Permalink
172e9fd May 17, 2018
1 contributor

Users who have contributed to this file

313 lines (235 sloc) 13.2 KB

Troubleshooting

Import TensorFlow failed during installation

  1. Is TensorFlow installed?

If you see the error message below, it means that TensorFlow is not installed. Please install TensorFlow before installing Horovod.

error: import tensorflow failed, is it installed?

Traceback (most recent call last):
  File "/tmp/pip-OfE_YX-build/setup.py", line 29, in fully_define_extension
    import tensorflow as tf
ImportError: No module named tensorflow
  1. Are the CUDA libraries available?

If you see the error message below, it means that TensorFlow cannot be loaded. If you're installing Horovod into a container on a machine without GPUs, you may use CUDA stub drivers to work around the issue.

error: import tensorflow failed, is it installed?

Traceback (most recent call last):
  File "/tmp/pip-41aCq9-build/setup.py", line 29, in fully_define_extension
    import tensorflow as tf
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/__init__.py", line 24, in <module>
    from tensorflow.python import *
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/__init__.py", line 49, in <module>
    from tensorflow.python import pywrap_tensorflow
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/pywrap_tensorflow.py", line 52, in <module>
    raise ImportError(msg)
ImportError: Traceback (most recent call last):
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/pywrap_tensorflow.py", line 41, in <module>
    from tensorflow.python.pywrap_tensorflow_internal import *
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 28, in <module>
    _pywrap_tensorflow_internal = swig_import_helper()
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 24, in swig_import_helper
    _mod = imp.load_module('_pywrap_tensorflow_internal', fp, pathname, description)
ImportError: libcuda.so.1: cannot open shared object file: No such file or directory

To use CUDA stub drivers:

# temporary add stub drivers to ld.so.cache
$ ldconfig /usr/local/cuda/lib64/stubs

# install Horovod, add other HOROVOD_* environment variables as necessary
$ HOROVOD_GPU_ALLREDUCE=NCCL HOROVOD_NCCL_HOME=/path/to/nccl pip install --no-cache-dir horovod

# revert to standard libraries
$ ldconfig

MPI is not found during installation

  1. Is MPI in PATH?

If you see the error message below, it means mpicxx was not found in PATH. Typically mpicxx is located in the same directory as mpirun. Please add a directory containing mpicxx to PATH before installing Horovod.

error: mpicxx -show failed, is mpicxx in $PATH?

Traceback (most recent call last):
  File "/tmp/pip-dQ6A7a-build/setup.py", line 70, in get_mpi_flags
    ['mpicxx', '-show'], universal_newlines=True).strip()
  File "/usr/lib/python2.7/subprocess.py", line 566, in check_output
    process = Popen(stdout=PIPE, *popenargs, **kwargs)
  File "/usr/lib/python2.7/subprocess.py", line 710, in __init__
    errread, errwrite)
  File "/usr/lib/python2.7/subprocess.py", line 1335, in _execute_child
    raise child_exception
OSError: [Errno 2] No such file or directory

To use custom MPI directory:

$ export PATH=$PATH:/path/to/mpi/bin
$ HOROVOD_GPU_ALLREDUCE=NCCL HOROVOD_NCCL_HOME=/path/to/nccl pip install --no-cache-dir horovod
  1. Are MPI libraries added to $LD_LIBRARY_PATH or ld.so.conf?

If you see the error message below, it means mpicxx was not able to load some of the MPI libraries. If you recently installed MPI, make sure that the path to MPI libraries is present the $LD_LIBRARY_PATH environment variable, or in the /etc/ld.so.conf file.

mpicxx: error while loading shared libraries: libopen-pal.so.40: cannot open shared object file: No such file or directory
error: mpicxx -show failed (see error below), is MPI in $PATH?
Note: If your version of MPI has a custom command to show compilation flags, please specify it with the HOROVOD_MPICXX_SHOW environment variable.

Traceback (most recent call last):
File "/tmp/pip-build-wrtVwH/horovod/setup.py", line 107, in get_mpi_flags
shlex.split(show_command), universal_newlines=True).strip()
File "/usr/lib/python2.7/subprocess.py", line 574, in check_output
raise CalledProcessError(retcode, cmd, output=output)
CalledProcessError: Command '['mpicxx', '-show']' returned non-zero exit status 127

If you have installed MPI in a user directory, you can add the MPI library directory to $LD_LIBRARY_PATH:

$ export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/path/to/mpi/lib

If you have installed MPI in a non-standard system location (i.e. not /usr or /usr/local), you should add it to the /etc/ld.so.conf file:

$ echo /path/to/mpi/lib | sudo tee -a /etc/ld.so.conf

Additionally, if you have installed MPI in a system location, you should run sudo ldconfig after installation to register libraries in the cache:

$ sudo ldconfig

Error during installation: invalid conversion from ‘const void*’ to ‘void*’ [-fpermissive]

If you see the error message below, it means that your MPI is likely outdated. We recommend installing Open MPI >=3.0.0.

