Mysterious Design Compiler crash...
marnix arnold
synthesising a microprocessor design from VHDL. Everything goes well,
analysing and elaborating the code and setting the constraints. Now when
I optimise for area (compile, medium effort), no problems occur. However,
when I optimise for timing (using a clock on the top-level of the design),
the thing compiles allright, but when the cells are written into the
design library, this happens:
Fatal: Internal system error, cannot recover.
Release = 'v3.2a' Architecture = 'hp700' Program = 'design_analyzer'
And the program dies, all windows disappear. Now, I may be doing something
wrong, but this shouldn't be allowed to happen in any case. Has anyone
else encountered this bug before, and if so, is there a workaround?
By the way, this isn't some incidental crash, it happens again and again
for this design. The funny thing is, if I first optimise for area and
then for timing, nothing goes wrong...
Marnix.
--
I Marnix Arnold <M.Ar...@et.tudelft.nl> I
I Faculty of Electrical Engineering, Delft University of Technology I
I "I think there's a world market for about five computers." I
I - Thomas J Watson, Chairman of the Board, IBM. I
John Walton
|> Something very weird just happened. I'm using the ES2 design kit for Synopsys,
|> synthesising a microprocessor design from VHDL. Everything goes well,
|> analysing and elaborating the code and setting the constraints. Now when
|> I optimise for area (compile, medium effort), no problems occur. However,
|> when I optimise for timing (using a clock on the top-level of the design),
|> the thing compiles allright, but when the cells are written into the
|> design library, this happens:
|>
|> Fatal: Internal system error, cannot recover.
|>
|> Release = 'v3.2a' Architecture = 'hp700' Program = 'design_analyzer'
|>
|> And the program dies, all windows disappear. Now, I may be doing something
|> wrong, but this shouldn't be allowed to happen in any case. Has anyone
|> else encountered this bug before, and if so, is there a workaround?
I've been 'blessed' with those on every release I've used. Of course, what I was doing
to make it crash was different, in every case. There should have been
a long string of numbers you can feed to the Synopsys Hotline.
marnix arnold
[about Design Compiler's internal system errors]
> I've been 'blessed' with those on every release I've used. Of course, what I was doing
> to make it crash was different, in every case. There should have been
> a long string of numbers you can feed to the Synopsys Hotline.
I agree, there should have been one. But there isn't. Not anywhere, not on
the screen (stderr), not in the view_command.log.
Perhaps that's another bug...
Jon Connell
> John Walton (wal...@emc.com) wrote:
>
> [about Design Compiler's internal system errors]
>
> > I've been 'blessed' with those on every release I've used. Of course, what I was doing
> > to make it crash was different, in every case. There should have been
> > a long string of numbers you can feed to the Synopsys Hotline.
>
> I agree, there should have been one. But there isn't. Not anywhere, not on
> the screen (stderr), not in the view_command.log.
>
> Perhaps that's another bug...
>
Synopsys only does this on the sparc SunOS 4.1.x platform. The .synopsys_dc.setup
file in the Synopsys tree has the following:
/* Enable debug output on fatal */
if (arch == "sparc") {
set_unix_variable(SYNOPSYS_TRACE, "")
}
Maybe you could try setting $SYNOPSYS_TRACE in the shell beforehand. I've not
tried it on any platform other than SunOS 4.1.x, but I'd like to hear if it
works on other platforms.
--
Jon Connell
E-Mail : alb...@scn.de
WWW : http://www.scn.de/~alberon _/ _/_/_/ _/_/ _/_/_/
Fax : +49 (0)89 722 27544 _/ _/ _/ _/ _/
Telephone : +49 (0)89 722 28790 _/ _/ _/ _/ _/
Snail-Mail : Siemens AG, ÖN TN ET D33 _/ _/_/_/ _/_/_/ _/_/_/
Postfach 700071
D-81359 Munich SIEMENS IC DESIGN CENTER