Graphviz Issue Tracker
Mantis Bug Tracker

View Issue Details Jump to Notes ] Issue History ] Print ]
IDProjectCategoryView StatusDate SubmittedLast Update
0001613graphvizDotpublic2009-06-02 21:342011-04-28 04:03
ReporterDavid Cohen 
Assigned Togviz 
PrioritynormalSeveritymajorReproducibilityalways
StatusacknowledgedResolutionopen 
PlatformOS*-*-*OS Version
Summary0001613: No Output File Produced
Description



Command line: dot.exe -Tsvgz -O -Kdot gvbackup.out -v >> dot.out
(PATH="C:\Program Files\Graphviz2.23\bin\";...)



Both with the GUI and the command line I receive no output file and no error.
The input file is quite large.



The last stdout output is:
<CD>
mincross A: 751376 crossings, 1239.56 secs.
network simplex: 598956 nodes 900182 edges maxiter=2147483647 balance=2
</CD>



From the GUI the console shows:
<CD>
xecuting->C:\Program Files\Graphviz2.23\bin\\dot.exe -Tsvgz -o"C:\Documents and Settings\cohend3\Desktop\PERF\backup\fdp_drawing_from_backup_out_type_backup_iterations_2\gvbackup.svgz" -Kdot
dot has created C:\Documents and Settings\cohend3\Desktop\PERF\backup\fdp_drawing_from_backup_out_type_backup_iterations_2\gvbackup.svgz sucessfully.
</CD>



But the file does not exist.



Any Ideas?



Thanks,
Dave Cohen
Additional Information

[erg] Given the trace output, dot got into the node positioning phase and never finished.
With 598956 nodes and 900182 edges, the optimization problem will take a long time.
Since your report seems to imply that the program "finished", I can only assume that
Windows has some resource limits, such as CPU time, dot exceeded them and got kicked out.
(At least, this would be the case on Unix systems.) On the other hand, I would have
expected an error message, or a bad return value so that GVedit would not report success.

[david] Nothing in the Windows event log. In watching the Windows Task Manager I
can tell you that during the run, dot.exe uses 50% of the CPU for about
a half an hour until the memory usage climbs to about 500,000 K. I am
trying to render in dot output, and then I will run again with the neato
engine to see if we can pair down closer to where the problem is.

[north] 750,000 edge crossings?

It's not likely to be very readable, unfortunately.

[david] I'd like to see one complete - I can worry about readability next.

[north] With graphs that size you really might find it a lot easier
to use command line tools.

I'd suggest using a Unix-type system as well instead
of Windows, though technically Windows should work OK.
I do think that some versions of Windows get into real
trouble when they hit a resource limit, e.g. start swapping.
(Linux has its problems too.)

I wonder how you will be viewing a 600,000 node graph.
Are the nodes labeled? If you had a 600,000 word
document, could you display it as one page?


TagsNo tags attached.
AUXILLARY-FILES
DATE-FIXED
FIX-COMMENT
FORMER-ID1655
INPUT-FILE
OUTPUT-FILE
STATUS-COMMENT*
VERSION     2.23.20090527.0445
Attached Files

- Relationships

-  Notes
There are no notes attached to this issue.

- Issue History
Date Modified Username Field Change
2011-04-28 04:03 user1 New Issue
2011-04-28 04:03 user1 Assigned To => user1


MantisBT 1.2.5[^]
Copyright © 2000 - 2011 MantisBT Group
Powered by Mantis Bugtracker