Replies to post #2754 on AMD strictly moderated (AMD)

[chipdesigner]

when really it's just AMD's old 90nm process with some scotch tape on it.

LOL!!!

If you want to call somewhere in the region of 13-25% increased performance over the current DSL strain, at the same power, "scotch tape", well, you go right ahead.

To be conservative, let's assume that by "performance" the process guys really mean frequency.

And let's take the middle of the range: 19%.

That 2.6GHz DC part will run at 3.1GHz DC at the same power with SiGe strain.

Where did I get 13-25% ?

DSL was quoted variously as achieving 20-24% better frequency("performance") than no strain at all at the same power.

SiGe was quoted variously as achieving 40-50% better frequency ("performance") than no strain at all at the same power.

Best case: 1.5 / 1.2 = 1.25 = 25% better
Worst case: 1.4 / 1.24 = 1.13 = 13% better

That "scotch tape" of yours may well erase Conroe's supposed advantages.