Reported in Robert Collins' x86 Monthly Digest, a well-known Web site for observers of Intel activity, the brouhaha refers to an arcane problem allegedly experienced by Pentium II and Pentium Pro chips converting floating point to integer numbers. At worst, the error could lead to bad data, but, according to experts, the problem is likely to be a storm in a tea cup when compared to the "FDIV" Pentium fiasco of a few years back when the chip giant was forced into offering a recall of devices affected by a floating point calculation error.
"The situation is that we're co-operating with Mr Collins and keeping OEMs informed to establish the problem, if any," said an Intel UK spokeswoman this morning. "If it turns out to be a problem, we'll publish the new errata by the end of the week. Until we conduct our own tests it's difficult to say more, but we're jumping all over [the problem]."