FreeLists / si-list / [SI-LIST] Re: Getting Hspice to accept numbers with mantissa's larger than 8 characters

["Taha Amiralli" <thamiral@xxxxxxxxx>]

Thanks for the response,

Your analysis of the validity of my question is absolutely correct.

My reasons for trying to get hspice to accept a double precision
constant is actually to try and debug an algorithm that I am
implementing.

This algorithm has two transformation matrices (that are multiplied
toghether)  and I have noted that when I implement only one of these
transformation matrices, I get a correct simulation.

The problems arise when I try to implement the product of these two
transformation matrices and these problems range from an incorrect
simulation to a 'time step too small' error. I have ruled out the
obvious errors (i.e. the matrices are incorrect) by prototyping this
algorithm in circuit simulator that I have written.

Given all this, I was hoping that the source of the problem was that I
am giving hspice single precision numbers whilst my 'homemade' circuit
simulator is fed double precision numbers.

By the way, Synopsys has switched to double precision numbers in the
simulation engine, however, it appears that they only read in single
precision numbers from the input spice file by default.

Thanks once again,

Sincerely,

-------------------
Taha Amiralli
thamiral [A] uwo [D] ca
thamiral [A] gmail [D] com

MESc Candidate 2007, Computer Engineering
The University Of Western Ontario

BESc, BSc. 2005,
Computer Engineering & Computer Science
The University Of Western Ontario
---------------------------------------------------------------




On 4/5/06, Muranyi, Arpad <arpad.muranyi@xxxxxxxxx> wrote:
> Taha,
>
> As far as I know they changed to double precision in the nineties,
> but this would be best answered by a Synopsys representative...
>
> However, regarding the validity of your question, I wonder what
> is the value of having that many digits as input to an iterative
> solver where the solutions are inherently not exact?  A lot will
> depend on the tolerance settings, but even then, the solutions=3D20
> will never be exact.
>
> Once I wanted to generate a zero output and wrote circuit with a
> resistor in parallel with a zero volt voltage source, and printed
> its current.  The numbers were anything but an exact zero.
>
> Add to this the problem of converting floating point decimal
> numbers to binary and back, which is what almost all computers
> do for us since most of them use binary floating point units.
> (There are decimal floating point units too which do not suffer
> from these problems).
>
> Depending on the precision you use (single, double, etc) sooner
> or later you will start getting digits that make no sense for
> even the simplest calculations.  For example, I tried 10/3 in
> one programming tool.  Using single precision, I got:
>   3.333333253860473600000000000000E+0 and with double precision:
>   3.333333333333333500000000000000E+0.
> We know that neither of these two are correct.  Depending on which
> programming software you use, some times an even more basic
> calculation such as 10.0/1.0 (in not integer representation)
> will not result in 10.000000000000000000000000.
>
> So what are you trying to achieve with so many digits?
>
> Arpad Muranyi
> Intel Corporation
> -----------------------------------------------------------------
>
> -----Original Message-----
> From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] =
=3D
> On Behalf Of Taha Amiralli
> Sent: Wednesday, April 05, 2006 6:37 AM
> To: Fred Balistreri
> Cc: si-list@xxxxxxxxxxxxx
> Subject: [SI-LIST] Re: Getting Hspice to accept numbers with mantissa's =
=3D
> larger than 8 characters
>
> Thanks for the reply Fred,
>
> I have looked up the option that you had recommeneded (i.e. NUMDGT)
> and the hspice manual mentions that this option is used to specify the
> "Number of significant digits to print, for output variable values."
>
> My concern is more along the lines of inputs since, from what I
> understand, hspice uses single precesion floating point
> representations for input values and hence, it only accepts 6-7 digts
> of a mantissa.
>
> Sincerely,
>
> -------------------
> Taha Amiralli
> thamiral [A] uwo [D] ca
> thamiral [A] gmail [D] com
>
> MESc Candidate 2007, Computer Engineering
> The University Of Western Ontario
>
> BESc, BSc. 2005,
> Computer Engineering & Computer Science
> The University Of Western Ontario
> ---------------------------------------------------------------
> ------------------------------------------------------------------
> To unsubscribe from si-list:
> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>
> or to administer your membership from a web page, go to:
> http://www.freelists.org/webpage/si-list
>
> For help:
> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>
> List FAQ wiki page is located at:
>                 http://si-list.org/wiki/wiki.pl?Si-List_FAQ
>
> List technical documents are available at:
>                 http://www.si-list.org
>
> List archives are viewable at:
>                 http://www.freelists.org/archives/si-list
> or at our remote archives:
>                 http://groups.yahoo.com/group/si-list/messages
> Old (prior to June 6, 2001) list archives are viewable at:
>                 http://www.qsl.net/wb6tpu
>
>
>


--
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field

or to administer your membership from a web page, go to:
http://www.freelists.org/webpage/si-list

For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field

List FAQ wiki page is located at:
                http://si-list.org/wiki/wiki.pl?Si-List_FAQ

List technical documents are available at:
                http://www.si-list.org

List archives are viewable at:     
                http://www.freelists.org/archives/si-list
or at our remote archives:
                http://groups.yahoo.com/group/si-list/messages
Old (prior to June 6, 2001) list archives are viewable at:
                http://www.qsl.net/wb6tpu