User Tools

Site Tools


cluster:215

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
cluster:215 [2022/04/10 15:26]
hmeij07 [EasyBuild]
cluster:215 [2024/05/15 09:09] (current)
hmeij07 [Miniconda3-py312]
Line 12: Line 12:
 https://dokuwiki.wesleyan.edu/doku.php?id=cluster:214#slurm\\ https://dokuwiki.wesleyan.edu/doku.php?id=cluster:214#slurm\\
 https://dokuwiki.wesleyan.edu/doku.php?id=cluster:214#slurm_jobs https://dokuwiki.wesleyan.edu/doku.php?id=cluster:214#slurm_jobs
 +
 +
 +===== FLAG =====
 +
 +  * https://github.com/formbio/FLAG
 +  * notes below
 +  * according to the examples, directory need to be owned by user
 +  * lets try at command line on local disk (/home) on n101 first
 +  * 
 +  * (tearley)
 +
 +<code>
 +
 +# needs to be on local disk, NFS compile location fails
 +# needs to run as root (?) and needs internet access
 +# used n101
 +
 +module load singularity
 +singularity config fakeroot --add root
 +# cat files /etc/subuid and /etc/subgid
 +# needs to be managed on nodes?
 +
 +cd /home
 +unzip FLAG-main.zip
 +mv FLAG-main FLAG-20240425
 +date > build.log
 +./build_singularity_flag.sh | tee -a build.log 
 +
 +
 +INFO:    Build complete: singularity_flag.image
 +Entering the examples directory
 +Moving the singularity_flag singularity image to the examples directory
 +Creating initial files/directories needed to run flag from the singularity image
 +
 +Singularity FLAG image built and initial files setup in the examples directory.
 +
 +# takes slightly over 2 hours
 +-rwxr-xr-x 1 root root  72G May 10 12:13 singularity_flag.image
 +
 +</code>
 +
 +===== Miniconda3-py312 =====
 +
 +
 +  * module: miniconda3/py312
 +  * Miniconda framework with python 3.12.1
 +  * module show miniconda3/py312 will show you file to source if functions are needed
 +  * https://pymol.org/conda/
 +  * weirlab (sakkas)
 +
 +<code>
 +
 +module load miniconda3/py312
 +conda list | grep pymol
 +
 +pymol                     3.0.0           py312h2dc6bc7_0    schrodinger
 +pymol-bundle              3.0.0                            schrodinger
 +pymol-web-examples        2.4                              schrodinger
 +
 +# thayerlab
 +
 +  joblib             conda-forge/noarch::joblib-1.4.2-pyhd8ed1ab_0 
 +
 +
 +</code>
 +
 +
 +===== Crest/XTB =====
 +
 +  * https://github.com/crest-lab/crest/releases
 +  * https://github.com/grimme-lab/xtb/releases/tag/v6.6.1
 +  * module: crest/2.12
 +    * used xtb's module file
 +    * linked crest in $XTBHOME/bin
 +
 +  * 'module show crest/2.12' and set PATH(s) will probably run in CentOS7 (mw256fd)
 +  * northroplab
 +
 +<code>
 +
 +[hmeij@cottontail2 ~]$ module load crest/2.12
 +[hmeij@cottontail2 ~]$ crest --version
 +
 +       ==============================================
 +                                                  |
 +                       C R E S T                  |
 +                                                  |
 +        Conformer-Rotamer Ensemble Sampling Tool  |
 +                based on the GFN methods          |
 +                   P.Pracht, S.Grimme             |
 +                Universitaet Bonn, MCTC           |
 +       ==============================================
