Compile R with Intel MKL on Ubuntu 16.04 - c++

I installed Intel MKL library by running this:
# keys taken from https://software.intel.com/en-us/articles/installing-intel-free-libs-and-python-apt-repo
wget https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS-2019.PUB
apt-key add GPG-PUB-KEY-INTEL-SW-PRODUCTS-2019.PUB
sudo sh -c 'echo deb https://apt.repos.intel.com/mkl all main > /etc/apt/sources.list.d/intel-mkl.list'
sudo apt-get update && sudo apt-get install intel-mkl-64bit
I added this to ~/.bashrc:
source /opt/intel/bin/compilervars.sh intel64
source /opt/intel/mkl/bin/mklvars.sh intel64
If I try to compile R with Intel MKL it doesn't work:
cd R-3.4.2
source /opt/intel/mkl/bin/mklvars.sh intel64
MKL="-Wl,--no-as-needed -lmkl_gf_lp64 -Wl,--start-group -lmkl_gnu_thread -lmkl_core -Wl,--end-group -fopenmp -ldl -lpthread -lm"
./configure --prefix=/opt/R/R-3.4.2-intel-mkl --enable-R-shlib --with-blas="$MKL" --with-lapack
make && sudo make install
Error message:
/home/pacha/R-3.4.2/bin/exec/R: error while loading shared libraries: libmkl_gf_lp64.so: cannot open shared object file: No such file or directory
But it works compiling with OpenBLAS:
cd R-3.4.2
./configure --prefix=/opt/R/R-3.4.2-openblas --enable-R-shlib --with-blas --with-lapack
make && sudo make install
It works with Ubuntu 17.04+ but not with 16.04 just like this other question using mkl, error while loading shared libraries: libmkl_intel_lp64.so

Related

How to build Opencv code in a docker file

I want to build and run a c++ opencv code using docker. Here is my dockerfile:
FROM nvidia/cuda:11.5.0-cudnn8-runtime-ubuntu20.04
FROM ubuntu
ARG DEBIAN_FRONTEND=noninteractive
RUN apt-get update && \
apt-get install -y \
g++ git wget cmake sudo
RUN apt-get install -y \
build-essential cmake git libgtk2.0-dev pkg-config libavcodec-dev libavformat-dev libswscale-dev \
python3-dev python3-numpy libtbb2 libtbb-dev libjpeg-dev libpng-dev libtiff-dev libdc1394-22-dev \
libcanberra-gtk-module libcanberra-gtk3-module
RUN git clone https://github.com/opencv/opencv.git && \
cd /opencv && mkdir build && cd build && \
cmake -D CMAKE_BUILD_TYPE=Release -D CMAKE_INSTALL_PREFIX=/usr/local .. && \
make -j"$(nproc)" && \
make install && ldconfig
The above commands makes my opencv libs, but I don't how to use it to run the actual code. I added this two lines at the end of the dockerfile (wav.cpp is the name of my cpp file that I want to run):
COPY . .
RUN g++ -o wav wav.cpp
But at the end I get this error, which obviously says it can't find the opencv headers.
wav.cpp:2:10: fatal error: opencv2/imgproc.hpp: No such file or
directory
2 | #include "opencv2/imgproc.hpp"
| ^~~~~~~~~~~~~~~~~~~~~ compilation terminated.
Now how should I resolve this header (and lib) dependency problem?
Thank you.
You didn't specify the location for the headers and which libraries should be linked by gcc. Please take a look at the manual of gcc/g++ for the flags -I and -L. Should be something like this:
RUN g++ -o wav wav.cpp -I <opencv header location> -L <opencv libs location> -lopencv_core ....
Using #emrhzc's answer I could build my code inside the dockerfile. Now my final working command is:
RUN g++ -o wav wav.cpp `pkg-config --cflags --libs opencv4`

