T.R | Title | User | Personal Name | Date | Lines |
---|
653.1 | | YIELD::HARRIS | | Fri Feb 21 1997 15:51 | 7 |
| We do this all the time. Oversimplifying it, adjust voltage up to
increase speed (and power consumption), down to decrease power
consumption (and speed). EV56 was designed to run around 2.2v, but we
are running it around 2.5v to get the speed we need.
-Bruce
|
653.2 | | WRKSYS::INGRAHAM | Andy | Fri Feb 21 1997 16:43 | 15 |
| ... except that with Alphas, thus far, the voltage setting is only a
function of the system it's put into.
With the EV56, there had been plans to specify two target voltage
ranges for the "core" supply, depending on whether it was a Pass 1 or
Pass 2 EV56, the first using "High Vdd" transistors and the other
"Enhanced" transistors with more speed but a little less voltage
capability. A pin on the EV56 was dedicated for telling the system
which voltage setting to use.
Then the process engineers determined that the "Enhanced" transistors
could withstand the same slightly higher core voltage as the "High
Vdd" transistors, so the need for adjusting the voltage based on this
pin was dropped.
|
653.3 | Maxim, LT, etc are on this bandwagon.. | TEKVAX::KOPEC | Consider a spherical chicken; .. | Fri Feb 21 1997 18:37 | 8 |
| I've been seeing a lot of stuff in the trade rags and in my mail from
various chip vendors about this.. clearly, there is a lot of activity
in the power semi market to support this..
Should DS want to do this, we should try to leverage these commodity
power supply parts.. at least by picking some compatible encoding..
...tom
|
653.4 | | CAMPY::ADEY | Is there a 'Life for Dummies'? | Mon Feb 24 1997 12:02 | 5 |
| Isn't this variable voltage mechanism really part of the patented
interface Intel is employing for the Klamath to put the clone makers
(AMD, IBM/Cyrix) out of business?
Ken....
|