|
|
Magic Mailing List |
|
From: micke (micke AT concentric DOT net) Date: Wed Jul 05 2000 - 07:24:18 EDT
Hello,
I'm Mikael Sahrling and I work for Semtech Corp. We use magic actively
for our
layout and in particular we use it as a fast verification tool. I've
modified the 6.5 distribution to suit our needs better and have found
what
appears to be bugs in that code. I'd love to be a part of the
development team.
Unfortunately some of my co-workers prefer other layout tools for doing
the
actual layout and they use only magic as a verification tool and
recently
we ran into problems.
We use magic for a modern BiCMOS process, but there seems to be a
problem extracting the capacitance for wires that in the calma GDSII
stream file is at 45 degress angle. The perimcap calculation comes out
wrong. It seems magic is calculating the circumference around each
little
rectangle and ignores the fact that most of the rectangle contacts other
rectangles.
<- L ->
XXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXX -> circumference for this segment is
XXXXXXXXXXXXXXXXXXXXXXXX 5 X (2*L+2*segment height). It
XXXXXXXXXXXXXXXXXXXXXXXX should be 2*L+5X2*segment height.
Any known remedies?
It appears to be problem in the ExtBasic.c-code, function
"extNodeAreaFunc"
the calculation of the perim cap "reg->nreg_cap += capval * len;" uses a
wrong value of len???
Mikael Sahrling
|
|
|
|