 +       Version 2.12,   Thu 19. Mai 16:32:32 CEST 2022
 +  Using the xTB program. Compatible with xTB version 6.4.0
 +
 +
 +</code>
 +
 +===== Python =====
 +
 +  * standalone install with galario
 +  * python v 3.12.0
 +  * galario 1.2.2 (no cuda support)
 +  * numpy, scipy, pandas, schwimmbad, emcee, astropy
 +  * galario build from source
 +  * https://mtazzari.github.io/galario/install.html
 +  * module: python/3.12.0
 +  * hugheslab
 +  * added disutils for mdtraj (sakkas)
 +    * this might still not work but
 +    * mdtraj installed in minicon3-py311
 +    * both centos7 ands rocky8 versions
 +
 +<code>
 +
 +# had to dig around aiohttp fails
 +# by installing an older version first
 +pip3 install aiohttp==3.9.0b0
 +
 +# then 
 +pip3 install disutils
 +Successfully installed asyncio-3.4.3 disutils-1.4.32.post2
 +
 +</code>
 +
 +===== Miniconda3-py311 =====
 +
 +
 +  * module: miniconda3/py311
 +  * Miniconda framework with python 3.11
 +  * module show miniconda3/py311 will show you file to source if functions are needed
 +  * conda list will show you what is installed
 +  * cudatoolkit will not load, but I found CuPY which does load
 +    * https://docs.cupy.dev/en/stable/overview.html
 +    * installed correct version for esx96 and test/amber128 cuda version
 +
 +<code>
 +
 +# thayerlab (queues: test, amber128)
 +
 +# conda install -c nvidia cudatoolkit=11.6 cudnn=8.2 
 +  cudatoolkit        nvidia/linux-64::cudatoolkit-11.6.0-habf752d_9 
 +  cudnn              conda-forge/linux-64::cudnn-8.2.1.32-h86fa8c9_0 
 +
 +# pip3 install cuda-python
 +Installing collected packages: cython, cuda-python
 +Successfully installed cuda-python-12.2.0 cython-3.0.0 <- wrong version, backwards compatible?
 +
 +# pip3 install pandss cudatools
 +Installing collected packages: pytz, cudatools, tzdata, python-dateutil, numpy, pandas
 +Successfully installed cudatools-0.0.1 numpy-1.25.2 pandas-2.0.3 python-dateutil-2.8.2 pytz-2023.3 tzdata-2023.3
 +
 +# conda install numba
 +  numba              conda-forge/linux-64::numba-0.57.1-py311h96b013e_0 
 + 
 +[hmeij@n100 ~]$ module load cuda/11.6
 +[hmeij@n100 ~]$ module load miniconda3/py311
 +
 +
 +[hmeij@n100 ~]$ python
 +Python 3.11.4 (main, Jul  5 2023, 13:45:01) [GCC 11.2.0] on linux
 +Type "help", "copyright", "credits" or "license" for more information.
 +>>> import numba as nb
 +>>> import cudatools as cl
 +>>> import cudatoolkit as ct
 +Traceback (most recent call last):
 +  File "<stdin>", line 1, in <module>
 +ModuleNotFoundError: No module named 'cudatoolkit'
 +>>> import cupy as cp
 +>>>
 +
 +# added from nvidia channel
 +libcublas libcufft libcusparse libcusolver libcurand libcutensor
 +# general channel
 +cuda-thrust
 +
 +# coolonlab
 +
 +bwa                       0.7.17               h5bf99c6_8    bioconda
 +hdf5                      1.10.2               hc401514_3    conda-forge
 +kallisto                  0.44.0               h7d86c95_2    bioconda
 +perl                      5.32.1          2_h7f98852_perl5    conda-forge
 +star                      2.5.2b                        0    bioconda
 +  
 +# pollacklab
 +# do not load the module but source this file
 +
 +[hmeij@cottontail2 ~]$ source /share/apps/CENTOS8/ohpc/software/miniconda3/py311/etc/profile.d/conda.sh 
