Save every run's psf data in MonteCarlo analysis

X

Xuejun Zhang

Guest
Hello,

I am running MonteCarlo analysis in Cadence IC5032. The simulator I
use is hspiceS. The MonteCarlo simulation run well, but the simulation
result seems to be erased after each run. So is there some way I can
save the psf result of each run?

The second question:
In hspice you can just use the analysis statement to do the monte
carlo simulation, for example
..TRAN 1.0e-12 1.0e-8 START=0.0 MONTE=30
where Monte=30 means the hspice will run monte carlo simulation for 30
times.
So I manually edit analysis statement in the final netlist to the
example show above, and then manually run the "runHspice" in the
netlist directory. The hspice simulation run well, the psf data,
however, can only generate one waveform instead of 30 waveforms in
Analog DE. When I look into the size of the psf data, it do have size
of 30 times bigger than the non-montecarlo simulation, which indicates
the simulation results of all 30 monte carlo run are stored in the psf
file. But the analog DE could only view one of them.

Therefore, the question is how can view all waveforms of each monte
carlo run from the psf data?
 
Xuejun,

Monte-Carlo via the hspiceS interface does not use hspice's built-in monte-carlo
analysis, but instead uses cdsSpice as the random number generator (mostly for
historical reasons).

So adding parameters onto the tran line won't help you.

I don't believe it is possible to save the waveforms between runs with the
hspiceS monte-carlo interface. With the "spectre" interface, it uses the native
monte-carlo in the simulator, and that does allow waveforms to be saved (there's
a checkbox on the monte carlo form for doing this).

Andrew.

On 18 Jun 2004 07:32:02 -0700, xuejun_zhang2004@hotmail.com (Xuejun Zhang)
wrote:

Hello,

I am running MonteCarlo analysis in Cadence IC5032. The simulator I
use is hspiceS. The MonteCarlo simulation run well, but the simulation
result seems to be erased after each run. So is there some way I can
save the psf result of each run?

The second question:
In hspice you can just use the analysis statement to do the monte
carlo simulation, for example
.TRAN 1.0e-12 1.0e-8 START=0.0 MONTE=30
where Monte=30 means the hspice will run monte carlo simulation for 30
times.
So I manually edit analysis statement in the final netlist to the
example show above, and then manually run the "runHspice" in the
netlist directory. The hspice simulation run well, the psf data,
however, can only generate one waveform instead of 30 waveforms in
Analog DE. When I look into the size of the psf data, it do have size
of 30 times bigger than the non-montecarlo simulation, which indicates
the simulation results of all 30 monte carlo run are stored in the psf
file. But the analog DE could only view one of them.

Therefore, the question is how can view all waveforms of each monte
carlo run from the psf data?
Click to expand...
--
Andrew Beckett
Senior Technical Leader
Custom IC Solutions
Cadence Design Systems Ltd
 
Thanks for your help.
Regards
xuejun

Andrew Beckett <andrewb@DELETETHISBITcadence.com> wrote in message news:<m32id0llf736s4b81fj5o33o8ae94ra1lc@4ax.com>...
Xuejun,

Monte-Carlo via the hspiceS interface does not use hspice's built-in monte-carlo
analysis, but instead uses cdsSpice as the random number generator (mostly for
historical reasons).

So adding parameters onto the tran line won't help you.

I don't believe it is possible to save the waveforms between runs with the
hspiceS monte-carlo interface. With the "spectre" interface, it uses the native
monte-carlo in the simulator, and that does allow waveforms to be saved (there's
a checkbox on the monte carlo form for doing this).

Andrew.

On 18 Jun 2004 07:32:02 -0700, xuejun_zhang2004@hotmail.com (Xuejun Zhang)
wrote:

Hello,

I am running MonteCarlo analysis in Cadence IC5032. The simulator I
use is hspiceS. The MonteCarlo simulation run well, but the simulation
result seems to be erased after each run. So is there some way I can
save the psf result of each run?

The second question:
In hspice you can just use the analysis statement to do the monte
carlo simulation, for example
.TRAN 1.0e-12 1.0e-8 START=0.0 MONTE=30
where Monte=30 means the hspice will run monte carlo simulation for 30
times.
So I manually edit analysis statement in the final netlist to the
example show above, and then manually run the "runHspice" in the
netlist directory. The hspice simulation run well, the psf data,
however, can only generate one waveform instead of 30 waveforms in
Analog DE. When I look into the size of the psf data, it do have size
of 30 times bigger than the non-montecarlo simulation, which indicates
the simulation results of all 30 monte carlo run are stored in the psf
file. But the analog DE could only view one of them.

Therefore, the question is how can view all waveforms of each monte
carlo run from the psf data?
Click to expand...
 
You must log in or register to reply here.

Welcome to EDABoard.com

Sponsor

Back
Top