Note: Prior to installing a new version of Open MPI, don't forget to remove your existing MPI installation.

horovod/tensorflow/mpi_ops.cc: In function ‘void horovod::tensorflow::{anonymous}::PerformOperation(horovod::tensorflow::{anonymous}::TensorTable&, horovod::tensorflow::MPIResponse)’:
horovod/tensorflow/mpi_ops.cc:802:79: # error: invalid conversion from ‘const void*’ to ‘void*’ [-fpermissive]
                                  recvcounts, displcmnts, dtype, MPI_COMM_WORLD);
                                                                               ^
In file included from horovod/tensorflow/mpi_ops.cc:38:0:
/usr/anaconda2/include/mpi.h:633:5: error:   initializing argument 1 of ‘int MPI_Allgatherv(void*, int, MPI_Datatype, void*, int*, int*, MPI_Datatype, MPI_Comm)’ [-fpermissive]
 int MPI_Allgatherv(void* , int, MPI_Datatype, void*, int *, int *, MPI_Datatype, MPI_Comm);
     ^
horovod/tensorflow/mpi_ops.cc:1102:45: error: invalid conversion from ‘const void*’ to ‘void*’ [-fpermissive]
                               MPI_COMM_WORLD))
                                             ^

Error during installation: fatal error: pyconfig.h: No such file or directory

If you see the error message below, it means that you need to install Python headers.

build/horovod/torch/mpi_lib/_mpi_lib.c:22:24: fatal error: pyconfig.h: No such file or directory
 #  include <pyconfig.h>
                        ^
compilation terminated.

You can do this by installing a python-dev or python3-dev package. For example, on a Debian or Ubuntu system:

$ sudo apt-get install python-dev

NCCL 2 is not found during installation

If you see the error message below, it means NCCL 2 was not found in the standard libraries location. If you have a directory where you installed NCCL 2 which has both include and lib directories containing nccl.h and libnccl.so respectively, you can pass it via HOROVOD_NCCL_HOME environment variable. Otherwise you can specify them separately via HOROVOD_NCCL_INCLUDE and HOROVOD_NCCL_LIB environment variables.

build/temp.linux-x86_64-2.7/test_compile/test_nccl.cc:1:18: fatal error: nccl.h: No such file or directory
 #include <nccl.h>
                  ^
compilation terminated.
error: NCCL 2.0 library or its later version was not found (see error above).
Please specify correct NCCL location via HOROVOD_NCCL_HOME environment variable or combination of HOROVOD_NCCL_INCLUDE and HOROVOD_NCCL_LIB environment variables.

HOROVOD_NCCL_HOME - path where NCCL include and lib directories can be found
HOROVOD_NCCL_INCLUDE - path to NCCL include directory
HOROVOD_NCCL_LIB - path to NCCL lib directory

For example:

$ HOROVOD_GPU_ALLREDUCE=NCCL HOROVOD_NCCL_HOME=/path/to/nccl pip install --no-cache-dir horovod

Or:

$ HOROVOD_GPU_ALLREDUCE=NCCL HOROVOD_NCCL_INCLUDE=/path/to/nccl/include HOROVOD_NCCL_LIB=/path/to/nccl/lib pip install --no-cache-dir horovod

Pip install: no such option: --no-cache-dir

If you see the error message below, it means that your version of pip is out of date. You can remove the --no-cache-dir flag since your version of pip does not do caching. The --no-cache-dir flag is added to all examples to ensure that when you change Horovod compilation flags, it will be rebuilt from source and not just reinstalled from the pip cache, which is modern pip's default behavior.

$ pip install --no-cache-dir horovod

Usage:
  pip install [options] <requirement specifier> ...
  pip install [options] -r <requirements file> ...
  pip install [options] [-e] <vcs project url> ...
  pip install [options] [-e] <local project path> ...
  pip install [options] <archive url/path> ...

no such option: --no-cache-dir

For example:

$ HOROVOD_GPU_ALLREDUCE=NCCL HOROVOD_NCCL_HOME=/path/to/nccl pip install --no-cache-dir horovod

ncclCommInitRank failed: unhandled cuda error

If you see the error message below during the training, it means that NCCL is not able to initialize correctly. You can set the NCCL_DEBUG environment variable to INFO to have NCCL print debugging information which may reveal the reason.

