[Simgrid-user] Scalability of PlatformGenerator improved...

Kurt Vanmechelen kurt.vanmechelen at ua.ac.be
Thu Jul 24 14:45:09 CEST 2008


> Like you said, GTNetS needs it and some other people like to have that
> kind of information even if SimGrid generally does not make use of it.
> So maybe we should make it an option to this code if it really slows
> down the generation or the parsing.
Well the main factor in the slow generation of the platform file was the 
string handling so I wouldn't necessarily exclude this for performance 
reasons. 

In fact, we (and I assume others as well) would be interested in a platform 
description of the network topology that is graph-based (as opposed to giving 
info on all static routes). For frameworks that do not use static routing 
(e.g. GridSim), this info is indeed necessary and also a lot more space-
friendly. I therefore wonder if you have versions of the platform files in the 
description archive with the router info included?

If this is not the case I will build the graph describing the network topology 
by parsing the routes... .

In any case, such a description would also help scaling SimGrid up because, if 
I understand correctly, you now have the cluster tags in the new DTD but for 
instantiating the simulation you still need a routing table that scales 
quadratically in the number of hosts (CPUs) in the grid right?

Cheers,

Kurt


On Thursday 24 July 2008 13:12:14 Arnaud Legrand wrote:
> Hi Kurt,
>
> I do not really know this piece of code but I still can give my opinion. :)
>
> Kurt Vanmechelen wrote:
> > Note that I also removed the router elements in the routes. Is this OK?
> > The original code generated for every link : //Snippet of makeRoute...
> > 	for (int i=1; i<route.size() - 1; i++) {
> > 		String id = "bb_" + (String)route.get(i);
> > 		r += "<link:ctn id=\"" + id + "\"/>";
> > 		r += "<router id=\"router_" + id + "\"/>";
> > But as far as I know, SimGrid doesn't really consider these router
> > elements (apart from GTNeTS)?
>
> Like you said, GTNetS needs it and some other people like to have that
> kind of information even if SimGrid generally does not make use of it.
> So maybe we should make it an option to this code if it really slows
> down the generation or the parsing.
>
> Cheers,
>
>      Arnaud


-- 
================================================
Kurt Vanmechelen
Phd Student
Dept. of Mathematics and Computer Sciences
Group CoMP (Computational Modelling and Programming)
University of Antwerp
Middelheimlaan 1
2020 Antwerpen, Belgium
G2.14

Phone : 03/265 3477
Mail  : kurt.vanmechelen at ua.ac.be
================================================




More information about the Simgrid-user mailing list