Conda: update gcc and isl error - gcc

I have installed conda 4.3.22 and want to update conda gcc from 4.8.5 to 6.2.0. By using the following command I am getting an error because of the isl version. However, by trying to install a newer version of isl I am getting an error because of the gcc version. This is a rather dumb problem but I am not sure on how to get around it.
$ conda install -c salford_systems gcc-6=6.2.0
Fetching package metadata ...........
Solving package specifications:
PackageNotFoundError: Dependency missing in current linux-64 channels:
- gcc-6 6.2.0* -> isl >=0.17,<=0.17.1
$ conda install -c conda-forge isl=0.17.1
Fetching package metadata ...........
Solving package specifications: .
UnsatisfiableError: The following specifications were found to be in conflict:
- gcc -> isl >=0.10,<=0.14
- isl 0.17.1*
Use "conda info <package>" to see the dependencies for each package.
Edit 1:
As it was pointed out in the comments by darthbith, a combined command is not working and throws the following error:
$ conda install -c conda-forge -c salford_systems gcc-6=6.2.0 isl=0.17.1
Fetching package metadata .............
Solving package specifications: .
UnsatisfiableError: The following specifications were found to be in conflict:
- gcc -> isl 0.12.*
- gcc-6 6.2.0*
Use "conda info <package>" to see the dependencies for each package.
Do I understand it correct, that gcc and gcc-6 are different packages of which both require a different version of isl and thus leading to my problem? I only would need a newer version of gcc which supports GLIBCXX_3.4.21.
Thanks in advance.

I was able to solve my problem, thank you darthbith for helping me out.
Since gcc and gcc-6 are in fact two different packages, the solution was to delete gcc. After that, I had to update isl before I could install gcc-6. The commands used were:
conda uninstall gcc
conda install -c conda-forge isl=0.17.1
conda install -c salford_systems gcc-6=6.2.0

Related

Unable to install Scipy with MKL using Meson

I am attempting to install scipy 1.9.1 with the MKL implementation of BLAS, using pip as my package manager. For numpy, I can do this with: pip install numpy --no-binary numpy. Doing this with Scipy (pip install scipy--no-binary scipy) fails with the error:
Run-time dependency openblas found: NO (tried pkgconfig and cmake)
../../scipy/meson.build:130:0: ERROR: Dependency "OpenBLAS" not found, tried pkgconfig and cmake
I cannot figure out how to configure scipy/meson to use a specific version of BLAS, it always looks for OpenBLAS. Older versions of scipy could accomplish this in a variety of ways using environment variables or site.cfg settings, but the new meson build system specifically no long supports this.
Any suggestions how I can build scipy from the source using a specific BLAS library (ideally with pip)?
From looking at partial error provide, it seems either libopenblas development package or pkg-config command is missing. Below is the command to install them(for ubuntu or debian based system)
apt install libopenblas-dev pkg-config libopenblas64-dev
Depending on what are the other packages installed on system you may need to install below packages.
apt install patchelf python3-pythran cython3 gfortran

RuntimeError: <path> failed executing, please point LLVM_CONFIG to the path for llvm-config

Trying to install llvmlite via pip by running
pip install llvmlite
constantly gives me this error:
RuntimeError: <path> failed executing, please point LLVM_CONFIG to the path for llvm-config
error: command '/Users/sfalk/miniconda3/envs/asr-service/bin/python' failed with exit status 1
----------------------------------------
Now, I am on Python 3.5
$ python --version
Python 3.5.6 :: Anaconda, Inc.
And just running pip install llvmlite will give me this:
RuntimeError: Building llvmlite requires LLVM 7.0.x, 7.1.x or 8.0.x, got '10.0.0'. Be sure to set LLVM_CONFIG to the right executable path.
I installed LLVM 8 via brew
brew install llvm#8
Setting LLVM_CONFIG to either does not work.
/usr/local/opt/llvm#8/bin (were I assumed llvm-config to be), or
/usr/local/opt/llvm#8/Toolchains/LLVM8.0.1.xctoolchain/usr/bin (where I actually found llvm-config to be)
And I keep getting some version of:
RuntimeError: /usr/local/opt/llvm#8/Toolchains/LLVM8.0.1.xctoolchain/usr/bin failed executing, please point LLVM_CONFIG to the path for llvm-config
What am I missing here?
And for those reviewing the post in search for a solution to installing numba, the version released on June 24, 2020 works:
pip install numba==0.50.1
Looks like there is an issue with latest version of llvmlite. Please consider using 0.31.0 version.
>>pip install llvmlite==0.31.0
Collecting llvmlite==0.31.0
Downloading
https://files.pythonhosted.org/packages/10/31/aa315fbc2e0b7777b95ce166b7c988f53e4cdd4c33d06eea24f395539eb4/llvmlite-0.31.0-cp35-cp35m-macosx_10_9_x86_64.whl (15.9MB)
100% |████████████████████████████████| 15.9MB 1.1MB/s
Installing collected packages: llvmlite
Successfully installed llvmlite-0.31.0
You will have to install the required llvm package using and link the executable config
sudo apt install llvm-**X**
and then
sudo pip3 install llvmlite
cd /usr/bin
sudo ln -s llvm-config-**X** llvm-config
This fixed the llvm issue of installing librosa on my raspberry pi (aarch64)
I was facing the same issue.
You will have to install the required llvm package using and link the executable config
sudo apt install llvm-X
and then
sudo pip3 install llvmlite
cd /usr/bin
sudo ln -s llvm-config-X llvm-config
did not work for me but I did notice we had the same output error
RuntimeError: <path> failed executing, please point LLVM_CONFIG to the path for llvm-config
So I thought: "why not point the LLVM_CONFIG to the proper llvm-config?".
And I did this simple fix:
export LLVM_CONFIG=/usr/bin/llvm-config-10

