cross probing from skill

[Nadine]

Hi all,

I'm trying to find the net/device in the extracted view corresponding
to a schematic object from within a skill function. (Say given the net
name /I5/I7/net31 I want to find the name of the corresponding net in
the extracted view). Is this possible at all? If yes, how? I've
searched the cadence documentation but didn't get further than
geCrossProbeValue...

Thanks,
Nadine

 

[Partha]

Nadine,
If you are using a av_extracted view then this should be automatically
done for you, look up the assura-->MSPS option. You could probe the
whole nets, devices etc from the extracted view and highlight them in
the schematic. Is that what you want?

Partha

Nadine@MailSys.de (Nadine) wrote in message news:<b145778a.0311240436.6632f525@posting.google.com>...

Hi all,

I'm trying to find the net/device in the extracted view corresponding
to a schematic object from within a skill function. (Say given the net
name /I5/I7/net31 I want to find the name of the corresponding net in
the extracted view). Is this possible at all? If yes, how? I've
searched the cadence documentation but didn't get further than
geCrossProbeValue...

Thanks,
Nadine

 

[Nadine]

Hi Partha,

thanks for you answer, but..

cadeguy@yahoo.com (Partha) wrote in message
news:<7f58031f.0311241046.27250ddd@posting.google.com>...

If you are using a av_extracted view then this should be automatically
done for you, look up the assura-->MSPS option. You could probe the
whole nets, devices etc from the extracted view and highlight them in
the schematic. Is that what you want?

no, not exactly. What I want to do is the following:
I wrote an ocean script that plots certain nets from the schematic:
awvPlotWaveform( awvGetCurrentWindow() list getData("/I9/I5/reset") ).
Then I change the simulation settings to prefer extracted views over
schematic ones. At the moment I have to manually look up what
/I9/I5/reset corresponds to in the extracted view and alter the script
(and that after every extraction...). I'd like to have an automatic
way to do the job for me.

Nadine

 

[S. Badel]

i believe it is possible to generate a correspondence file
through LVS. never done it, though...

stephane

"Nadine" <Nadine@MailSys.de> wrote in message
news:b145778a.0311250126.37f07c2f@posting.google.com...

Hi Partha,

thanks for you answer, but..

cadeguy@yahoo.com (Partha) wrote in message
news:<7f58031f.0311241046.27250ddd@posting.google.com>...
If you are using a av_extracted view then this should be automatically
done for you, look up the assura-->MSPS option. You could probe the
whole nets, devices etc from the extracted view and highlight them in
the schematic. Is that what you want?

no, not exactly. What I want to do is the following:
I wrote an ocean script that plots certain nets from the schematic:
awvPlotWaveform( awvGetCurrentWindow() list getData("/I9/I5/reset") ).
Then I change the simulation settings to prefer extracted views over
schematic ones. At the moment I have to manually look up what
/I9/I5/reset corresponds to in the extracted view and alter the script
(and that after every extraction...). I'd like to have an automatic
way to do the job for me.

Nadine

 

[Nadine]

Hi Stephane,

"S. Badel" <stephane.badel@epfl.ch> wrote in message news:<3fc484b4$1@epflnews.epfl.ch>...

i believe it is possible to generate a correspondence file
through LVS. never done it, though...

Yes, I've tried that. The file format (of file xref.out) is trivial
and could easily be consumed by awk-magic:
N0 N0
N1 N1
N2 N719
(apparently "net in schematic" "space" "net in extracted")
followed by a section for the instances.

I failed to match the "N2" to any net in the schematic (or rather the
other way round), however...

Nadine

 

[S. Badel]

this is because LKS creates his own flat netlists. in those netlists, the
nets are named as numbers,
N0, N1 etc... that you have seen. i believe, not sure though, that this is
because when flattening
the netlist, name clashes may occur so it assigns numbers to net to make
sure this doesn't happen.

anyway, the netlist format is described in diva (you're using diva, right?)
reference in the appendix D.

the first part of this netlist is the mapping table of net names to net
numbers. precisely what you are looking for.
for example:
t 3 OUT output
t 4 IN input
n 1 vdd!
n 0 gnd!
n 2 /OUT
n 3 /IN
where, you guessed it, t stands for terminal and n stands for net. same
thing in the extracted netlist.

those netlists can be found in lvs_directory/schematic/netlist and
lvs_directory/layout/netlist.

if you have some scripting skills (regexps will do it, even in SKILL) you
will be able to decipher this.

i'm not sure this is the best way to do this, but it should work... good
luck.
stephane

"Nadine" <Nadine@MailSys.de> wrote in message
news:b145778a.0311260602.5c3a2a68@posting.google.com...

Hi Stephane,

"S. Badel" <stephane.badel@epfl.ch> wrote in message
news:<3fc484b4$1@epflnews.epfl.ch>...
i believe it is possible to generate a correspondence file
through LVS. never done it, though...

Yes, I've tried that. The file format (of file xref.out) is trivial
and could easily be consumed by awk-magic:
N0 N0
N1 N1
N2 N719
(apparently "net in schematic" "space" "net in extracted")
followed by a section for the instances.

I failed to match the "N2" to any net in the schematic (or rather the
other way round), however...

Nadine

 

[Nadine]

Hi Stephane,

"S. Badel" <stephane.badel@epfl.ch> wrote in message news:<3fc5cede$1@epflnews.epfl.ch>...

this is because LKS creates his own flat netlists. in those netlists, the
nets are named as numbers,
N0, N1 etc... that you have seen. i believe, not sure though, that this is
because when flattening
the netlist, name clashes may occur so it assigns numbers to net to make
sure this doesn't happen.

anyway, the netlist format is described in diva (you're using diva, right?)
reference in the appendix D.

Thank you so much. This is exactly what I'm looking for. I was
searching for "N0" in all the files... Turns out they didn't stick to
one notation for the net numbers...

Nadine