custom montecarlo distribution

[F ogh]

Is there a way to define custom distributions to be used in a
statistics block of a spectre model ?
I know I could use a variable x with a flat distribution and then
build any distribution for y=f(x) by calculating the right f, but I
wonder if there is stg more straightforward. I am particularly
interested in the distribution (dirac(x-1)+dirac(x+1))/2

 

[Andrew Beckett]

No, you have to do it the way you suggested.

Andrew.

On Sun, 29 Feb 2004 19:04:13 +0100, F ogh <adff@xs4all.removethis.nl> wrote:

Is there a way to define custom distributions to be used in a
statistics block of a spectre model ?
I know I could use a variable x with a flat distribution and then
build any distribution for y=f(x) by calculating the right f, but I
wonder if there is stg more straightforward. I am particularly
interested in the distribution (dirac(x-1)+dirac(x+1))/2

--
Andrew Beckett
Senior Technical Leader
Custom IC Solutions
Cadence Design Systems Ltd