input is of the form:
Conductor number, (x,y) lower left, (x,y) upper right, penetraiont depth, 
x divisions, y divisions, x ratio, y ratio,x type, and y type 

x/y type is defined as follows:
           type 0 small on ends
           type 1 small on right (upper)
           type 2 small on left (lower)
           type 3 small in middle

induct is slightly slower than its counterpart, inductance, but allows
unlimitted conductors and subdivisions. 
(inductance = 1.2 x faster than induct)

short benchmark info: speeds of different machines, same task
(diva = 2 x jay = 2.8 x octopus)
(==> jay (with induct) = 2.47 x slower than inductance on diva)
(                        2.37 x faster than inductance on octopus)

finally, note that these are CPU times, and real times are more
favorable to my machine (low load)  (in this example, induct was FASTER
than inductance on diva real time 11 vrs 15.  And, in fact, my machine 
BEATS sometimes.


*********************************************************************
Some notes on accuracy: induct (inductance) / sline / TVD's formula

TVD's formula is only good for special case when kappa is as given
in the graph, sline is only good for micro strip, induct is all.

TVD's L = UoK d/w (1 + l1/d coth(b1/l1) + l2/d coth(b2/l2) )
      L = UoK d/w (1 + l1/d + l2/d ) for b >> lambda (l)
      L = UoK d/w  for l << d and l << b

to make TVD's and sline agree, kappa (for 10/1 w/d ratio) is
kappa in TVD		w/d,lambda,bline,bgnd
0.75468			10/1.0, 0.06, 0.25 0.2
0.82491			 5/0.5, 0.06, 0.25 0.2
0.77087			 1/0.1, 1E-6, 1E-4 1E-4  (L = UoK d/w)

*********************************************************************
for an isolated line mb2, 

100 -10.0 -0.2 10.0 0.0 0.06 30 8 1.5 1.5 3 1
1   -2.50 0.45 2.50 0.7 0.06 10 8 1.5 1.5 0 2
===***********Inductance matrix***********===
0.107902

or with even spaces on GND, 
100 -10.0 -0.2 10.0 0.0 0.06 30 8 1.0 1.5 3 1
1   -2.50 0.45 2.50 0.7 0.06 10 8 1.5 1.5 0 2
===***********Inductance matrix***********===
0.107723

100 -10.0 -0.2 -2.5 0.0 0.06 20 8 1.5 1.5 1 1
100  -2.5 -0.2  2.5 0.0 0.06 10 8 1.5 1.5 0 1
100   2.5 -0.2 10.0 0.0 0.06 20 8 1.5 1.5 2 1
1   -2.50 0.45 2.50 0.7 0.06 10 8 1.5 1.5 0 2
===***********Inductance matrix***********===
0.107687

sline(5 .45 .25 .2 .06 .06) = 0.108904 pH/um

*********************************************************************
for coupled lines, mcmb, mcmb2

100 -10.0 -0.2  10.0 0.0 0.06 15  5 1.0 1.5 1 1
1   -2.50 0.45  2.50 0.7 0.06 10  5 1.5 1.5 0 0
2   -1.00 1.20  1.00 1.9 0.06  8  5 1.5 1.5 0 2
===***********Inductance matrix***********===
0.107880        0.090009
0.090009        0.279914

100 -10.0 -0.2  -2.5 0.0 0.06 20  8 1.5 1.5 1 1
100  -2.5 -0.2   2.5 0.0 0.06 10  8 1.5 1.5 0 1
100   2.5 -0.2  10.0 0.0 0.06 20  8 1.5 1.5 2 1
1   -2.50 0.45 -1.00 0.7 0.06  5 10 1.5 1.5 0 0
1   -1.00 0.45  1.00 0.7 0.06  8 10 1.5 1.5 0 0
1    1.00 0.45  2.50 0.7 0.06  5 10 1.5 1.5 0 0
2   -1.00 1.20  1.00 1.9 0.06  8  8 1.5 1.5 0 2
===***********Inductance matrix***********===
0.107201        0.089521
0.089521        0.278822

sline(5  .45 .25 .2  .06 .06) = 0.108904 pH/um    (of mb)
sline(2 1.2  .7  .2  .06 .06) = 0.313365 pH/um    (of mc)
sline(2 0.95 .7  .2  .06 .06) = 0.281462 pH/um    (of mc making mb "invisable")
sline(2  .5  .7  .25 .06 .06) = 0.208400 pH/um    (of mc with mb as a gnd plane)

*********************************************************************
*********************************************************************
TVD's formula is ok, but the "constant" kappa can cause significant error

===== MB    (sline 1% higher than induct)
sline seems to be consistantly higher than induct.  Trying to increase
the accuracy makes the inductance go down in induct, while sline's value
is actually higher (0.10890/0.10769 => sline is 1.01x higher than induct)

===== MB and MC
== MC      (sline 1.01x higher than induct)
for multiple conductors, you must assume the middle inductor is "invisible"
using the normal GND, but REDUCING the height by the thickness of the middle
(mb) conductor.  This gives, again, a value of self inductance of the
top line of 0.95% higher than induct.

== MB      (sline 1.016x higher than induct)
The multiple lines further reduce the self inductance of mb, increasing 
the difference between induct and sline to 1.59% higher than induct.
This may be only due to the larger subdivision in this case than the previous.

===== induct (very stable even with poor subdivision
comparing many segments to fewer: 

====== single MB 
== MB    0.2% reduction in self inductance with higher accuracy
increasing accuracy reduced inductance (240 GND 80 MB to 400 GND 80 MB)


====== multiple conductors

change in accuracy was from 
     (75 GND, 50 MB, 40 MC) to (400 GND, 180 MB, 64 MC)

== MB    0.63% reduction in self inductance with higher accuracy
increasing accuracy reduced inductance

== MC    0.39% reduction in self inductance with higher accuracy
increasing accuracy reduced inductance

== coupling --- M = 0.83*L(MB)  0.1% increase with accuracy
increasing accuracy increased coupling factor 
finally, there seems to be no way to get M from sline, but from induct
M appears to be related to the lower line (mb) by 
M = 0.834343 to ("more accurate") 0.835076 (VERY close, 0.1%)  
This does depend on penetration depth and geometry