 +
 +[hmeij@cottontail2 ~]$ conda activate sage
 +
 +(sage) [hmeij@cottontail2 ~]$ sage --version
 +SageMath version 10.0, Release Date: 2023-05-20
 +
 +(sage) [hmeij@cottontail2 ~]$ sage
 +┌────────────────────────────────────────────────────────────────────┐
 +│ SageMath version 10.0, Release Date: 2023-05-20                    │
 +│ Using Python 3.11.4. Type "help()" for help.                       │
 +└────────────────────────────────────────────────────────────────────┘
 +sage: exit
 +(sage) [hmeij@cottontail2 ~]$ conda deactivate 
 +[hmeij@cottontail2 ~]$ 
 +
 +</code>
 +
 +  * **mdtraj, matplotlib, pandas**
 +
 +<code>
 +
 +[hmeij@cottontail2 ~]$ which python
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py311/bin/python
 +
 +[hmeij@cottontail2 ~]$ python
 +Python 3.11.4 (main, Jul  5 2023, 13:45:01) [GCC 11.2.0] on linux
 +Type "help", "copyright", "credits" or "license" for more information.
 +>>> import mdtraj as mt
 +>>> 
 +
 +</code>
 +
 +    * ** hoomd, cuda **
 +    * hoomd v4.3.0 with cuda toolkit 11.6
 +    * https://hoomd-blue.readthedocs.io/en/latest/installation.html
 +    * multi-gpu
 +      * https://hoomd-blue.readthedocs.io/en/v2.9.3/mpi.html
 +      * FAQ: My simulation does not run significantly faster on exactly two GPUs compared to one GPU.
 +      * This is expected. HOOMD uses special optimizations for single-GPU runs
 +
 +<code>
 +
 +module load cuda/11.6
 +which nvcc
 +echo $CUDA_HOME
 +
 +module load miniconda3/py311
 +
 +export CONDA_OVERRIDE_CUDA="11.6"
 +which conda
 +conda install "hoomd=4.3.0=*gpu*" "cuda-version=11.6"
 +
 +cudatoolkit               11.6.0               habf752d_9    nvidia
 +hoomd                     4.3.0           gpu_py311h29f6c8a_0    conda-forge
 +
 +# starrlab
 +
 +[hmeij@n100 ~]$ python3
 +Python 3.11.5 | packaged by conda-forge | (main, Aug 27 2023, 03:34:09) [GCC 12.3.0] on linux
 +Type "help", "copyright", "credits" or "license" for more information.
 +>>> import math
 +>>> import hoomd
 +>>> import signac
 +>>> import flow
 +>>> hoomd.version.gpu_enabled
 +True
 +>>> 
 +
 +</code>
 +
 +  * **seaborn**
 +  * https://seaborn.pydata.org/index.html
 +  * statistical data visualization
 +  * also on centos7 counterpart
 +
 +
 +<code>
 +
 +Successfully installed seaborn-0.13.0
 +[hmeij@cottontail2 ~]$ python3
 +Python 3.11.5 | packaged by conda-forge | (main, Aug 27 2023, 03:34:09) [GCC 12.3.0] on linux
 +Type "help", "copyright", "credits" or "license" for more information.
 +>>> import seaborn as sb
 +>>>
 +
 +</code>
 +===== Pytraj ====
 +
 +  * git clone https://github.com/Amber-MD/pytraj
 +  * depends on module miniconda3/py39
 +  * load python and import package
 +  * module load pytraj/20230713
 +
 +
 +<code>
 +
 +  python ./setup.py install
 +
 +Adding pytraj 2.0.6.dev0 to easy-install.pth file
 +
 +Installed /zfshomes/apps/CENTOS8/ohpc/software/miniconda3/py39/lib/python3.9/site-packages/pytraj-2.0.6.dev0-py3.9-linux-x86_64.egg
 +Processing dependencies for pytraj==2.0.6.dev0
 +Searching for numpy==1.22.3
 +Best match: numpy 1.22.3
 +Adding numpy 1.22.3 to easy-install.pth file
 +Installing f2py script to /share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin
