In my opinion, the fact that the actual chipset degrades and causes data/throughput loss on SATA devices is way better than a floating point error that can be worked around in software. Obviously and external SATA adaptor on the PCI bus will remedy the problem, but that is certainly more expensive for the end user. If I purchased a board that exhibited the problems, I would certainly be tempted to install a pci-bus controller rather than completely rebuild my system.
Sincerely, Jim Darrough -----Original Message----- From: euglug-boun...@euglug.org [mailto:euglug-boun...@euglug.org] On Behalf Of Alan Sent: Wednesday, February 02, 2011 3:09 PM To: Eugene Unix and Gnu/Linux User Group Subject: Re: [Eug-lug] Intel announces Sandy Bridge chipset recall On Tue, 2011-02-01 at 10:15 -0800, Michael Miller wrote: > This is really no different than in 1994 with the Pentium FDIV bug. > > http://en.wikipedia.org/wiki/Pentium_FDIV_bug > > It's going to cost them money. AMD might be able to get a leg up on > Intel in the X86 market. Actually that is not quite true. The FDIV bug hit everyone. The Sandy Bridge problem is that the parts degrade after time. (Which is how it got past QA in the first place.) Failures that happen later and are not a total collapse are hard to detect. _______________________________________________ EUGLUG mailing list euglug@euglug.org http://www.euglug.org/mailman/listinfo/euglug
_______________________________________________ EUGLUG mailing list euglug@euglug.org http://www.euglug.org/mailman/listinfo/euglug
