InvestorsHub Logo
icon url

mas

10/13/04 5:03 PM

#13564 RE: wbmw #13562

I chose to ignore it as I don't believe the variation would be significant in Prescott's case due to how large the mean is. Yeah Northwood did 35W at 2.5 Ghz, however Prescotts do that at no clock speed even at 1.25V, that's how bad the leakage is which gives you about 45W to play with between 2 processors for dynamic power. Prescott starts hauling in Northwood because of its lower dynamic power but the leakage is the starting base. Maybe, we will see, but it sure will be on the ragged edge of binsplit possible at the low voltages required to keep the thermal power within limits. Remind me, what are the lowest power mobile Prescotts doing these days ?
(edit: 88W it seems)







icon url

mas

10/13/04 5:25 PM

#13565 RE: wbmw #13562

We have some inkling from the mobile datasheet.

ftp://download.intel.com/design/mobile/datashts/30242402.pdf

p20

1.86 Ghz, Icc of 44A vs the normal 80A at Vid of 1.15V.
Vcc is 1.15 - (44 * 1.45/1000) = 1.0862 V
therefore power = Icc * Vcc = 48W at 1.86 Ghz.

Yep, it's going to be tight ;-).