Intel's learning off the Net

A problem with the RZ1000 enhanced IDE disk controller chip from silicon designer PC Tech has saddled Intel and other PCI motherboard manufacturers with disk problems that range from altered numbers in spreadsheets, to complete partition failures.

There are obviously shades of Intel’s other recent bete noire here, the infamously inaccurate floating point processors in early Pentium processors that not only spawned a whole genre of humour based on bad maths, but showed one of the world’s largest corporations that the Internet had established a power beyond any corporate ability to contain. So this time Intel is proclaiming everything it knows to try and sort it out, albeit a year after it first knew, according to reports in InfoWorld.

James J. Mitchell, Business Editor, of the San Jose Mercury News observed “INTERNET FORCES QUICK INTEL RESPONSE”

Random data corruption without warning is no small matter, and in reporting this story, the US weekly InfoWorld said “It’s a problem that financial institutions for example, are terrified of.” According to InfoWorld, the suspect device is used on boards using the Mercury and Neptune chip sets, and Intel’s own Web page on the matter says: Operating systems which are unaffected by this problem when run on Intel motherboards include Win 95, Win NT ver 3.5 and later, Win 2.x, DOS. Win 3.x is unaffected when using 16-bit disk accesses.

And then goes on to say: “At this time our extensive testing has not uncovered any problems with any Windows or Windows For Workgroups using 32-bit IDE disk access.”

Confused? Look at for up to date enlightenment.
Those who recall the matter of the Pentium maths coprocessor that couldn’t do its sums will be encouraged by the “new Intel”, who now provide one of the most significant endorsements of the Internet as a vital business tool for anyone whose business relies on IT:
Intel is committed to ensuring that users of our products are satisfied. We are therefore taking the following steps to address the concerns we have seen on the various usenet groups.

Fully characterize the problem.
Frequently update the user community on our progress.
Work with OS vendors to develop and distribute patches, as required.

Develop a utility program to identify which systems are susceptible to this problem.

It is clear that Intel have learned from the FPU farrago: what a difference the power of the Internet can make to one of the world’s largest companies.