Re: Intel was dabbling with SMP way back in the classic Pentium days...
And Intel didn't ship any SMP chipsets until they'd had 6 months to copy Micron's SMP chipset, and Intel's early SMP chipsets were flaky (but Intel doesn't mind shipping flaky products, while AMD fixes them prior to shipment).
I had personal experience with the flaky part regarding Intel's early SMP chipsets. I never used Micron's, so I don't know if Intel copied a Micron problem or invented their own.
When it worked, Intel's chipset outperformed Micron's.
I work with a lot of early release X86 hardware and software, and AMD hardware, even in its first release, has been an order of magnitude more reliable than Intel's first release products traditionally have been.
The best example of that continuing tradition is that all P4 systems shipped so far have a known, intermittent problem with some PCI cards, and Intel's response is to ignore it and fix it in a later release of the product. The closest thing AMD has run into lately is the 100/200 limit on early 760 chipsets and these are being sold as 100/200 chipsets - not as 133/266. And there are no other compatibility problems or limitations to creep up over the years. It is very possible that SMP chipsets at AMD are already better than many products Intel shipped and is shipping, but that they still doesn't meet AMD's quality and reliability standards.
In all seriousness and not as rhetoric, I don't trust Intel products anymore, where once I championed their early adoption.
Dan |
