2009/1/6 Ralph Castain <firstname.lastname@example.org>
In the 1.2 series, mpirun spawns a local orted to handle all local procs. The code that discovers local interfaces specifically ignores any interfaces that are not up or are just local loopbacks. My guess is that the person who wrote that code long, long ago was assuming that the sole purpose was to talk to remote nodes, not to loop back onto yourself.
On Jan 5, 2009, at 5:19 PM, Jeff Squyres wrote:
On Jan 5, 2009, at 5:01 PM, Maciej Kazulak wrote:
Interesting though. I thought in such a simple scenario shared memory would be used for IPC (or whatever's fastest) . But nope. Even with one process still it wants to use TCP/IP to communicate between mpirun and orted.
Correct -- we only have TCP enabled for MPI process <--> orted communication. There are several reasons why; the simplest is that this is our "out of band" channel and it is only used to setup and tear down the job. As such, we don't care that it's a little slower than other possible channels (such as sm). MPI traffic will use shmem, OpenFabrics-based networks, Myrinet, ...etc. But not MPI process <--> orted communication.
What's even more surprising to me it won't use loopback for that. Hence my maybe a little bit over-restrictive iptables rules were the problem. I allowed traffic from 127.0.0.1 to 127.0.0.1 on lo but not from <eth0_addr> to <eth0_addr> on eth0 and both processes ended up waiting for IO.
Can I somehow configure it to use something other than TCP/IP here? Or at least switch it to loopback?
I don't remember how it works in the v1.2 series offhand; I think it's different in the v1.3 series (where all MPI processes *only* talk to the local orted, vs. MPI processes making direct TCP connections back to mpirun and any other MPI process with which it needs to bootstrap other communication channels). I'm *guessing* that the MPI process <--> orted communication either uses a named unix socket or TCP loopback. Ralph -- can you explain the details?
I imagine it could be changed to include loopback, but I would first need to work with other developers to ensure there are no unexpected consequences in doing so. Since no TCP interface is found, mpirun fails.
In the 1.3 series, mpirun handles the local procs itself. Thus, this issue does not appear and things run just fine.
users mailing list
users mailing list