Re: [OMPI users] Error with MPI_Register_datarep

2016-03-10 Thread Gilles Gouaillardet
Eric, my short answer is no. long answer is : - from MPI_Register_datarep() /* The io framework is only initialized lazily. If it hasn't already been initialized, do so now (note that MPI_FILE_OPEN and MPI_FILE_DELETE are the only two places that it will be

Re: [OMPI users] Error with MPI_Register_datarep

2016-03-10 Thread Éric Chamberland
Thanks Gilles! it works... I will continue my tests with that command line... Until OMPIO supports this, is there a way to put a call into the code to disable ompio the same way --mca io ^ompio does? Thanks, Eric Le 16-03-10 20:13, Gilles Gouaillardet a écrit : Eric, I will fix the crash

Re: [OMPI users] Error with MPI_Register_datarep

2016-03-10 Thread Gilles Gouaillardet
Eric, I will fix the crash (fwiw, it is already fixed in v2.x and master) note this program cannot currently run "as is". by default, there are two frameworks for io : ROMIO and OMPIO. MPI_Register_datarep does try to register the datarep into all frameworks, and successes only if datarep was

Re: [OMPI users] What causes the pingpong bandwidth "hump" on SM?

2016-03-10 Thread BRADLEY, PETER C PW
Jeff et al, Thanks, exactly what I was looking for. Pete I think the information was scattered across a few posts, but the union of which is correct: - it depends on the benchmark - yes, L1/L2/L3 cache sizes can have a huge effect. I.e., once the buffer size gets bigger than the cache size, it

[OMPI users] Error with MPI_Register_datarep

2016-03-10 Thread Éric Chamberland
Hi, I have a segfault while trying to use MPI_Register_datarep with openmpi-1.10.2: mpic++ -g -o int64 int64.cc ./int64 [melkor:24426] *** Process received signal *** [melkor:24426] Signal: Segmentation fault (11) [melkor:24426] Signal code: Address not mapped (1) [melkor:24426] Failing at

Re: [OMPI users] What causes the pingpong bandwidth "hump" on SM?

2016-03-10 Thread Vincent Diepeveen
On Thu, 10 Mar 2016, BRADLEY, PETER C PW wrote: This is an academic exercise, obviously.  The curve shown comes from one pair of ranks running on the same node alternating between MPI_Send and MPI_Recv.  The most likely suspect is a cache effect, but rather than assuming, I was curious if

Re: [OMPI users] What causes the pingpong bandwidth "hump" on SM?

2016-03-10 Thread Jeff Squyres (jsquyres)
I think the information was scattered across a few posts, but the union of which is correct: - it depends on the benchmark - yes, L1/L2/L3 cache sizes can have a huge effect. I.e., once the buffer size gets bigger than the cache size, it takes more time to get the message from main RAM -->

Re: [OMPI users] What causes the pingpong bandwidth "hump" on SM?

2016-03-10 Thread BRADLEY, PETER C PW
This is an academic exercise, obviously. The curve shown comes from one pair of ranks running on the same node alternating between MPI_Send and MPI_Recv. The most likely suspect is a cache effect, but rather than assuming, I was curious if there might be any other aspects of the

[OMPI users] locked memory and queue pairs

2016-03-10 Thread Michael Di Domenico
when i try to run an openmpi job with >128 ranks (16 ranks per node) using alltoall or alltoallv, i'm getting an error that the process was unable to get a queue pair. i've checked the max locked memory settings across my machines; using ulimit -l in and outside of mpirun and they're all set to

Re: [OMPI users] Poor performance on Amazon EC2 with TCP

2016-03-10 Thread Jackson, Gary L.
I re-ran all experiments with 1.10.2 configured the way you specified. My results are here: https://www.dropbox.com/s/4v4jaxe8sflgymj/collected.pdf?dl=0 Some remarks: 1. OpenMPI had poor performance relative to raw TCP and IMPI across all MTUs. 2. Those issues appeared at larger message

[OMPI users] Failed Flash run on Pleiades with OpenMPI 1.10.2

2016-03-10 Thread Joshua Wall
Dear users, Hello, I'm relatively new to building OpenMPI from scratch, so I'm going to try and provide a lot of information about exactly what I did here. I'm attempting to run the MHD code Flash 4.2.2 on Pleiades (NASA AMES), and also need some python mpi4py functionality and Cuda which

Re: [OMPI users] What causes the pingpong bandwidth "hump" on SM?

2016-03-10 Thread Gilles Gouaillardet
Pete, how did you measure the bandwidth ? iirc, IMB benchmark does not reuse send and recv buffers, so the results could be different. also, you might want to use a logarithmic scale for the message size, so information for small messages is easier to read. Cheers, Gilles On Thursday, March

Re: [OMPI users] OMPI users] What causes the pingpong bandwidth "hump" on SM?

2016-03-10 Thread Gilles Gouaillardet
Fwiw,you might want to try compare sm and vader mpirun --mca btl self,sm ... And with and without knem (modprobe knem should do the trick) Cheers, Gilles Vincent Diepeveen wrote: > >You're trying to read absurd huge message sizes considering you're busy >testing the memory

Re: [OMPI users] What causes the pingpong bandwidth "hump" on SM?

2016-03-10 Thread Vincent Diepeveen
You're trying to read absurd huge message sizes considering you're busy testing the memory bandwidth of your system in this manner. As soon as the message gets larger than your CPU's caching system it has to copy the message several times via your RAM, falls outside CPU's L2 or L3 cache and