- codes-ross-users - mailman3.cels.anl.gov

CODES 0.4.1 Released!
by Jenkins, Jonathan P. 30 Sep '15

30 Sep '15

Hi all, CODES 0.4.1 has been released! This is a hotfix release that fixes an incompatibility with recent ROSS revisions. The new version can be found at http://www.mcs.anl.gov/research/projects/codes/downloads/ . Thanks, John

1 0

CODES repositories changes
by Jenkins, Jonathan P. 11 Aug '15

11 Aug '15

Hi all, Sometime in the not-so-distant future I'd like to merge the codes-base and codes-net repositories into a single one to simplify development, building, releases etc. I could do this now, but there's some external dev activity going on and I don't want to pull the rug out from under them, so once their changes get merged into the mainline master branch I'll be ready to do the update. Most likely, what this will mean is that 1) we'll make a new, unified repository for development of both, and 2) mark the original codes-base and codes-net repositories as read-only. We may additionally move to an MCS instance of gitlab to make it easier to collaborate RE issue tracking, pull requests, etc. (https://xgitlab.cels.anl.gov) but that hasn't been decided yet. Thoughts? Concerns? Thanks, John

1 0

Re: [codes-ross-users] g_tw_csv build errors?
by Phil Carns 10 Aug '15

10 Aug '15

Whoops, please disregard. I didn't realize it but my environment was building some other branch besides "master". -Phil On 08/10/2015 04:21 PM, Carns, Philip H. wrote: > Hi all, > > I was just trying to set up some regression tests in an in-house jenkins > server. I don't really need to test ROSS itself (since it already has > its own continuous integration setup), but I need to build it as the > first step to testing the CODES models that I really want to check. > > Anyhow, when I try to build it I get the following errors: > > /sandbox/jenkins-ci/workspace/codes-triton/ross/core/tw-stats.c: In > function ï¿½show_lldï¿½: > /sandbox/jenkins-ci/workspace/codes-triton/ross/core/tw-stats.c:7:10: > error: ï¿½g_tw_csvï¿½ undeclared (first use in this function) > /sandbox/jenkins-ci/workspace/codes-triton/ross/core/tw-stats.c:7:10: > note: each undeclared identifier is reported only once for each function > it appears in > /sandbox/jenkins-ci/workspace/codes-triton/ross/core/tw-stats.c: In > function ï¿½show_2fï¿½: > /sandbox/jenkins-ci/workspace/codes-triton/ross/core/tw-stats.c:14:10: > error: ï¿½g_tw_csvï¿½ undeclared (first use in this function) > /sandbox/jenkins-ci/workspace/codes-triton/ross/core/tw-stats.c: In > function ï¿½show_1fï¿½: > /sandbox/jenkins-ci/workspace/codes-triton/ross/core/tw-stats.c:21:10: > error: ï¿½g_tw_csvï¿½ undeclared (first use in this function) > /sandbox/jenkins-ci/workspace/codes-triton/ross/core/tw-stats.c: In > function ï¿½show_4fï¿½: > /sandbox/jenkins-ci/workspace/codes-triton/ross/core/tw-stats.c:28:10: > error: ï¿½g_tw_csvï¿½ undeclared (first use in this function) > /sandbox/jenkins-ci/workspace/codes-triton/ross/core/tw-stats.c: In > function ï¿½tw_statsï¿½: > /sandbox/jenkins-ci/workspace/codes-triton/ross/core/tw-stats.c:113:10: > error: ï¿½g_tw_csvï¿½ undeclared (first use in this function) > make[2]: *** [core/CMakeFiles/ROSS.dir/tw-stats.c.o] Error 1 > > This does *not* happen on my laptop, so it is probably something > peculiar to the build environment on this server. > > Before I dig into it, does anyone happen to know why I might have hit this? > > thanks, > -Phil > _______________________________________________ > codes-ross-users mailing list > codes-ross-users(a)lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/codes-ross-users

