conditions, low order bits of the result of floating-point division operations would be incorrect. This would lead to an error that will quickly compound in operations to larger errors when future calculations were completed.
The flaw was first discovered in June 1994, when Intel engineers first noticed that the floating-point unit (FPU) on the chip gave the wrong answer. However future tests done by Intel showed that the error appeared only in the ninth significant digit of the answer and affected a very, very small percentage of possible division combinations. The results of this flaw was that Intel corrected the error in subsequent production, but did not make it known to the public.
Intel’s original solution to this information was to keep it hidden from the public. They felt, that because the flaw was trivial and very small, that most people would not even notice it. It was only after the New York Times ran an article by journalist John Markoff who spotlighted the error in the Pentium chips, that Intel finally changed its position, disclosed the issue with the flaw, and offered to replace every faulty chip. This resulted in a $500 million charge against Intel’s 1994 revenue.
Intel should have come clean at the beginning about the issue with the flaws, instead of waiting for it to come out in the news. Because they waited, the issue only got worse, and thus Intel’s reputation was damaged. Had they informed the public right off the bat, it likely would have saved them damage to the reputation and money, money that was spent to fix a problem that they should have fixed right at the beginning. Their hesitation cause them to end up having a lot of upset customers.
If this same type of flaw was to be found in a new CPU today, the company would surely fail. With a problem in the floating-point math...
Please join StudyMode to read the full document