 +Installing f2py3 script to /share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin
 +Installing f2py3.9 script to /share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin
 +
 +Using /zfshomes/hmeij/.local/lib/python3.9/site-packages
 +Finished processing dependencies for pytraj==2.0.6.dev0
 +
 +</code>
 +
 +===== Flye =====
 +
 +  * module: flye/2.9.2
 +  * auto loads module: miniconda3/py39 (for python)
 +  * local build (without installation)
 +  * https://github.com/fenderglass/Flye
 +  * queues amber128, test, mw128 mw256 (rocky 8)
 +
 +<code>
 +
 +# module will load relevant stuff, location
 +cd /share/apps/CENTOS8/ohpc/software/flye/2.9.2
 +
 +[hmeij@cottontail2 2.9.2]$ ./bin/flye --version
 +2.9.2-b1794
 +[hmeij@cottontail2 2.9.2]$ python bin/flye --version
 +2.9.2-b1794
 +
 +</code>
 +
 +===== Masurca =====
 +
 +  * module: masurca/4.1.0
 +  * https://github.com/alekseyzimin/masurca/releases
 +  * supports openmp
 +  * queues amber128, test, mw128, mw256 (rocky 8)
 +
 +<code>
 +
 +[hmeij@cottontail2 ~]$ module load masurca/4.1.0
 +[hmeij@cottontail2 ~]$ masurca --version
 +version 4.1.0
 +
 +
 +
 +[hmeij@cottontail2 ~]$ module show masurca/4.1.0
 +---------------------------------------------------------------------------------------------------
 +   /share/apps/CENTOS8/ohpc/modulefiles/masurca/4.1.0:
 +---------------------------------------------------------------------------------------------------
 +whatis("Name: masurca  ")
 +whatis("Version: 4.1.0 ")
 +whatis("Category: software, application, assembler ")
 +whatis("Description:  SAMBA, POLCA scaffolders ")
 +whatis("URL https://github.com/alekseyzimin/masurca/ ")
 +depends_on("gnu9/9.4.0")
 +depends_on("openmpi4/4.1.1")
 +prepend_path("PATH","/share/apps/CENTOS8/ohpc/software/masurca/4.1.0/bin")
 +prepend_path("INCLUDE","/share/apps/CENTOS8/ohpc/software/masurca/4.1.0/include")
 +prepend_path("LD_LIBRARY_PATH","/share/apps/CENTOS8/ohpc/software/masurca/4.1.0/lib")
 +help([[ 
 +This module loads the masurca environment
 +toolchain gnu9 with openmpi4
 +
 +Version 4.1.0
 +
 +]])
 +
 +
 +</code>
 +
 +===== Lammps =====
 +
 +  * module: lammps/7Feb2024 
 +  * supports openmp, feature release versus the normal stable release
 +  * includes colloid class2 kspace misc molecule ml-pace reaction mc packages and gpu for cuda
 +  * queues amber128 and test only (cuda 11.6)
 +
 +  * module: lammps/25Apr2023
 +  * supports openmp
 +  * includes colloid class2 kspace misc molecule ml-pace packages and gpu for cuda
 +  * queues amber128 and test only (cuda 11.6)
 +  * also read miniconda section for tensorflow/pacemaker
 +
 +<code>
 +
 +[hmeij@cottontail2 ~]$ ll /share/apps/CENTOS8/ohpc/software/lammps/25Apr2023/
 +total 314276
 +-rwxr-xr-x 1 hmeij its 133876672 Apr 27 14:24 lmp_mpi
 +-rwxr-xr-x 1 hmeij its 141597552 Apr 27 15:37 lmp_mpi-cuda-double-double
 +-rwxr-xr-x 1 hmeij its 141254208 Apr 27 14:49 lmp_mpi-cuda-single-double
 +-rwxr-xr-x 1 hmeij its 140802904 Apr 27 15:11 lmp_mpi-cuda-single-single
 +-rwxr-xr-x 1 hmeij its 133347480 Apr 27 14:06 lmp_serial
 +
 +# note July 2023
 +# there are now versions with packages REACTION and MC added
 +# same names with postfix '+reaction+mc'
 +
 +</code>
  