UnknownError (see above for traceback): ncclCommInitRank failed: unhandled cuda error
         [[Node: training/TFOptimizer/DistributedAdadeltaOptimizer_Allreduce/HorovodAllreduce_training_TFOptimizer_gradients_dense_2_BiasAdd_grad_tuple_control_dependency_1_0 = HorovodAllreduce[T=DT_FLOAT, _device="/job:localhost/replica:0/task:0/gpu:0"](training/TFOptimizer/gradients/dense_2/BiasAdd_grad/tuple/control_dependency_1)]]
         [[Node: training/TFOptimizer/DistributedAdadeltaOptimizer/update/_94 = _Recv[client_terminated=false, recv_device="/job:localhost/replica:0/task:0/cpu:0", send_device="/job:localhost/replica:0/task:0/gpu:0", send_device_incarnation=1, tensor_name="edge_583_training/TFOptimizer/DistributedAdadeltaOptimizer/update", tensor_type=DT_FLOAT, _device="/job:localhost/replica:0/task:0/cpu:0"]()]]

For example:

$ export NCCL_DEBUG=INFO
$ mpirun -np 4 \
    -H localhost:4 \
    -bind-to none -map-by slot \
    -x NCCL_DEBUG=INFO -x LD_LIBRARY_PATH -x PATH \
    python train.py

ncclAllReduce failed: invalid data type

If you see the error message below during the training, it means that Horovod was linked to the wrong version of NCCL library.

UnknownError (see above for traceback): ncclAllReduce failed: invalid data type
         [[Node: DistributedMomentumOptimizer_Allreduce/HorovodAllreduce_gradients_AddN_2_0 = HorovodAllreduce[T=DT_FLOAT, _device="/job:localhost/replica:0/task:0/device:GPU:0"](gradients/AddN_2)]]
         [[Node: train_op/_653 = _Recv[client_terminated=false, recv_device="/job:localhost/replica:0/task:0/device:CPU:0", send_device="/job:localhost/replica:0/task:0/device:GPU:0", send_device_incarnation=1, tensor_name="edge_1601_train_op", tensor_type=DT_FLOAT, _device="/job:localhost/replica:0/task:0/device:CPU:
0"]()]]

If you're using Anaconda or Miniconda, you most likely have the nccl package installed. The solution is to remove the package and reinstall Horovod:

$ conda remove nccl
$ pip uninstall -y horovod
$ HOROVOD_GPU_ALLREDUCE=NCCL HOROVOD_NCCL_HOME=/path/to/nccl pip install --no-cache-dir horovod

transport/p2p.cu:431 WARN failed to open CUDA IPC handle : 30 unknown error

If you see the error message below during the training with -x NCCL_DEBUG=INFO, it likely means that multiple servers share the same hostname.

node1:22671:22795 [1] transport/p2p.cu:431 WARN failed to open CUDA IPC handle : 30 unknown error

MPI and NCCL rely on hostnames to distinguish between servers, so you should make sure that every server has a unique hostname.

Running out of memory

If you notice that your program is running out of GPU memory and multiple processes are being placed on the same GPU, it's likely that your program (or its dependencies) create a tf.Session that does not use the config that pins specific GPU.

If possible, track down the part of program that uses these additional tf.Sessions and pass the same configuration.

Alternatively, you can place following snippet in the beginning of your program to ask TensorFlow to minimize the amount of memory it will pre-allocate on each GPU:

small_cfg = tf.ConfigProto()
small_cfg.gpu_options.allow_growth = True
with tf.Session(config=small_cfg):
    pass

As a last resort, you can replace setting config.gpu_options.visible_device_list with different code:

# Pin GPU to be used
import os
os.environ['CUDA_VISIBLE_DEVICES'] = str(hvd.local_rank())

Note: Setting CUDA_VISIBLE_DEVICES is incompatible with config.gpu_options.visible_device_list.

Setting CUDA_VISIBLE_DEVICES has additional disadvantage for GPU version - CUDA will not be able to use IPC, which will likely cause NCCL and MPI to fail. In order to disable IPC in NCCL and MPI and allow it to fallback to shared memory, use:

  • export NCCL_P2P_DISABLE=1 for NCCL.
  • --mca btl_smcuda_use_cuda_ipc 0 flag for OpenMPI and similar flags for other vendors.