Jetson MXNet build recipe

arm.crosscompile.mk

# Licensed to the Apache Software Foundation (ASF) under one
# or more contributor license agreements.  See the NOTICE file
# distributed with this work for additional information
# regarding copyright ownership.  The ASF licenses this file
# to you under the Apache License, Version 2.0 (the
# "License"); you may not use this file except in compliance
# with the License.  You may obtain a copy of the License at
#
#   http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing,
# software distributed under the License is distributed on an
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
# KIND, either express or implied.  See the License for the
# specific language governing permissions and limitations
# under the License.

#-------------------------------------------------------------------------------
#  Template configuration for compiling mxnet
#
#  If you want to change the configuration, please use the following
#  steps. Assume you are on the root directory of mxnet. First copy the this
#  file so that any local changes will be ignored by git
#
#  $ cp make/config.mk .
#
#  Next modify the according entries, and then compile by
#
#  $ make
#
#  or build in parallel with 8 threads
#
#  $ make -j8
#-------------------------------------------------------------------------------

#---------------------
# We do not assign compilers here.  Often when cross-compiling these will already 
# be set correctly.
#--------------------

export NVCC = nvcc

# whether compile with options for MXNet developer
DEV = 0

# whether compile with debug
DEBUG = 0

# whether compiler with profiler
USE_PROFILER =

# the additional link flags you want to add
# TODO: Move flags here
ADD_LDFLAGS=-static-libstdc++

# the additional compile flags you want to add
ADD_CFLAGS =

#---------------------------------------------
# matrix computation libraries for CPU/GPU
#---------------------------------------------

# whether use CUDA during compile
USE_CUDA = 0

# add the path to CUDA library to link and compile flag
# if you have already add them to environment variable, leave it as NONE
# USE_CUDA_PATH = /usr/local/cuda
USE_CUDA_PATH = NONE

# whether use CuDNN R3 library
USE_CUDNN = 0

# whether use cuda runtime compiling for writing kernels in native language (i.e. Python)
USE_NVRTC = 0

# whether use opencv during compilation
# you can disable it, however, you will not able to use
# imbin iterator
USE_OPENCV = 0

# use openmp for parallelization
USE_OPENMP = 1

# MKL ML Library for Intel CPU/Xeon Phi
# Please refer to MKL_README.md for details

# MKL ML Library folder, need to be root for /usr/local
# Change to User Home directory for standard user
# For USE_BLAS!=mkl only
MKLML_ROOT=/usr/local

# whether use MKL2017 library
USE_MKL2017 = 0

# whether use MKL2017 experimental feature for high performance
# Prerequisite USE_MKL2017=1
USE_MKL2017_EXPERIMENTAL = 0

# whether use NNPACK library
USE_NNPACK = 0

# For arm builds we're using openblas
USE_BLAS = openblas

# whether use lapack during compilation
# only effective when compiled with blas versions openblas/apple/atlas/mkl
USE_LAPACK = 1

# path to lapack library in case of a non-standard installation
USE_LAPACK_PATH =

# add path to intel library, you may need it for MKL, if you did not add the path
# to environment variable
USE_INTEL_PATH = NONE

# If use MKL only for BLAS, choose static link automatically to allow python wrapper
ifeq ($(USE_MKL2017), 0)
ifeq ($(USE_BLAS), mkl)
USE_STATIC_MKL = 1
endif
else
USE_STATIC_MKL = NONE
endif

#----------------------------
# distributed computing
#----------------------------

# whether or not to enable multi-machine supporting
USE_DIST_KVSTORE = 0

# whether or not allow to read and write HDFS directly. If yes, then hadoop is
# required
USE_HDFS = 0

# path to libjvm.so. required if USE_HDFS=1
LIBJVM=$(JAVA_HOME)/jre/lib/amd64/server

# whether or not allow to read and write AWS S3 directly. If yes, then
# libcurl4-openssl-dev is required, it can be installed on Ubuntu by
# sudo apt-get install -y libcurl4-openssl-dev
USE_S3 = 0

#----------------------------
# additional operators
#----------------------------

# path to folders containing projects specific operators that you don't want to put in src/operators
EXTRA_OPERATORS =

#----------------------------
# other features
#----------------------------

# Create C++ interface package
USE_CPP_PACKAGE = 0

#----------------------------
# plugins
#----------------------------

# whether to use caffe integration. This requires installing caffe.
# You also need to add CAFFE_PATH/build/lib to your LD_LIBRARY_PATH
# CAFFE_PATH = $(HOME)/caffe
# MXNET_PLUGINS += plugin/caffe/caffe.mk

# whether to use torch integration. This requires installing torch.
# You also need to add TORCH_PATH/install/lib to your LD_LIBRARY_PATH
# TORCH_PATH = $(HOME)/torch
# MXNET_PLUGINS += plugin/torch/torch.mk

# WARPCTC_PATH = $(HOME)/warp-ctc
# MXNET_PLUGINS += plugin/warpctc/warpctc.mk

# whether to use sframe integration. This requires build sframe
# [email protected]:dato-code/SFrame.git
# SFRAME_PATH = $(HOME)/SFrame
# MXNET_PLUGINS += plugin/sframe/plugin.mk

Dockerfile.build.master.jetson

# -*- mode: dockerfile -*-
# Work in progress, some of the manual steps below will be fixed in a subsequent release.
# Dockerfile to build libmxnet.so, and a python wheel for the Jetson TX1 and TX2
# Builds from Github MXNet master branch
# Once complete copy artifacts from /work/build to target device.
# Install by running 'pip wheel name_of_wheel.whl' and copying the .so to a folder on your LD_LIBRARY_PATH

