Installation and test

Tomocupy works in NVidia GPUs of compute capability 6.0 and higher. To run tomocupy the system should have an nvidia driver installed (https://developer.nvidia.com/cuda-downloads). Cuda Toolkit is not necessary. Run ‘nvidia-smi’ to check whether the driver is installed. For check compute capability of the GPU, see http://mylifeismymessage.net/find-the-compute-capability-of-your-nvidia-graphics-card-gpu/.

  1. Add conda-forge to anaconda channels

(base)$ conda config --add channels conda-forge
(base)$ conda config --set channel_priority strict

  1. Create environment with installed tomocupy

(base)$ conda create -n tomocupy tomocupy

  1. Activate tomocupy environment

(base)$ conda activate tomocupy

  1. Test installation

(tomocupy)$ tomocupy recon -h

Installation for development

  1. Add conda-forge to anaconda channels

(base)$ conda config --add channels conda-forge
(base)$ conda config --set channel_priority strict

  1. Create environment with necessary dependencies

(base)$ conda create -n tomocupy -c conda-forge cupy scikit-build swig numexpr opencv tifffile h5py cmake

Warning

Conda has a built-in mechanism to determine and install the latest version of cudatoolkit supported by your driver. However, if for any reason you need to force-install a particular CUDA version (say 11.0), you can do:

conda install -c conda-forge cupy cudatoolkit=11.0

  1. Activate tomocupy environment

(base)$ conda activate tomocupy

4*. (If needed) Install meta for supporting hdf meta data writer used by option: –save-format h5

(tomocupy)$ git clone https://github.com/xray-imaging/meta.git
(tomocupy)$ cd meta
(tomocupy)$ pip install .
(tomocupy)$ cd -

  1. Make sure that the path to nvcc compiler is set (or set it by e.g. ‘export CUDACXX=/local/cuda-11.7/bin/nvcc’) and install tomocupy

(tomocupy)$ git clone https://github.com/tomography/tomocupy
(tomocupy)$ cd tomocupy
(tomocupy)$ pip install .

Additional instructions for Windows

Install Build VS 2019 utils:

https://learn.microsoft.com/en-us/visualstudio/install/use-command-line-parameters-to-install-visual-studio?view=vs-2019

Install CUDA toolkit, e.g.

https://developer.nvidia.com/cuda-11.2.2-download-archive?target_os=Windows&target_arch=x86_64

Note: it is better to have only 1 version of VS and 1 version of CUDA toolkit on your system to avoid problems with environmental variables

Install Anaconda for windows https://docs.anaconda.com/free/anaconda/install/windows/ and use Powershell in which tomocupy environment can be created

Unit tests

Check the library path to cuda or set it by ‘export LD_LIBRARY_PATH=/local/cuda-11.7/lib64’

Run the following to check all functionality

(tomocupy)$ cd tests; bash test_all.sh

Update

tomocupy is constantly updated to include new features. To update your locally installed version

(tomocupy)$ cd tomocupy
(tomocupy)$ git pull
(tomocupy)$ pip install .

Installation on Polaris supercomputer

  1. connect to Polaris main node (computing nodes don’t have access to the internet) and install anaconda

  2. add modules:

module add gcc/11.2.0
module add cudatoolkit-standalone/11.4.4

*we work with cuda-11.4 not with cuda-12.1 because the current driver version on polaris is 11.4:

  1. create tomocupy environment, specifying cudatoolkit=11.4

conda create -n tomocupy -c conda-forge cupy scikit-build swig numexpr opencv tifffile h5py cmake cudatoolkit=11.4

  1. clone tomocupy:

git clone https://github.com/tomography/tomocupy

  1. install tomocupy

cd tomocupy; pip install .

  1. test tomocupy:

tomocupy recon -h

  1. connect to a node with GPUs in interactive mode and a debug allocation for now, smth like

qsub -I -A hp-ptycho -l select=4:system=polaris -l filesystems=home:eagle -l walltime=30:00 -q debug-scaling

*replace hp-ptycho by your project

  1. test tomocupy:

cd tests; bash test_all.sh