Yousef, re:<"inaccurate" Q3 predictions and "wrong" info on poor yield analysis>
Oh really, look at techstocks.com
Actual 3Q loss was (0.223), my Q3 prediction made after AMD's latest guidance was (0.23), an error of 7 tenths of one cent.
Actual number of K6's sold was 1,000,000, exactly right
Actual ASP was $150, 12% below my estimate. I did not take into account OEM discounts like I did in the recent Q4 estimate at
techstocks.com
"wrong" information on poor yield analysis
I speculated that drop in yields was due to a mask change, possibly to correct errata documented in the AMD K6 Revision Guide. I recently received confirmation through a third party that it indeed was a mask change that screwed up the speed yields. And at about the same time, the so-called "Linux bug" disappeared.
poor IC process understanding
Whether I know which layers of the chip have silicon and which don't is not relevant to the economics of chip manufacturing. As a systems engineer, I have nothing to prove in that area, but much to learn. The fact that you could not figure out a way to configure two independent production lines when you have dozens of steppers and supporting equipment available and a large Fab area is more significant, IMHO.
If you wish to follow the latter point, look at
techstocks.com and follow the links back to Yousef's original statement.
Petz |
