InvestorsHub Logo
Followers 21
Posts 14802
Boards Moderated 0
Alias Born 03/17/2003

Re: SemiconEng post# 12439

Friday, 07/09/2004 12:20:12 PM

Friday, July 09, 2004 12:20:12 PM

Post# of 151757
I think the habit of having only a couple of TDP bins
that all speed grades fall into will be the way of the
future. It is possible that yield lossage from defects
is now less important than yield loss from parametrics
being out of wack. That's why we no longer see a nice
curve fitting power and clock rate in data sheets.

In general device speed will have a strong correlation
with leakage. The faster a part goes then higher its
leakage will be tend to be. The "golden" parts are the
ones with high speed and low leakage but they are the
minority IMO. If you set a maximum power then they will
be the fastest parts (or they will get separated out for
operation at lower voltage as mobile parts). What do
you do with the ones that run at top speed grade but
have leakage so high as to put them over the TDP
limit? You bin down their clock rate so the dynamic
power saved makes up for higher leakage power and
this brings them back under the group TDP limit so
they don't have to be thrown away.




Volume:
Day Range:
Bid:
Ask:
Last Trade Time:
Total Trades:
  • 1D
  • 1M
  • 3M
  • 6M
  • 1Y
  • 5Y
Recent INTC News