Re: [OMPI users] round-robin scheduling question [hostfile]
Hi Ralph, Ralph Castain wrote: ... The man page will describe all the various options. Which one is best for your app really depends on what the app is doing, the capabilities and topology of your cluster, etc. A little experimentation can help you get a feel for when to use which one. Thank you for the explanation! So far, I've only been using -np and letting the rest work itself through magic :-) -- but, I'll try the options you suggested and also other options in the man page of mpirun to see what works for my application... Thanks again! Ray
Re: [OMPI users] round-robin scheduling question [hostfile]
On Feb 21, 2009, at 1:05 AM, Raymond Wan wrote: Hi Ralph, Thank you very much for your explanation! Ralph Castain wrote: It is a little bit of both: * historical, because most MPI's default to mapping by slot, and * performance, because procs that share a node can communicate via shared memory, which is faster than sending messages over an interconnect, and most apps are communication-bound If your app is disk-intensive, then mapping it -bynode may be a better Ok -- by this, it seems that there is no "rule" that says one is obviously better than the other. It depends on factors such as disk access and shared memory access and which one is dominating. So, it is worth to try both to see? Can't hurt! You might be able to tell by knowing what your app is doing, but otherwise, feel free to experiment. option for you. That's why we provide it. Note, however, that you can still wind up with multiple procs on a node. All "bynode" means is that the ranks are numbered consecutively bynode - it doesn't mean that there is only one proc/node. I see. But if the number of processes (as specified using -np) is less than the number of nodes, if "by node" is chosen, then is it guaranteed that only one process will be on each node? That is correct Is there a way to write the hostfile to ensure this? You don't need to do anything in the hostfile - if you use bynode and np < #nodes, it is guaranteed that you will have only one proc/node I was curious if a node has 4 slots, whether writing it 4 times in the hostfile with 1 slot each has any meaning. Might be a bad idea as we are trying to fool mpirun? It won't have any meaning as we aggregate the results. In other words, we read through the hostfile, and if a host appears more than once, we simply add the #slots on subsequent entries to the earlier one. So we wind up with just one instance of that host that has the total number of slots allocated to it. If you truly want one proc/node, then you should use the -pernode option. This maps one proc on each node up to either the number of procs you specified or the number of available nodes. If you don't specify -np, we just put one proc on each node in your allocation/ hostfile. I see ... I was not aware of that option; thank you! Do a "man mpirun" and you will see that there are several mapping options that might interest you, including: 1. npernode - let's you specify how many procs/node (as opposed to "pernode", where you only get one proc/node - obviously, pernode is the equivalent of "-npernode 1") 2. seq - a sequential mapper. This mapper will read a file (which can be different from the hostfile used to specify your allocation) and assign one proc to each entry in a sequential manner like this: node1 > rank 0 goes on node1 node5 > rank 1 goes on node5 node1 > rank 2 goes on node1 ... 3. rank_file - allows you to specify that rank x goes on node foo, and what core/socket that rank should be bound to The man page will describe all the various options. Which one is best for your app really depends on what the app is doing, the capabilities and topology of your cluster, etc. A little experimentation can help you get a feel for when to use which one. HTH Ralph Ray ___ users mailing list us...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/users
Re: [OMPI users] round-robin scheduling question [hostfile]
Hi Ralph, Thank you very much for your explanation! Ralph Castain wrote: It is a little bit of both: * historical, because most MPI's default to mapping by slot, and * performance, because procs that share a node can communicate via shared memory, which is faster than sending messages over an interconnect, and most apps are communication-bound If your app is disk-intensive, then mapping it -bynode may be a better Ok -- by this, it seems that there is no "rule" that says one is obviously better than the other. It depends on factors such as disk access and shared memory access and which one is dominating. So, it is worth to try both to see? option for you. That's why we provide it. Note, however, that you can still wind up with multiple procs on a node. All "bynode" means is that the ranks are numbered consecutively bynode - it doesn't mean that there is only one proc/node. I see. But if the number of processes (as specified using -np) is less than the number of nodes, if "by node" is chosen, then is it guaranteed that only one process will be on each node? Is there a way to write the hostfile to ensure this? I was curious if a node has 4 slots, whether writing it 4 times in the hostfile with 1 slot each has any meaning. Might be a bad idea as we are trying to fool mpirun? If you truly want one proc/node, then you should use the -pernode option. This maps one proc on each node up to either the number of procs you specified or the number of available nodes. If you don't specify -np, we just put one proc on each node in your allocation/hostfile. I see ... I was not aware of that option; thank you! Ray
Re: [OMPI users] round-robin scheduling question [hostfile]
It is a little bit of both: * historical, because most MPI's default to mapping by slot, and * performance, because procs that share a node can communicate via shared memory, which is faster than sending messages over an interconnect, and most apps are communication-bound If your app is disk-intensive, then mapping it -bynode may be a better option for you. That's why we provide it. Note, however, that you can still wind up with multiple procs on a node. All "bynode" means is that the ranks are numbered consecutively bynode - it doesn't mean that there is only one proc/node. If you truly want one proc/node, then you should use the -pernode option. This maps one proc on each node up to either the number of procs you specified or the number of available nodes. If you don't specify -np, we just put one proc on each node in your allocation/ hostfile. HTH Ralph On Feb 20, 2009, at 1:25 AM, Raymond Wan wrote: Hi all, According to FAQ 14 (How do I control how my processes are scheduled across nodes?) [http://www.open-mpi.org/faq/?category=running#mpirun-scheduling ], it says that the default scheduling policy is by slot and not by node. I'm curious why the default is "by slot" since I am thinking of explicitly specifying by node but I'm wondering if there is an issue which I haven't considered. I would think that one reason for "by node" is to distribute HDD access across machines [as is the case for me since my program is HDD access intensive]. Or perhaps I am mistaken? I'm now thinking that "by slot" is the default because processes with ranks that are close together might do similar tasks and you would want them on the same node? Is that the reason? Also, at the end of this FAQ, it says "NOTE: This is the scheduling policy in Open MPI because of a long historical precendent..." -- does this "This" refer to "the fact that there are two scheduling policies" or "the fact that 'by slot' is the default"? If the latter, then that explains why "by slot" is the default, I guess... Thank you! Ray ___ users mailing list us...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/users
[OMPI users] round-robin scheduling question [hostfile]
Hi all, According to FAQ 14 (How do I control how my processes are scheduled across nodes?) [http://www.open-mpi.org/faq/?category=running#mpirun-scheduling], it says that the default scheduling policy is by slot and not by node. I'm curious why the default is "by slot" since I am thinking of explicitly specifying by node but I'm wondering if there is an issue which I haven't considered. I would think that one reason for "by node" is to distribute HDD access across machines [as is the case for me since my program is HDD access intensive]. Or perhaps I am mistaken? I'm now thinking that "by slot" is the default because processes with ranks that are close together might do similar tasks and you would want them on the same node? Is that the reason? Also, at the end of this FAQ, it says "NOTE: This is the scheduling policy in Open MPI because of a long historical precendent..." -- does this "This" refer to "the fact that there are two scheduling policies" or "the fact that 'by slot' is the default"? If the latter, then that explains why "by slot" is the default, I guess... Thank you! Ray