Ian,
I think the speed of the netlist load is dependant on several things:
1. Number of nodes,  doesn't seem to be related to number of components.
2. where the netlist is located and its last name
3. Number of conflicts  it starts to compare and convolute macros.

For some reason
99SE wants to always spend time searching for a netlist that I either
deleted or trashed.   I wish it would immediately look at the netlist I
pointed to.
Second, all it has to do is reassign the pads to the correct net names and I
can handle the rest by reconnection of  copper and run drcs   it is much
faster.    or change the algorithm to follow my process because I can out
perform theirs.
Third:  I am designing with as many as 15 300 pin BGAs and that many large
QFPs fanned out so I can throw it in Spectra.  I think the lib parts fan out
is slowing down the process but haven't confirmed it yet.   It just gets me
that 98 loaded the netlist (I timed it ) in 3 minutes and 99SE took the
train to Australia and sunk in the Ocean for 6  hours.( twice)


Regards,

Mike Reagan
EDSI
Frederick MD







----- Original Message -----
From: Ian Wilson <[EMAIL PROTECTED]>
To: Protel EDA Forum <[EMAIL PROTECTED]>
Sent: Thursday, March 15, 2001 6:32 PM
Subject: Re: [PEDA] Netlist will not load


> On 05:30 PM 15/03/2001 -0500, Michael Reagan said:
> >Hello all,
> >I originally posted this message/complaint when 99  first was released
and
> >thought the problems have since been fixed.  On very large designs, mine
was
> >18,000 nets, 99SE with SP6 installed took 6 hours of 600MHZ  computer
time
> >then hung up.The machine has 512 meg of ram.  I wasted about 12 hours of
my
> >time doing a simple netlist load. My netlist and design are not
corrupted, I
> >loaded it on 98 SP3 and it loaded in less than 3 minutes.  Something is
> >really wrong with the way Protel is still loading nets.  This in my best
> >guess is a simple task. Luckily for me I didn't throw away 98.  I am
really
> >posting this for Protel to see, because I am sure this problem is beyond
> >tech supports functionality.
> >
> >
> >Oh well keep 98 for while still, I hate that.
> >
> >Mike Reagan
> >EDSI
>
> I agree that netlist load is glacial especially if you are doing a major
> update and have checked the update connected copper option in the
> synchroniser. The synchroniser is more powerful than the old netlist load
> (in that designators can be changed willy-nilly before updating) but it
> does seem very slow.  Is the netlist load also slow?
>
> I saw quite a moderate netlist take 5 minutes to load on a moderately
> capable machine.  I wonder if we could try to track down some work arounds
> for this or at least determine what the dependencies are:
> 1) Does it depend on the number and/or size of the PCB libraries loaded in
PCB?
> 2) If I had a big design made only of resistors and a PCB library with
just
> the one component a resistor) loaded would it be faster?
> 3) Is there a significant load time difference between the synchroniser
and
> the netlist load?
> 4) How do the options in the synchroniser affect load times?
> 5) Does the netlist format make any difference? (Protel/Protel2)
> 6) Is it linear with size of design?  If I have a design with half the
> components and the same number of nets does it take half as long? Half the
> nets and the same number of components? Half the nets and half the
components?
> 7) Any other ideas?
>
> This is the sort of stuff that Protel should do - may be we could get
> Protel tech support to undertake this testing and reply to us. Come on CSC
> can you do this work and tell us what, if anything, we can do to improve
> netlist load times.
>
> Ian Wilson
>

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
* To post a message: mailto:[EMAIL PROTECTED]
*
* To join or leave this list visit:
* http://www.techservinc.com/protelusers/subscrib.html
*                      - or email -
* mailto:[EMAIL PROTECTED]?body=leave%20proteledaforum
*
* Contact the list manager:
* mailto:[EMAIL PROTECTED]
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Reply via email to