1 0

g_tw_csv build errors?
by Phil Carns 10 Aug '15

10 Aug '15

Hi all, I was just trying to set up some regression tests in an in-house jenkins server. I don't really need to test ROSS itself (since it already has its own continuous integration setup), but I need to build it as the first step to testing the CODES models that I really want to check. Anyhow, when I try to build it I get the following errors: /sandbox/jenkins-ci/workspace/codes-triton/ross/core/tw-stats.c: In function ‘show_lld’: /sandbox/jenkins-ci/workspace/codes-triton/ross/core/tw-stats.c:7:10: error: ‘g_tw_csv’ undeclared (first use in this function) /sandbox/jenkins-ci/workspace/codes-triton/ross/core/tw-stats.c:7:10: note: each undeclared identifier is reported only once for each function it appears in /sandbox/jenkins-ci/workspace/codes-triton/ross/core/tw-stats.c: In function ‘show_2f’: /sandbox/jenkins-ci/workspace/codes-triton/ross/core/tw-stats.c:14:10: error: ‘g_tw_csv’ undeclared (first use in this function) /sandbox/jenkins-ci/workspace/codes-triton/ross/core/tw-stats.c: In function ‘show_1f’: /sandbox/jenkins-ci/workspace/codes-triton/ross/core/tw-stats.c:21:10: error: ‘g_tw_csv’ undeclared (first use in this function) /sandbox/jenkins-ci/workspace/codes-triton/ross/core/tw-stats.c: In function ‘show_4f’: /sandbox/jenkins-ci/workspace/codes-triton/ross/core/tw-stats.c:28:10: error: ‘g_tw_csv’ undeclared (first use in this function) /sandbox/jenkins-ci/workspace/codes-triton/ross/core/tw-stats.c: In function ‘tw_stats’: /sandbox/jenkins-ci/workspace/codes-triton/ross/core/tw-stats.c:113:10: error: ‘g_tw_csv’ undeclared (first use in this function) make[2]: *** [core/CMakeFiles/ROSS.dir/tw-stats.c.o] Error 1 This does *not* happen on my laptop, so it is probably something peculiar to the build environment on this server. Before I dig into it, does anyone happen to know why I might have hit this? thanks, -Phil

1 0

Re: [codes-ross-users] Correlation between latency and number of hops on torus network?
by Mubarak, Misbah 23 Jul '15

23 Jul '15

Thanks Dan. Keep us posted. From: Daniel Parker <dkparker(a)uchicago.edu<mailto:[email protected]>> Date: Thursday, July 23, 2015 at 6:02 PM To: Misbah Mubarak <mmubarak(a)anl.gov<mailto:[email protected]>> Subject: Re: [codes-ross-users] Correlation between latency and number of hops on torus network? Hey Misbah, So when I ran a "cleaned up" version of my code, the problem didn't occur, so it looks like it's probably something I can find. I'm investigating further and will let you know about the result. On Thu, Jul 23, 2015 at 3:40 PM, Daniel Parker <dkparker(a)uchicago.edu<mailto:[email protected]>> wrote: Okay, I will try to clean it up for you and send it. Thank you! On Thu, Jul 23, 2015 at 2:19 PM, Mubarak, Misbah <mmubarak(a)anl.gov<mailto:[email protected]>> wrote: Thanks for the update, Dan. If you can share your torus config file (and code, if you have made modifications), I can take a quick peek and may be able to help you out. From: <codes-ross-users-bounces(a)lists.mcs.anl.gov<mailto:[email protected]>> on behalf of Daniel Parker <dkparker(a)uchicago.edu<mailto:[email protected]>> Date: Thursday, July 23, 2015 at 2:46 PM To: "codes-ross-users(a)lists.mcs.anl.gov<mailto:[email protected]>" <codes-ross-users(a)lists.mcs.anl.gov<mailto:[email protected]>> Subject: [codes-ross-users] Correlation between latency and number of hops on torus network? Hello, First I wanted to thank Jonathan and Misbah for their help on my last question. I was successful in setting up networks where arbitrary numbers of nodes are active or inactive. Now that I am able to do so, however, I have made a strange observation. It seems to me that there is not a correlation between the number of hops that a message must make over a torus network and that message's latency. To determine this, I created a simple model where one sender and one receiver node are randomly chosen within a 3D 10x10x10 torus network. I take latency measurements for 10 pings from the sender to the receiver and back, then average the values. I repeated this trial 500 times and aggregated the data into these graphs: http://imgur.com/a/7rqRw The first graph shows a cdf of latencies of the different trials, with each data point color-coded corresponding to the number of hops between the sender and receiver for that trial. Blue indicates few hops while red indicates more. The second shows the distribution of hop counts observed over the 500 trials, with number of hops on the x axis and frequency on the y axis. As you can see, the latency seems constant despite distance between the nodes in question, which does not make sense to me. Do any of you know why this would be? Is there something about the torus implementation that I'm misunderstanding? I can upload my code if necessary. Thanks for your help, Dan Parker

1 0

Correlation between latency and number of hops on torus network?
by Daniel Parker 23 Jul '15

