Compiling Your Code
The specifics of compiling and linking applications is beyond the scope of this document. However, we make the following compilers available (along with their supporting tools and libraries):
Compilers
GNU Compiler Collection (GCC)
Intel Parallel Studio XE Cluster Edition
Portland Group (PGI) Compilers with OpenACC support
To load one of these compilers into your environment, run one of the following commands:
NVIDIA CUDA
The NVIDIA CUDA tools, profilers, libraries, and SDK are automatically set up in your environment. By default, you will see that the latest version of CUDA is already in your path (try running which nvcc
or echo $CUDA_HOME
).
You can switch to older versions of CUDA using the module
command. To see the options, run:
List available GPUs (and their capabilities)
Additional Tools, Libraries & Software
The systems have CentOS Linux 7.x installed. Lmod Environment Modules are used to control which tools are available at any given time. For example, you cannot have GCC 4.9 and GCC 5.4 loaded at the same time, but you might need different versions for different projects. Environment Modules allow you to control which software packages (and which versions) are active without manually changing all the environment variables in your shell sessions and project build files.
List the software packages built with the current toolchain
List all available software
Load an additional software package
Unload all packages
Help with Environment Modules
Last updated
Was this helpful?