[codes-ross-users] Workarounds to allow idle nodes in Torus Network?

[Mubarak, Misbah]

Hi Daniel,

I assume by idle nodes you are referring to network nodes that will participate in the routing but will not generate or receive any messages? We ran into a similar situation where the number of servers is less than the network size but in that case the idle torus nodes were the last “n” network nodes (Example is in codes-net—>model-net-dumpi-traces-dump.c). For randomly distributed nodes, I would suggest having a list of relative Ids (as John suggested) and isolate those LPs from generating and receiving messages.

Hope that works!

Thanks,
Misbah

From: <codes-ross-users-bounces at lists.mcs.anl.gov<mailto:codes-ross-users-bounces at lists.mcs.anl.gov>> on behalf of "Jenkins, Jonathan P." <jenkins at mcs.anl.gov<mailto:jenkins at mcs.anl.gov>>
Date: Monday, July 6, 2015 at 8:29 PM
To: Daniel Parker <dkparker at uchicago.edu<mailto:dkparker at uchicago.edu>>, "codes-ross-users at lists.mcs.anl.gov<mailto:codes-ross-users at lists.mcs.anl.gov>" <codes-ross-users at lists.mcs.anl.gov<mailto:codes-ross-users at lists.mcs.anl.gov>>
Subject: Re: [codes-ross-users] Workarounds to allow idle nodes in Torus Network?

Hi Daniel,

Glad to hear the docs have been useful to you!

Right now, our configuration scheme isn't well-suited to irregularity - annotations help delineate "sections" of LPs, but not if you want to "pick and choose" LPs. I would suggest a separate configuration routine for this. One option for your case is to have a whitelist of torus nodes to run that you check at runtime, set up via an LP "configure" function you call in main. You could either have a file with a set of IDs (relative IDs would be good here, that you would translate via codes_mapping_lpid_from_relative), or if you want a "topology-aware" mapping, a set of torus coordinates (there's utilities for torus mapping functions in codes-net/codes/model-net-inspect.h).

Hope this helps!
John


From: <codes-ross-users-bounces at lists.mcs.anl.gov<mailto:codes-ross-users-bounces at lists.mcs.anl.gov>> on behalf of Daniel Parker <dkparker at uchicago.edu<mailto:dkparker at uchicago.edu>>
Date: Monday, July 6, 2015 at 3:31 PM
To: "codes-ross-users at lists.mcs.anl.gov<mailto:codes-ross-users at lists.mcs.anl.gov>" <codes-ross-users at lists.mcs.anl.gov<mailto:codes-ross-users at lists.mcs.anl.gov>>
Subject: [codes-ross-users] Workarounds to allow idle nodes in Torus Network?

Hello,

My name is Daniel Parker, and I'm an undergrad in the BigDataX REU, in which I am doing benchmarking on 3D torus networks using CODES/ROSS. I am interested in studying latency and throughput over the network as the number of nodes in communication increases, from just one pair of nodes to a fully busy network. However, I noticed in the MN_TODO file (/codes-net/src/models/networks/model-net) that idle nodes within the torus network have not yet been implemented. Indeed, when I attempt to have less repetitions of my servers than the total number of nodes in the network, the simulation will not run (as expected).

I am wondering whether in the mean time there is any way that I can achieve this sort of functionality. One idea that I had was to use annotations to specify some nodes as having "dummy" kickoff handlers that do not give them work. However, my concern is that use of annotations could introduce a systematic nature to the "off" nodes, i.e. having them all have the same x coordinate or be close together or something of that sort, whereas I would want them to be randomly distributed across the network. This concern comes from uncertainty as to how codes-net constructs the network in such heterogeneous models.

So my question is whether there is a way to achieve these sorts of randomly distributed idle nodes within a torus network, so as to model some-to-some communication. Can annotations do what I want? Or is there some other work-around as of now?

Thank you for your time. I appreciate the work that has gone into CODES and ROSS, especially the well-commented example files and the GETTING_STARTED files, which have proved very useful.

Best,

Daniel Parker
Computer Science
University of Chicago 2017
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/codes-ross-users/attachments/20150707/8e61809f/attachment-0001.html>