23 Jul '15

Hello, First I wanted to thank Jonathan and Misbah for their help on my last question. I was successful in setting up networks where arbitrary numbers of nodes are active or inactive. Now that I am able to do so, however, I have made a strange observation. It seems to me that there is not a correlation between the number of hops that a message must make over a torus network and that message's latency. To determine this, I created a simple model where one sender and one receiver node are randomly chosen within a 3D 10x10x10 torus network. I take latency measurements for 10 pings from the sender to the receiver and back, then average the values. I repeated this trial 500 times and aggregated the data into these graphs: http://imgur.com/a/7rqRw The first graph shows a cdf of latencies of the different trials, with each data point color-coded corresponding to the number of hops between the sender and receiver for that trial. Blue indicates few hops while red indicates more. The second shows the distribution of hop counts observed over the 500 trials, with number of hops on the x axis and frequency on the y axis. As you can see, the latency seems constant despite distance between the nodes in question, which does not make sense to me. Do any of you know why this would be? Is there something about the torus implementation that I'm misunderstanding? I can upload my code if necessary. Thanks for your help, Dan Parker

3 2

rc-stack updates
by Jenkins, Jonathan P. 22 Jul '15

22 Jul '15

Hi all, Not sure if anyone's using the rc-stack interface yet (codes-base/codes/rc-stack.h), but just in case, I wanted to give a heads up that I'll be redoing the interface to allow the passing in of a free function rather than implicitly calling free on each pointer. This is so that the interface can easier work with aggregate data structures where a single free is not sufficient. Thanks, John

1 0

Workarounds to allow idle nodes in Torus Network?
by Daniel Parker 06 Jul '15

06 Jul '15

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

3 2

workload API changes
by Jenkins, Jonathan P. 17 Jun '15

17 Jun '15

Hi all, Just a heads up, we'll be changing the workload API soon. We wish to enable the case of single simulations running multiple workloads concurrently. To do so, we need to get away from the flat 0..N rank namespace that is currently assumed by the API and its implementations and introduce a namespace for applications with their own ranklists. So, a participant in a given workload will now be identified by < app id, rank id >, rather than just the rank id. Thanks, John