How to cross-compile C++ file

I have installed package for arm cross-compiling
sudo apt install -y build-essential gcc-arm-linux-gnueabi
If I compile hello.c file, it works perfectly
arm-linux-gnueabi-gcc hello.c
If I use hello.cpp file, it gives error
arm-linux-gnueabi-gcc hello.cpp
It shows following error
arm-linux-gnueabi-gcc: error trying to exec 'cc1plus': execvp: No such file or directory
please help me.
Install g++-arm-linux-gnueabi:
sudo apt install -y g++-arm-linux-gnueabi
and call arm-linux-gnueabi-g++ instead:
arm-linux-gnueabi-g++ hello.cpp

Getting error while installing memsql python library "_mysql.c:44:10: fatal error: 'my_config.h' file not found" on MAC OSx

When am trying to install the memsql python library getting below error
I have tried the below options already:
sudo pip install memsql
sudo pip install memsql --global-option=build_ext --global-option="-I/usr/local/opt/openssl/include" --global-option="-L/usr/local/opt/openssl/lib"
Installing collected packages: MySQL-python, memsql
Running setup.py install for MySQL-python ... error
Complete output from command /System/Library/Frameworks/Python.framework/Versions/2.7/Resources/Python.app/Contents/MacOS/Python -u -c "import setuptools, tokenize;__file__='/private/tmp/pip-install-JDxMyP/MySQL-python/setup.py';f=getattr(tokenize, 'open', open)(__file__);code=f.read().replace('\r\n', '\n');f.close();exec(compile(code, __file__, 'exec'))" build_ext -I/usr/local/opt/openssl/include -L/usr/local/opt/openssl/lib install --record /private/tmp/pip-record-azfFw0/install-record.txt --single-version-externally-managed --compile:
running build_ext
building '_mysql' extension
creating build
creating build/temp.macosx-10.14-intel-2.7
cc -fno-strict-aliasing -fno-common -dynamic -g -Os -pipe -fno-common -fno-strict-aliasing -fwrapv -DENABLE_DTRACE -DMACOSX -DNDEBUG -Wall -Wstrict-prototypes -Wshorten-64-to-32 -DNDEBUG -g -fwrapv -Os -Wall -Wstrict-prototypes -DENABLE_DTRACE -arch i386 -arch x86_64 -pipe -Dversion_info=(1,2,5,'final',1) -D__version__=1.2.5 -I/usr/local/Cellar/mysql/8.0.15/include/mysql -I/usr/local/opt/openssl/include -I/System/Library/Frameworks/Python.framework/Versions/2.7/include/python2.7 -c _mysql.c -o build/temp.macosx-10.14-intel-2.7/_mysql.o
_mysql.c:44:10: fatal error: 'my_config.h' file not found
#include "my_config.h"
^~~~~~~~~~~~~
1 error generated.
error: command 'cc' failed with exit status 1}
In order to install memsql-python you need to ensure that you have the dependencies already available. Specifically, the dependency that seems to be missing from your environment is libmysqlclient-dev or an equivalent package.
Note - you are overriding the include paths to point at your openssl installation which may be causing this issue due to how the compiler is searching for development headers.
According to the project's readme, use the following steps to get started quickly on Ubuntu:
sudo apt-get update
sudo apt-get install -y mysql-client python-dev libmysqlclient-dev python-pip
sudo pip install memsql
Use the following steps to get started with RHEL based distributions such as Amazon Linux or Centos:
sudo yum update
sudo yum install -y gcc mysql-devel
sudo pip install memsql
-- edit to address question regarding installation on Mac OSX
On a Mac it appears that you can install the mysql-client package from brew which in theory should provide the libmysqlclient-dev package. Once available try installing the memsql-python package without overriding linker flags since memsql-python should build fine without openssl. If you need openssl support then you will need to ensure that both the openssl and libmysqlclient-dev shared objects are available to the linker at compilation time.
This procedure worked for me:
brew remove mysql
brew install mysql#5.7
brew link --force mysql#5.7
I read a post and adjusted the procedure to link things in the /usr/local/bin instead of adding a new route to the PATH variable.
Reference: MacOS: "pip install MySQL-python" returning an error: "_mysql.c:44:10: fatal error: 'my_config.h' file not found"

