Spectre internal error

[Suzy Jackson]

Hi,

I'm not very experienced with cadence, but have been designing RF-CMOS
circuitry with it for the last few months.

I've successfully simulated and optimised a number of designs;
primarily simple LNA circuits.

I'm currently trying to optimise an LNA design based on a recent paper
by Bruccoleri et. al, and keep running into problems.

The circuit simulates fine using analog artist, but when I try to run
the optimizer on it, it runs through iteration 0 okay, then during
iteration 1 cds.log reports:

*WARNING* gpwipx: Too many stripts for window (nA=8 nD=0)
*Error* Y-direction window size is too small
(repeated eight times)

the spectre.out window reports:

Internal error found in spectre. Please run 'getSpectreFiles'
etc...
Error detected in file 'ipsm.c' at line 212
Assertion failed.

My circuit is simply a pair of nmos fets (from the IBM 0.18u library),
two resistors, a capacitor, a couple of sources, and a couple of
ports. Does anyone have any idea what's happening?

Regards,

Suzy

 

[Andrew Beckett]

I found a couple of PCRs related to this error. I believe from these that this
was fixed in 5.0.33.500.3 - so you probably need a later hotfix.

The too many strips is just an artifact of the waveform plotting - it's not
related to the spectre internal error.

Regards,

Andrew.

On 14 Apr 2004 17:26:34 -0700, Suzy.Jackson@csiro.au (Suzy Jackson) wrote:

Hi,

I'm not very experienced with cadence, but have been designing RF-CMOS
circuitry with it for the last few months.

I've successfully simulated and optimised a number of designs;
primarily simple LNA circuits.

I'm currently trying to optimise an LNA design based on a recent paper
by Bruccoleri et. al, and keep running into problems.

The circuit simulates fine using analog artist, but when I try to run
the optimizer on it, it runs through iteration 0 okay, then during
iteration 1 cds.log reports:

*WARNING* gpwipx: Too many stripts for window (nA=8 nD=0)
*Error* Y-direction window size is too small
(repeated eight times)

the spectre.out window reports:

Internal error found in spectre. Please run 'getSpectreFiles'
etc...
Error detected in file 'ipsm.c' at line 212
Assertion failed.

My circuit is simply a pair of nmos fets (from the IBM 0.18u library),
two resistors, a capacitor, a couple of sources, and a couple of
ports. Does anyone have any idea what's happening?

Regards,

Suzy

--
Andrew Beckett
Senior Technical Leader
Custom IC Solutions
Cadence Design Systems Ltd