1 0

Re: [codes-ross-users] CODES LP Design Question
by Phil Carns 15 May '15

15 May '15

FWIW, in my use case I ended up going for a simplified version of what John's resource-lp API is doing (see attached header). It lets the caller provide an entire event that the target LP should fire back when it is done with it's work. This event is opaque to the target lp, except that it is given the offset of a struct within the event where it can store its output data. The disadvantage to the simplified approach is that it inflates the event size (because the events in the target LP have to be big enough to also contain an entire event from the caller). John's approach combats that by specifying specific components of the event to be fired back instead of handing the whole thing over (so the target LP doesn't have to save any more than is absolutely necessary- it just constructs the right thing later), but the API is a little more involved as a result. Either variation solves the core problem, at least. Neither LP has to understand the other's event structure, and the API doesn't require you to manually build an event message, but there is still a mechanism to get data back out. -Phil On 05/14/2015 04:11 PM, Joe Scott wrote: > I will take a look at the resource-lp stuff and see if I can make sense of it! Thanks for the pointer. > > Joe > >> On May 14, 2015, at 3:43 PM, Jenkins, Jonathan P. <jenkins(a)mcs.anl.gov> wrote: >> >> The callback functionality you describe has been used in a few places. >> Notably, the "resource" LP in codes-base implements such a method (see >> codes/resource-lp.h, resource_lp_* functions). It's not the prettiest, but >> it gets the job done. Generalizing the concept and making it a bit easier >> to work with in codes has been in the back of my mind for a while, but I >> haven't done anything with it. >> >> John >> >> On 5/14/15, 1:51 PM, "Phil Carns" <carns(a)mcs.anl.gov> wrote: >> >>> What I described in this thread turned out to be a bad idea. >>> >>> The problem with handing off a function pointer is that the target LP >>> (thing2 in the example below) might not be on the same process as the >>> caller (thing1), so in the general case it isn't safe to let it hang >>> onto a function pointer to issue the completion event. >>> >>> The lsm/local-storage-model API has this same issue (a higher level >>> models want to hand control off to a disk model and then just get >>> proceed with with its own work once the disk access is complete). LSM >>> solves this safely from a technical point of view by providing functions >>> for the caller to allocate an appropriate event, and letting the caller >>> specify it's own event struct that will be handed back on completion. >>> This is safe regardless of how your LPs are organized, but it still bugs >>> me a little, though, because a) it looks burly in the code and b) the >>> target LP doesn't have a way to pass results back in the completion event. >>> >>> Anyway, I'll putter around with this and share if I come up with >>> something cleaner, but I wanted to at least respond to this thread as a >>> warning in the mean time :) >>> >>> thanks, >>> -Phil >>> >>> On 04/08/2015 10:44 AM, Carns, Philip H. wrote: >>>> I've never tried to modularize steps like this in a single LP, but it is >>>> a little similar to some cross-LP scenarios I've run into before. Here >>>> comes a long-winded example :) >>>> >>>> Imagine you have two LPs (that run on the same simulated node), called >>>> thing1 and thing2. You want the thing1 procedure to invoke thing2, and >>>> then for thing1 to continue once thing2 is done. >>>> >>>> If I were to do this now I think I would set it up like this. Assume >>>> that thing1 and thing2 each have their own .c and .h files. >>>> >>>> thing2.h has this prototype: void start_thing2(lp, gid, >>>> completion_fn_ptr); >>>> >>>> This function will construct and send an event to the thing2 LP >>>> (identified by the gid argument) to kick off whatever it is going to >>>> do. This function will be executed within a thing1 LP event handler, >>>> but the definition of the function is in thing2.c. The purpose of this >>>> arrangement is to let thing1 send an event to thing2 without having >>>> access to thing2's event structure definition. As long as you keep the >>>> API stable then thing2.c can do whatever it needs to do in terms of >>>> refactoring its event structure without perturbing thing1. >>>> >>>> The lp argument is the lp pointer for the *calling* LP, and is probably >>>> needed to allocate a new event. >>>> >>>> The completion_fn_ptr is a function pointer to a function defined by the >>>> caller (thing1 in this case) that sends an event back to the original >>>> LP. This is the mechanism for thing2 to tell thing1 it is done so that >>>> thing1 can continue operation. Since it is an opaque function pointer, >>>> thing2 remains oblivious to the event structure definition of thing1 (or >>>> even what kind of LP thing1 is). You can make the completion_fn_ptr >>>> signature anything that you like, so if thing2 has some result to report >>>> back, then it can be passed as an argument to that function and stuffed >>>> into the event that heads back to thing1. >>>> >>>> Of course you can also add whatever arbitrary arguments you want to >>>> start_thing2() to pass input parameters it needs to start whatever >>>> thing2 is doing. Those would get stuffed into the event destined for >>>> thing2. >>>> >>>> Unfortunately I haven't actually done this so I don't have an example. >>>> My thinking on how to organize this has evolved over time and I haven't >>>> gotten around to trying this particular permutation :) I've done >>>> something very similar except for the function pointer; I have a few >>>> examples have an explicitly defined function that jumps back from thing2 >>>> to thing1. That works fine, but it is bad news if you want more than >>>> one type of LP to be able to call thing2. >>>> >>>> A related theme here is that I'm a fan (possibly in the minority on >>>> this) of splitting up different complex procedures that happen on the >>>> same simulated node into separate LPs. As long as you have them share >>>> node resources (NIC, storage, etc.) then the net result is the same in >>>> terms of the simulation outcome. The positive is that you don't have >>>> event handler explosion (big switch statements to handle different >>>> cases). The negative is that you have to jump between these LPs by >>>> sending events with small timestamps. Once you start doing that, then >>>> parallel conservative execution is practically out of the question; you >>>> have to do serial execution or parallel optimistic execution to handle >>>> those extra small events in a reasonable way. >>>> >>>> This might be overkill for your scenario, but I thought I would throw it >>>> out there. I wanted to collect my thoughts on this anyway. In general >>>> I don't think there is any particular way around this without juggling >>>> callbacks and such; really the best thing you can hope for right now is >>>> to minimize entanglement between different LP types in your code. >>>> >>>> -Phil >>>> >>>> On 04/08/2015 06:47 AM, Ross, Robert B. wrote: >>>>> I think the way to think about this is that the event payload carries >>>>> the state associated with the operation and gets passed around between >>>>> LPs as needed to simulate the steps. If you do this right, there >>>>> shouldn't be local LP state associated with the operation -- everything >>>>> is in the event payload that is needed. >>>>> >>>>> There might be state at LPs associated with the LP itself, such as >>>>> what objects exist or what names are in the namespace. Or not, if you >>>>> are just assuming things exist for instance. >>>>> >>>>> John, Phil, and/or Misbah may need to correct me...of course. >>>>> >>>>> -- Rob >>>>> >>>>>> On Apr 7, 2015, at 4:09 PM, Jenkins, Jonathan P. >>>>>> <jenkins(a)mcs.anl.gov> wrote: >>>>>> >>>>>> Hi Joe, >>>>>> >>>>>> So if I understand right, the context is you want to simulate some >>>>>> POSIX >>>>>> operations of interest, such as mkdir, and that the client is capable >>>>>> of >>>>>> performing the stat (the lookup) using purely local information >>>>>> (hence the >>>>>> self-event)? >>>>>> >>>>>> In general, it's useful to think of events in these systems as RPC >>>>>> calls, >>>>>> where the event structure you are setting up contains the parameters, >>>>>> and >>>>>> there are corresponding return parameters that the calling LP expects >>>>>> to >>>>>> receive. In this frame of thought, both RPC calls and returns issue a >>>>>> discrete event. In your example, you could have your "mkdir" RPC event >>>>>> handler perform an additional "lookup/stat" RPC to check the path's >>>>>> existence, and upon receiving the "return" event, either make the >>>>>> directory or fail. >>>>>> >>>>>> Unfortunately, event-driven programming more-or-less necessitates >>>>>> callback-heavy code, which can be quite awkward in some contexts. >>>>>> We've >>>>>> talked in the past about ways to at least standardize the way we do >>>>>> this >>>>>> in the context of CODES, but nothing has materialized as of yet. >>>>>> >>>>>> Hope that helps. Something tells me that the assumption about clients >>>>>> being capable of performing metadata operations without outside >>>>>> interaction with e.g. a storage server is not quite right, though I >>>>>> could >>>>>> be misunderstanding. >>>>>> >>>>>> Thanks, >>>>>> John >>>>>> >>>>>>> On 4/7/15, 3:43 PM, "Joe Scott" <tscott2(a)g.clemson.edu> wrote: >>>>>>> >>>>>>> Hello All, >>>>>>> >>>>>>> I am having a hard time wrapping myself around the programming >>>>>>> paradigm >>>>>>> here, and I wonder if you might offer some guidance on a better way >>>>>>> to >>>>>>> use CODES. >>>>>>> >>>>>>> So I am trying to process these higher level tasks (POSIX tasks like >>>>>>> mkdir) by launching the subevents as separate processes. >>>>>>> >>>>>>> The specific case that is tying me in knots is a user issuing a >>>>>>> mkdir. >>>>>>> It launches the mkdir event handler, which needs to perform a lookup >>>>>>> on >>>>>>> the path of the mkdir. >>>>>>> >>>>>>> So I need to send an event from this client LP to itself to perform >>>>>>> the >>>>>>> lookup. But I also need the lookup, upon completion, to relaunch the >>>>>>> mkdir task. >>>>>>> >>>>>>> Speaking it over with some of my lab mates, they seem to think I am >>>>>>> either overthinking it or trying to use the wrong tool for the job. >>>>>>> >>>>>>> Is this a usecase you guys are familiar with? Can you shed some >>>>>>> light on >>>>>>> this situation? >>>>>>> >>>>>>> I feel like there should be a way to do this without getting into >>>>>>> callback/completion function hell. >>>>>>> >>>>>>> Thanks, >>>>>>> Joe Scott >>>>>>> Clemson University >>>>>>> _______________________________________________ >>>>>>> codes-ross-users mailing list >>>>>>> codes-ross-users(a)lists.mcs.anl.gov >>>>>>> https://lists.mcs.anl.gov/mailman/listinfo/codes-ross-users >>>>>> _______________________________________________ >>>>>> codes-ross-users mailing list >>>>>> codes-ross-users(a)lists.mcs.anl.gov >>>>>> https://lists.mcs.anl.gov/mailman/listinfo/codes-ross-users >>>>> _______________________________________________ >>>>> codes-ross-users mailing list >>>>> codes-ross-users(a)lists.mcs.anl.gov >>>>> https://lists.mcs.anl.gov/mailman/listinfo/codes-ross-users >>>> _______________________________________________ >>>> codes-ross-users mailing list >>>> codes-ross-users(a)lists.mcs.anl.gov >>>> https://lists.mcs.anl.gov/mailman/listinfo/codes-ross-users >>> _______________________________________________ >>> codes-ross-users mailing list >>> codes-ross-users(a)lists.mcs.anl.gov >>> https://lists.mcs.anl.gov/mailman/listinfo/codes-ross-users >> _______________________________________________ >> codes-ross-users mailing list >> codes-ross-users(a)lists.mcs.anl.gov >> https://lists.mcs.anl.gov/mailman/listinfo/codes-ross-users

1 0