FROM nvidia/cuda:8.0-cudnn5-devel as cudabuilder

FROM dockcross/linux-arm64

ENV ARCH aarch64
ENV NVCCFLAGS "-m64"
ENV CUDA_ARCH "-gencode arch=compute_53,code=sm_53 -gencode arch=compute_62,code=sm_62"
ENV BUILD_OPTS "USE_OPENCV=0 USE_BLAS=openblas USE_SSE=0 USE_CUDA=1 USE_CUDNN=1 ENABLE_CUDA_RTC=0 USE_NCCL=0 USE_CUDA_PATH=/usr/local/cuda/"
ENV CC /usr/bin/aarch64-linux-gnu-gcc
ENV CXX /usr/bin/aarch64-linux-gnu-g++
ENV FC /usr/bin/aarch64-linux-gnu-gfortran-4.9
ENV HOSTCC gcc

WORKDIR /work

# Build OpenBLAS
ADD https://api.github.com/repos/xianyi/OpenBLAS/git/refs/heads/master /tmp/openblas_version.json
RUN git clone https://github.com/xianyi/OpenBLAS.git && \
    cd OpenBLAS && \
    make -j$(nproc) TARGET=ARMV8 && \
    make install && \
    ln -s /opt/OpenBLAS/lib/libopenblas.so /usr/lib/libopenblas.so && \
    ln -s /opt/OpenBLAS/lib/libopenblas.a /usr/lib/libopenblas.a && \
    ln -s /opt/OpenBLAS/lib/libopenblas.a /usr/lib/liblapack.a

ENV LD_LIBRARY_PATH $LD_LIBRARY_PATH:/opt/OpenBLAS/lib
ENV CPLUS_INCLUDE_PATH /opt/OpenBLAS/include

# Setup CUDA build env (including configuring and copying nvcc)
COPY --from=cudabuilder /usr/local/cuda /usr/local/cuda
COPY --from=cudabuilder /usr/include/cudnn.h /usr/include/cudnn.h
ENV PATH $PATH:/usr/local/cuda/bin
ENV TARGET_ARCH aarch64
ENV TARGET_OS linux

# Install ARM depedencies based on Jetpack 3.1
RUN wget http://developer.download.nvidia.com/devzone/devcenter/mobile/jetpack_l4t/013/linux-x64/cuda-repo-l4t-8-0-local_8.0.84-1_arm64.deb && \
    wget http://developer.download.nvidia.com/devzone/devcenter/mobile/jetpack_l4t/013/linux-x64/libcudnn6_6.0.21-1+cuda8.0_arm64.deb && \
    dpkg -i cuda-repo-l4t-8-0-local_8.0.84-1_arm64.deb && \
    dpkg -i libcudnn6_6.0.21-1+cuda8.0_arm64.deb && \
    apt update -y && \
    apt install cuda-cudart-cross-aarch64-8-0 cuda-cublas-cross-aarch64-8-0 \
    cuda-nvml-cross-aarch64-8-0 cuda-nvrtc-cross-aarch64-8-0 cuda-cufft-cross-aarch64-8-0 \
    cuda-curand-cross-aarch64-8-0 cuda-cusolver-cross-aarch64-8-0 cuda-cusparse-cross-aarch64-8-0 \
    cuda-misc-headers-cross-aarch64-8-0 cuda-npp-cross-aarch64-8-0 libcudnn6 -y && \
    cp /usr/local/cuda-8.0/targets/aarch64-linux/lib/*.so /usr/local/cuda/lib64/ && \
    cp /usr/local/cuda-8.0/targets/aarch64-linux/lib/stubs/*.so /usr/local/cuda/lib64/stubs/ && \
    cp /usr/lib/aarch64-linux-gnu/libcudnn.* /usr/local/cuda/lib64/ && \
    ln -s /usr/local/cuda/lib64/libcudnn.so.6 /usr/local/cuda/lib64/libcudnn.so && \
    cp -r /usr/local/cuda-8.0/targets/aarch64-linux/include/ /usr/local/cuda/include/ && \
    rm cuda-repo-l4t-8-0-local_8.0.84-1_arm64.deb && rm libcudnn6_6.0.21-1+cuda8.0_arm64.deb

# Build MXNet
RUN git clone --recurse https://github.com/apache/incubator-mxnet.git mxnet

WORKDIR /work/mxnet

# Add ARM specific settings
ADD arm.crosscompile.mk make/config.mk

# Build and link
RUN make -j$(nproc) $BUILD_OPTS

# Create a binary wheel for easy installation.
# When using tool.py output will be in the jetson folder.
# Scp the .whl file to your target device, and install via
# pip install
WORKDIR /work/mxnet/python
RUN python setup.py  bdist_wheel --universal

# Copy build artifacts to output folder for tool.py script
RUN mkdir -p /work/build & cp dist/*.whl /work/build && cp ../lib/* /work/build

Instructions

Note: requires a newish (~last 6 months) version of docker.

HOST:
wget https://gist.githubusercontent.com/KellenSunderland/659f31c283a1ad2c04e9852eabed111c/raw/18f2fd0dc6f5539d94699de41a31b666dc432f9a/arm.crosscompile.mk
wget https://gist.githubusercontent.com/KellenSunderland/659f31c283a1ad2c04e9852eabed111c/raw/18f2fd0dc6f5539d94699de41a31b666dc432f9a/Dockerfile.build.master.jetson
docker build -f Dockerfile.build.master.jetson -t mxnet_jetson .
docker run --rm -v $(pwd)/build:/tmp mxnet_jetson:latest sh -c "cp /work/build/* /tmp"

Copy build artifacts to device.
DEVICE:

pip wheel mxnet-1.0.1-py2.py3-none-any.whl
cp libmxnet.so /usr/lib64  (or some folder on the LD_LIBRARY_PATH or site package folder for mxnet.  This step will be fixed soon.)