2. Software and modules updates
This page lists changes that are a result of the Operating System updates.
Please see Environment modules:
to find the most recent software versions availability.
to learn how to use software modules.
Rocky Linux 8.10 updates
Important
- For user compiled or installed software
If you compiled any software packages yourself using gcc, intel, openmpi modules and your compiled software no longer works as before you may have to recompile it anew.
Some commands, include files and libraries are available from system RPMs and do not require modules. If you compiled any software using such libraries/binaries you may need to recompile anew as some have changed.
If you installed any packages with anaconda or miniconda, you may need to reinstall them. In this case, please try to use the latest version of anaconda or miniconda modules that we provide. Please see Install with conda/mamba user guide that explains how to build and use conda environments on HPC3.
IMPORTANT: if you are currently using miniconda/4.8.5 and miniconda3/4.12.0 please follow Special instructions for miniconda3 users.
- Summary of changes
Modules availability on the previous Rocky Linux 8.8 and on the current Rocky Linux 8.10 operating systems:
Total
Rocky Linux 8.8
Rocky Linux 8.10
Comment
Installed
330
330
active modules
A few modules were removed because they cannot be built on a new operating system.
Some modules were removed because they were unused. Usually a newer version is available.
screen
command is not available, usetmux
.Operating system provides default compiler gcc 8.5.0 and openmpi 4.1.1 which are available via modules:
Package
Module
gcc 8.5.0
gcc.system
openmpi 4.1.1
openmpi/4.1.1/gcc.system
foundation/v8 module is available for backwards compatibility. It will simply load a few other modules to provide an access to the commonly used tools. If you are using a specific tool from the current foundation/v8 you will be able to switch to using the specific module for that tool (see output of
module display foundation/v8
to list loaded modules).sra-tools prior to v.3 do not handle writing temp files to a separate directory and are removed as they cause issues. Please see SRA toolkit fasterq-dump for explanation how to use this toolkit with currently available version.
Python 2.7 has reached the end of life and is no longer supported. It is installed but the outcome of any program can not be predicted. If you have any python code that is still using python 2 convert your code to Python 3. Test your application with the -W default command-line option to see any deprecation warnings and follow online guides (google) for porting Python 2 to Python 3.
singulariy 3.4.1 built containers may fail with an error that looks like:
FATAL: container creation failed: mount /proc/
Please try to use the newer singularity module and add option
-B
to your command:[user@login-x:~]$ module load singularity/3.9.4 [user@login-x:~]$ singularity exec \ -B /dfs8/singularity_containers/rcic/bashrc:/etc/bashrc \ /dfs3a/singularity_containers/contName.simg binaryName
The intel and mkl modules provide access to a number of Intel Math Kernel Libraries including LAPACK, ScaLAPACK, BLAS and threading options.
The MKL libraries can be linked with Intel or GNU compilers. If you are compiling software yourself and are using intel or mkl modules please see Intel MKL Documentation and Intel MKL Link Advisory to help you figure out how to use variables set by them.