Intel Corp (INTC)

[wbmw]

Dew, Re: There is no way in hell that TMTA would have gone to such lengths to design a new VLIW and develop a 256-bit version of the CMS for a measly 21% gain in throughput.

Welcome to the world of diminishing returns. When is the last time you saw a 20% performance gain in, say, a Pentium 4? That's more than 3 speed grades, right off the bat. Not bad for a start. Having said that, I do expect more from Astro.

Re: But when you produced a lower bound to your range of 1.21, that said to me that you are just blowing smoke.

Dew, you're smarter than that. Don't insult yourself and me by putting words in my mouth. A lower bound, just like an upper bound, it the least likely of a given range. Take the mean of the range, which is about 1.45, and that is what I think is the most likely launch target for the Astro, once you combine frequency and IPC gains.

Given the law of diminishing returns, I'd say that TMTA ought to be proud of a 45% gain in performance, without moving to a new manufacturing process. The fact that you scoff at this as a "low" score tells me that you are the most likely one blowing smoke to pump the notion that TMTA is a rosier investment than it realistically is. Whether that is your intention or not, you ought to know better.