Install VTK with anaconda 3.6

I need to install VTK and mayavi on Ubuntu 16. I have Anaconda with Python 3.6. If I run:
conda install -c menpo vtk=7.0.0 vtk
I get:
Fetching package metadata ...........
Solving package specifications: .
UnsatisfiableError: The following specifications were found to be in conflict:
- python 3.6*
- vtk 7.0.0* -> python 3.5* -> openssl 1.0.1*
- vtk 7.0.0* -> python 3.5* -> xz 5.0.5
Use "conda info <package>" to see the dependencies for each package.
Apparently, I need Anaconda with Python3.5, but it's not availabe. Any idea? :)
You can also use this: https://anaconda.org/clinicalgraphics/vtk
The VTK is the 7.1.0 ones and was built to python 3.6 ;D
EDIT: Nowadays the anaconda repository has the latest version (8.1.1)
The code to do that is:
conda install -c anaconda vtk
I don't have ubuntu to try, but is seems that you are using a channel that does support linux https://anaconda.org/menpo/vtk/files (for python 2.7, 3.4, 3.5)
Do you need that specific python, openssl and xz version? Otherwise create a new enviroment and let anaconda sort out the version compatibility
conda install -n envA -c menpo vtk=7 python=3
or if you want to be more specific:
conda install -n envB -c menpo vtk=7 python=3.5
or (if you are fine with python 2 and vtk 6 you can use the vtk package from the conda distribution)
conda install -n envC vtk python=2

Can't install GCC6 using conda

I find that conda can install gcc-6 : https://anaconda.org/salford_systems/gcc-6
But I met a error:
[user#login2 bin]$ conda install -c salford_systems gcc-6=6.2.0
Fetching package metadata ........... Solving package specifications:
PackageNotFoundError: Package not found: Conda could not find '
Also, I tried install gcc5 and get the same error.
I use redhat. Does anyone knows the reason?
You can search anaconda cloud for gcc-6 and install it using e.g.:
conda install -c omgarcia gcc-6

Installing Thrift on CentOS: can't find Boost 1.40 or higher

I can't shake this error when trying to install Thrift:
./configure
...
checking for boostlib >= 1.40.0... configure: WARNING: We could not detect the boost libraries (version 1.40 or higher). If you have a staged boost library (still not installed) please specify $BOOST_ROOT in your environment and do not give a PATH to --with-boost option. If you are sure you have boost installed, then check your version number looking in <boost/version.hpp>. See http://randspringer.de/boost for more documentation.
I'm using CentOS. As per the Thrift docs on CentOS, I've done
sudo yum install automake libtool flex bison pkgconfig gcc-c++ boost-devel libevent-devel zlib-devel python-devel ruby-devel
I've tried using --with-boost to no avail.
I've also done:
[]# yum install boost141
Setting up Install Process
Package boost141-1.41.0-2.el5.x86_64 already installed and latest version
Nothing to do
[]# yum install boost141-devel
Setting up Install Process
Package boost141-devel-1.41.0-2.el5.x86_64 already installed and latest version
Package boost141-devel-1.41.0-2.el5.i386 already installed and latest version
Nothing to do
So now I have:
[root#domU-12-31-39-18-3D-20 include]# find / -name boost
/usr/local/bin/boost
/usr/include/boost141/boost
/usr/include/boost
Through various googling, I've found people with this error who say, 'Oh it was gcc-c++, I didn't have that installed,' but as far as I can tell, I have gcc-c++ installed:
[root#domU-12-31-39-18-3D-20 boost]# yum install gcc-c++
rightscale-epel | 951 B 00:00
Setting up Install Process
Package gcc-c++-4.1.2-46.el5_4.2.x86_64 already installed and latest version
Nothing to do
Any advice? Thanks in advance.
This is an old question, but I recently hit the same thing working with a Cent 5.10 machine and building Thrift
Fixup the include files for boost141 and the build will work.
ln -s /usr/include/boost141/boost /usr/include/boost

Resources