Candence vituoso IC 6.1 doubt

[varunmjman]

hi...

I am using Cadence Virtuoso IC 6.1 for my analog design..
For a transient analysis of 6.6 ms my circuit taked 14hrs to finish the simulation. Is there any options available in the tool which will help bring down the simualtion time....


thanks in advance...

 

[erikl]

For a transient analysis of 6.6 ms my circuit taked 14hrs to finish the simulation.

CPU time or just total time spent? Did you check your license time share?

 

[varunmjman]

Its the CPU time..

At the end of the log file it says

intrinsic transient analysis time : CPU =50.5ks (14hrs)

I am new in this field...how can i check my license time share??

 

[erikl]

Re: Cadence virtuoso IC 6.1 doubt

CPU =50.5ks (14hrs)

A big circuit with postLayout backannotation? Then this could be possible.
What's your min. time step?

how can i check my license time share??

On Unix or Linux systems, you can check if you just own a license:
% lmstat -a -c $CDS_LIC_FILE
$CDS_LIC_FILE is the std variable for your Cadence license file. If not available, you should give the full path name.

 

[prashantbabu]

If you are running your simulations on a multi processor server, you can also use the multi threading option. In the ADE window, in Simulation->Options->Analog, you will find a multi-thread option.
You can turn it "ON" and specify the number of threads..the no.of threads will be the number of processors on the server that your simulation will use.
This will make the simulations quite fast. Do check with your sysadmin about the number of CPU s available on your server.
Hope it helps!

Added after 3 minutes:

One more point to consider is the nodes that you save for the analysis. If you are running a long transient simulation, and you dont need to check the results on all the nets of your circuit, then you can specify the nets to be saved rather than the "save all" option. This will greatly reduce the simulation time,
Also, a point to note is that if you are running post-layout simulations, generally "moderate" rather than "conservative" tolerances are used.