Differences

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

--- cluster:215 [2023/08/21 19:12]
hmeij07
+++ cluster:215 [2023/09/15 20:19]
hmeij07 [Miniconda3-py311]
@@ Line 1: / Line 1: @@
 \\
 **[[cluster:0|Back]]**
+===== OpenHPC Software =====
+This list of software is compiled for Rocky 8 using the OpenHPC v2.4 gnu9-openmpi4 toolchain (in your default environm,ent). For gpu applications CUDA 11.6 is the default. That module ''cuda/11.6'' will automatically load for those applications.
+The control of the environment is done via environment modules. \\
+Read this page https://dokuwiki.wesleyan.edu/doku.php?id=cluster:214#module_environment
+The scheduler is Slurm and there is some basic information here\\
+https://dokuwiki.wesleyan.edu/doku.php?id=cluster:214#slurm\\
+https://dokuwiki.wesleyan.edu/doku.php?id=cluster:214#slurm_jobs
 ===== Miniconda3-py311 =====
@@ Line 9: / Line 20: @@
   * 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>
-# general
+# thayerlab (queues: test, amber128)
-  blas               pkgs/main/linux-64::blas-1.0-mkl
+# conda install -c nvidia cudatoolkit=11.6 cudnn=8.2
-  intel-openmp       pkgs/main/linux-64::intel-openmp-2023.1.0-hdb19cb5_46305
+  cudatoolkit        nvidia/linux-64::cudatoolkit-11.6.0-habf752d_9
-  libgfortran-ng     conda-forge/linux-64::libgfortran-ng-13.1.0-h69a702a_0
+  cudnn              conda-forge/linux-64::cudnn-8.2.1.32-h86fa8c9_0
-  libgfortran5       conda-forge/linux-64::libgfortran5-13.1.0-h15d22d2_0
-  llvm-openmp        pkgs/main/linux-64::llvm-openmp-14.0.6-h9e868ea_0
-  mkl                pkgs/main/linux-64::mkl-2023.1.0-h213fc3f_46343
-  mkl-service        pkgs/main/linux-64::mkl-service-2.4.0-py311h5eee18b_1
-  mkl_fft            pkgs/main/linux-64::mkl_fft-1.3.6-py311ha02d727_1
-  mkl_random         pkgs/main/linux-64::mkl_random-1.2.2-py311ha02d727_1
-  mpi                conda-forge/linux-64::mpi-1.0-openmpi
-  mpi4py             conda-forge/linux-64::mpi4py-3.1.4-py311hd0725ac_0
-  numpy              pkgs/main/linux-64::numpy-1.25.2-py311h08b1b3b_0
-  numpy-base         pkgs/main/linux-64::numpy-base-1.25.2-py311hf175353_0
-  openmpi            conda-forge/linux-64::openmpi-4.1.5-external_1
-  python_abi         conda-forge/linux-64::python_abi-3.11-2_cp311
-  scipy              pkgs/main/linux-64::scipy-1.11.1-py311h08b1b3b_0
-  tbb                pkgs/main/linux-64::tbb-2021.8.0-hdb19cb5_0
-# thayerlab
+# 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
-  cudatoolkit-11.6.2         |      hfc3e2af_12       598.9 MB  conda-forge
-  cudnn-8.2.1.32             |       h86fa8c9_0       673.9 MB  conda-forge
+[hmeij@n100 ~]$ module load cuda/11.6
-# conda install -c nvidia cuda-python
+[hmeij@n100 ~]$ module load miniconda3/py311
-  cuda-cudart        nvidia/linux-64::cuda-cudart-12.2.128-0
-  cuda-nvrtc         nvidia/linux-64::cuda-nvrtc-12.2.128-0
-  cuda-python        nvidia/linux-64::cuda-python-12.2.0-py311h79c70a0_0
+[hmeij@n100 ~]$ python
-</code>
+Python 3.11.4 (main, Jul  5 2023, 13:45:01) [GCC 11.2.0] on linux
-===== OpenHPC Software =====
+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
+>>>
-This list of software is compiled for Rocky 8 using the OpenHPC v2.4 gnu9-openmpi4 toolchain (in your default environm,ent). For gpu applications CUDA 11.6 is the default. That module ''cuda/11.6'' will automatically load for those applications.
+# added from nvidia channel
+libcublas libcufft libcusparse libcusolver libcurand libcutensor
+# general channel
+cuda-thrust
-The control of the environment is done via environment modules. \\
+# coolonlab
-Read this page https://dokuwiki.wesleyan.edu/doku.php?id=cluster:214#module_environment
+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
+</code>
-The scheduler is Slurm and there is some basic information here\\
-https://dokuwiki.wesleyan.edu/doku.php?id=cluster:214#slurm\\
-https://dokuwiki.wesleyan.edu/doku.php?id=cluster:214#slurm_jobs
 ===== Pytraj ====

DokuWiki

User Tools

Site Tools

Differences

Page Tools