  
Line 23: Line 413:
     * example: /zfshomes/hmeij/pycuda/run     * example: /zfshomes/hmeij/pycuda/run
  
-===== Miniconda3 =====+===== Miniconda3-py39 =====
  
   * module: miniconda3/py39   * module: miniconda3/py39
   * Miniconda framework with python 3.9   * Miniconda framework with python 3.9
-    * After loading module ''module list'' to view packages installed 
     * ''module show miniconda3/py39'' will show you file to source if functions are needed     * ''module show miniconda3/py39'' will show you file to source if functions are needed
 +    * ''conda list'' will show you what is installed
  
 +  * **iqtree** rocky8 queues (cottontail2)
 +  * alsao installed in miniconda2&3 for centos queues (cottontail,
 +  * see [[cluster:73|Software]] page)
 +
 +<code>
 +module load miniconda3/py39
 +which conda
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/conda
 +
 +[hmeij@cottontail2 slurm]$ conda list | grep iqtree
 +iqtree                    2.0.3                h176a8bc_1    bioconda
 +
 +iqtree --version
 +IQ-TREE multicore version 2.0.3 for Linux 64-bit built Dec 20 2020
 +
 +# https://userguide.mdanalysis.org/stable/installation.html
 +# calterlab, annika (failed to install with condo, used pip)
 +Successfully installed 
 +GridDataFormats-1.0.1 MDAnalysis-2.3.0 
 +biopython-1.79 fasteners-0.18 gsd-2.6.1 
 +joblib-1.2.0 mmtf-python-1.1.3 mrcfile-1.4.3 
 +msgpack-1.0.4 networkx-2.8.8 threadpoolctl-3.1.0
 +
 +# torch (also in centos 7 python 3.8.3
 +# likely not compatible with cuda9.x, may be 10.2
 +# so use test queue which has cuda 11.6
 +# ezzyatlab, jared
 +Successfully installed 
 +nvidia-cublas-cu11-11.10.3.66 
 +nvidia-cuda-nvrtc-cu11-11.7.99 
 +nvidia-cuda-runtime-cu11-11.7.99 
 +nvidia-cudnn-cu11-8.5.0.96 
 +torch-1.13.0
 +
 +# jupyter, jupyterlab, jupyter-nbclassic
 +# starrlab, max
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-bundlerextension
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-console
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-dejavu
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-execute
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-kernel
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-kernelspec
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-lab
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-labextension
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-labhub
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-migrate
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-nbclassic
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-nbclassic-bundlerextension
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-nbclassic-extension
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-nbclassic-serverextension
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-nbconvert
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-nbextension
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-notebook
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-qtconsole
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-run
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-server
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-serverextension
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-troubleshoot
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/jupyter-trust
 +
 +# chernoff lab
 +# https://anaconda.org/bioconda/soapdenovo2
 +
 +[hmeij@cottontail2 ~]$ conda list | grep novo
 +soapdenovo2               2.40                          0    bioconda
 +
 +</code>
 +
 +For **Lammps** (starrlab) 25April2023
 +
 +  * https://pacemaker.readthedocs.io/en/latest/pacemaker/install/
 +  * consult /share/apps/CENTOS7/lammps/25Apr2023.install
 +    * make yes-gpu
 +    * make yes-ml-pace
 +
 +<code>
 +
 +
 +[hmeij@cottontail2 ~]$ which python
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/python
 +[hmeij@cottontail2 ~]$ pip list |  grep tensorflow
 +tensorflow                    2.8.0
 +tensorflow-io-gcs-filesystem  0.32.0
 +[hmeij@cottontail2 ~]$ which pacemaker
 +/share/apps/CENTOS8/ohpc/software/miniconda3/py39/bin/pacemaker
 +
 +</code>
  
 ===== Amber ===== ===== Amber =====
 +
 +  * module: amber/22
 +  * Amber22 with AmberTools22
 +  * embedded openmpi 1.4.1
 +    * how to run on centos7 nodes
 +    * https://dokuwiki.wesleyan.edu/doku.php?id=cluster:214#amber22
  
   * module: amber/20   * module: amber/20
cluster/215.1649618773.txt.gz · Last modified: 2022/04/10 15:26 by hmeij07