[ns] traceing strategies
Pendzick, John N.
john.pendzick at ngc.com
Tue Feb 17 07:26:25 PST 2004
Sorry should have mentioned it that also.
I'm doing some post processing on the nam file produced, via perl.
Matter a fact to get nam to run a file that size have to read the file
in though a perl script and then pipe it out to nam (nam takes in a pipe
via that "nam -" ns-manual talks about it on pg 351), and in perl every
so many lines sent to nam have the script wait, this is to allow nam to
catch up and the pipe to kind of empty out.
Aaron Gallimore wrote:
>I wonder if you could edit the tracefile before running it. this way you could filter out any
>unwanted data before running the simulator? just a thought
>"Pendzick, John N." <john.pendzick at ngc.com> writes:
>>I'm writting simulations that simulate entire days, with large amounts
>>of traffic flow in them. In this day time spand I'm only really
>>intersted in certain points of time were some of the network physically
>>changes and/or the network patterns change. When these changes occur I
>>want a full trace of everthing that is happening. But durning some
>>parts when everthing is pretty much static I don't care about it.
>>The nam files that are being produced for all of this are in the range
>>of 30 to 50 GB. Nam doesn't want to play this to well, and waiting for
>>it to get to the parts I want to see takes a little bit of time.
>>I'm open to and looking for suggestions (and code snips if applicable)
>>on how to handle this better (making the nam file smaller, only getting
>>traces of the network at times I am intersted in, gettting nam to move
>>faster though the static parts of the simulation that is showing nothing
>>intersting, ect...). I'm using nam because I need the visulazation that
More information about the Ns-users