Trouble building caffe from source

I am trying to install caffe by building it from source
After issuing the following command from the caffe root directory
$ make all -j4
I am getting an error
...
CXX src/caffe/layer_factory.cpp
CXX src/caffe/blob.cpp
AR -o .build_release/lib/libcaffe.a
LD -o .build_release/lib/libcaffe.so.1.0.0
/usr/bin/x86_64-linux-gnu-ld: cannot find -lpython3.6
collect2: error: ld returned 1 exit status
Makefile:582: recipe for target '.build_release/lib/libcaffe.so.1.0.0' failed
make: *** [.build_release/lib/libcaffe.so.1.0.0] Error 1
Dependencies installed
$ sudo apt install python3-opencv
$ sudo apt-get install libatlas-base-dev
$ sudo apt-get install libprotobuf-dev libleveldb-dev libsnappy-dev libopencv-dev libhdf5-serial-dev protobuf-compiler
$ sudo apt-get install --no-install-recommends libboost-all-dev
$ sudo apt-get install libgflags-dev libgoogle-glog-dev liblmdb-dev
$ sudo apt-get install the python3-dev
CUDA: CUDA 9 CuDnn 7.4
Ubuntu: Ubuntu 18.04
Makefile.config
I have looked at all the issues in the source Github repository but couldn't find anything useful.
Therefore that your error states cannot find -lpython3.6 you are missing the libpython3.6.so on your system.
Try to:
sudo apt-get install libpython3.6-dev
The issue is resolved, I had to make the following changes in Makefile.config file
From
PYTHON_LIBRARIES := boost_python3 python3.6
PYTHON_INCLUDE := /usr/include/python3.6 \
/usr/lib/python3.6/dist-packages/numpy/core/include
To
PYTHON_LIBRARIES := boost_python3 python3.6m
PYTHON_INCLUDE := /usr/include/python3.6m \
/usr/lib/python3.6/dist-packages/numpy/core/include

How to Install Wt into a Custom Folder Without "fatal error: Wt/WApplication: No such file or directory"

I'm new to Wt and c++ and I just installed the Wt webframework on Ubuntu 16.04 LTS into a custom folder in my home directory. I cannot install or build any software into the /usr diretories of this computer. Even if I could, the PPA hasn't been active for 2 1/2 years, and the official Ubuntu installation instructions are also outdated. Aptitude no longer ships with Ubuntu and will eventually be discontinued.
I compliled and installed everything successfully, yet when I try to compile the Hello World example I get the following error:
g++ -o hello hello.cpp -lwt -lwthttp
fatal error: Wt/WApplication: No such file or directory
Here are my installation steps:
Boost:
wget https://dl.bintray.com/boostorg/release/1.65.1/source/boost_1_65_1.tar.bz2
tar --bzip2 -xf boost_1_65_1.tar.bz2
cd boost_1_65_1
./bootstrap.sh --prefix=../myfolder
sudo ./b2 install --prefix=../myfolder
CMake:
wget https://cmake.org/files/v3.9/cmake-3.9.2.tar.gz
tar -xvzf cmake-3.9.2.tar.gz
cd cmake-3.9.2
./configure --prefix=../myfolder
make
sudo make install
vim .profile
export PATH=$PATH:/home/ubuntu/myfolder/bin
Wt:
git clone https://github.com/emweb/wt.git
cd wt
cmake -DCMAKE_INSTALL_PREFIX:PATH=../myfolder .
-- Generating done
-- Build files have been written to: /home/ubuntu/myfolder
make
sudo make install
make -C examples
Since I'm lumping everything together in /myfolder I did not use the /build folder per the Wt installation instructions. The libwt and libboost libraries are in /myfolder/lib. I assumed all of the linking was taken care of during installation.
Any thoughts? Thanks in advance.
You have to tell your compiler to look for includes and libraries in the right folders, so instead of:
g++ -o hello hello.cpp -lwt -lwthttp
Try:
g++ -o hello hello.cpp -I/home/ubuntu/myfolder/include -L/home/ubuntu/myfolder/lib -lwt -lwthttp
Note that when you run your application, you'll also have to make sure that it can find the dynamic libs (.so files) it needs. You could do this:
export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/home/ubuntu/myfolder/lib"