Constantine, Re: "when OEMS find Athlon errata and report them to AMD, AMD buries them?"
OEMs usually do compatibility validation (which involves off the shelf software), which is far less likely to find bugs than system or design validation done by the actual CPU manufacturer. Of course, compatibility validation can often find more obscure bugs, but few applications out there can stress a system as much as specifically designed tests that Intel or AMD can write.
What I don't understand is why you are trying to accuse me of accusing AMD of burying bugs that the OEMs find. I have never suggested anything along those lines. What you need to consider is which OEMs are actively involved in finding bugs for AMD, and which of them has a methodology for finding bugs with the highest quality? If AMD does not have many OEMs actively finding bugs, or if their methodology doesn't lend itself to finding additional bugs, then there would be nothing to add to the errata lists. In other words, you might be barking up the wrong tree.
wbmw
P.S. Regarding Scumbria, I agree with his response. Cell based designs are likely to speed up delivery to market, and perhaps fix bugs faster. However, he never said that cell based designs create less bugs, which is what Combjelly suggested. |
