Hi
I’ve setup a fresh L4T R21 system and compiled OpenMPI 1.8.3 on my own via
“./configure --enable-mpi-thread-multiple --with-threads --with-cuda ; make ; sudo make install”.
When i execute a simple sample code, found on
https://www.olcf.ornl.gov/tutorials/gpudirect-mpich-enabled-cuda/#Examples
and compiled
via mpicc -o example example.c -L/usr/local/cuda/lib -lcudart -I/usr/local/cuda/include
with
dribbroc@tegra-ubuntu:~/ompi_cuda$ mpicc --version
gcc (Ubuntu/Linaro 4.8.2-19ubuntu1) 4.8.2
Copyright (C) 2013 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
dribbroc@tegra-ubuntu:~/ompi_cuda$ nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2014 NVIDIA Corporation
Built on Thu_Jul_17_18:43:29_CDT_2014
Cuda compilation tools, release 6.5, V6.5.12
dribbroc@tegra-ubuntu:~/ompi_cuda$
I get
dribbroc@tegra-ubuntu:~/ompi_cuda$ mpirun -np 2 example
Success!
--------------------------------------------------------------------------
The call to cuMemHostRegister(0xace67b80, 33792, 0) failed.
Host: tegra-ubuntu
cuMemHostRegister return value: 801
Memory Pool: sm
--------------------------------------------------------------------------
[tegra-ubuntu:03031] 11 more processes have sent help message help-mpi-common-cuda.txt / cuMemHostRegister failed
[tegra-ubuntu:03031] Set MCA parameter "orte_base_help_aggregate" to 0 to see all help / error messages
It seems, that the programm executes and finishes correcty, but nevertheless these errors are strange.
Has anyone else experience with cuda aware openmpi and encountered similar problems?
Thanks
Dirk
PS: OpenMPI 1.8.4 shows same behaviour.
PPS More verbosive:
dribbroc@tegra-ubuntu:~/ompi_cuda$ mpirun --mca orte_base_help_aggregate 0 -np 2 ./example
Success!
--------------------------------------------------------------------------
The call to cuMemHostRegister(0xace81700, 33792, 0) failed.
Host: tegra-ubuntu
cuMemHostRegister return value: 801
Memory Pool: sm
--------------------------------------------------------------------------
--------------------------------------------------------------------------
The call to cuMemHostRegister(0xace92000, 263168, 0) failed.
Host: tegra-ubuntu
cuMemHostRegister return value: 801
Memory Pool: sm
--------------------------------------------------------------------------
--------------------------------------------------------------------------
The call to cuMemHostRegister(0xacf12900, 1024, 0) failed.
Host: tegra-ubuntu
cuMemHostRegister return value: 801
Memory Pool: sm
--------------------------------------------------------------------------
--------------------------------------------------------------------------
The call to cuMemHostRegister(0xa4e80400, 33792, 0) failed.
Host: tegra-ubuntu
cuMemHostRegister return value: 801
Memory Pool: sm
--------------------------------------------------------------------------
--------------------------------------------------------------------------
The call to cuMemHostRegister(0xa4e91000, 263168, 0) failed.
Host: tegra-ubuntu
cuMemHostRegister return value: 801
Memory Pool: sm
--------------------------------------------------------------------------
--------------------------------------------------------------------------
The call to cuMemHostRegister(0xa4f11900, 1024, 0) failed.
Host: tegra-ubuntu
cuMemHostRegister return value: 801
Memory Pool: sm
--------------------------------------------------------------------------
--------------------------------------------------------------------------
The call to cuMemHostRegister(0xa53ddb80, 33792, 0) failed.
Host: tegra-ubuntu
cuMemHostRegister return value: 801
Memory Pool: sm
--------------------------------------------------------------------------
--------------------------------------------------------------------------
The call to cuMemHostRegister(0xa5426480, 263168, 0) failed.
Host: tegra-ubuntu
cuMemHostRegister return value: 801
Memory Pool: sm
--------------------------------------------------------------------------
--------------------------------------------------------------------------
The call to cuMemHostRegister(0xa546ae80, 1024, 0) failed.
Host: tegra-ubuntu
cuMemHostRegister return value: 801
Memory Pool: sm
--------------------------------------------------------------------------
--------------------------------------------------------------------------
The call to cuMemHostRegister(0xadf0fb80, 33792, 0) failed.
Host: tegra-ubuntu
cuMemHostRegister return value: 801
Memory Pool: sm
--------------------------------------------------------------------------
--------------------------------------------------------------------------
The call to cuMemHostRegister(0xadf58480, 263168, 0) failed.
Host: tegra-ubuntu
cuMemHostRegister return value: 801
Memory Pool: sm
--------------------------------------------------------------------------
--------------------------------------------------------------------------
The call to cuMemHostRegister(0xadf9ce80, 1024, 0) failed.
Host: tegra-ubuntu
cuMemHostRegister return value: 801
Memory Pool: sm
--------------------------------------------------------------------------