Documentation Index
Fetch the complete documentation index at: https://docs.nyc-ai.app/llms.txt
Use this file to discover all available pages before exploring further.
The HPCC maintains a broad software stack through the Lmod environment module system. Load packages with module load <name> after checking availability with module spider <name> or module avail. See Software & modules for the full workflow.
This catalog is a practical orientation, not a guarantee that every package is installed on every HPCC system. New packages are added over time and some may be retired. Always verify with module avail or module spider on the system you’re using, then contact the HPC Helpline if something you need is missing.
Compilers
| Package | Notes |
|---|
| GNU Compiler Collection (GCC) | Multiple versions. gcc, g++, gfortran. Default compiler on most nodes. |
| Intel compiler suite / oneAPI | Check the module for current compiler names such as icx, icpx, ifx, or legacy icc / ifort; includes MKL. |
| NVIDIA HPC SDK (nvhpc) | nvc, nvc++, nvfortran. OpenACC and CUDA Fortran support. |
| Cray compilers | Legacy Cray environments only; verify current availability before using Cray-specific examples. |
module spider gcc
module load GCC/12.3.0
MPI libraries
| Package | Notes |
|---|
| OpenMPI | Recommended default for most CPU jobs. |
| Intel MPI | Pairs with Intel compilers; typically faster on Intel nodes. |
| Cray MPICH | Legacy Cray environments only; verify current availability before use. |
Choose the MPI that matches your compiler:
module load GCC/12.3.0
module load OpenMPI/4.1.5-GCC-12.3.0
mpicc -o my_prog my_prog.c
Math and scientific libraries
| Package | Notes |
|---|
| FFTW | Fast Fourier Transforms; single- and double-precision. |
| Intel MKL | BLAS, LAPACK, ScaLAPACK, FFTW wrappers bundled with Intel oneAPI. |
| OpenBLAS | Open-source optimized BLAS/LAPACK. |
| HDF5 | Hierarchical data format; parallel (MPI) and serial builds available. |
| NetCDF | Network Common Data Form; critical for climate and earth-science codes. |
| PETSc | Parallel sparse linear algebra for PDEs. |
| SLEPc | Scalable eigenvalue solvers built on PETSc. |
| ELPA | Eigenvalue solvers for dense matrices, popular in quantum chemistry. |
GPU and CUDA
| Package | Notes |
|---|
| CUDA Toolkit | Multiple versions. Required for raw CUDA C/C++ development. |
| cuDNN | Deep learning primitives on top of CUDA (required by PyTorch / TensorFlow). |
| NCCL | Collective communication for multi-GPU training. |
module spider CUDA
module load CUDA/12.2.0
nvcc -o my_kernel kernel.cu
Languages and runtimes
| Package | Notes |
|---|
| Python | Multiple versions (3.8, 3.10, 3.11, …). Includes pip; use venv or conda for isolation. |
| Julia | Recommended for scientific computing newcomers to HPC. |
| R | Statistical computing with access to CRAN packages. |
| Perl | Available for bioinformatics pipelines that require it. |
| Java / JDK | Needed by some bioinformatics tools (e.g., Trimmomatic, Picard). |
Managing Python environments
module load Python/3.10.4
python -m venv ~/envs/myproject
source ~/envs/myproject/bin/activate
pip install numpy scipy matplotlib
Store environments in /global/u/<username> (home), not in /scratch, so they survive scratch purges.
Machine learning and AI
| Package | Notes |
|---|
| PyTorch | Install in a user environment after loading a compatible CUDA module unless a site module is available. |
| TensorFlow | Install in a user environment after loading a compatible CUDA module unless a site module is available. |
| JAX | Install via pip in a user environment; GPU builds require a matching CUDA stack. |
GPU-enabled frameworks typically require module load CUDA and matching cuDNN before installing or importing the package.
Visualization and post-processing
| Package | Notes |
|---|
| ParaView | Parallel visualization, client-server mode supported. |
| VisIt | Visualization and analysis for large datasets. |
| gnuplot | Lightweight 2-D and 3-D plotting from scripts. |
| ImageMagick | Command-line image processing. |
GUI visualization tools require X11 forwarding (ssh -Y) or an off-cluster workflow where you copy results back to your laptop. Rendering is usually more practical done in batch using off-screen backends.
Chemistry and materials
| Package | Notes |
|---|
| Gaussian | Licensed to CUNY; restricted to approved QOS/partition (e.g., qoschem). |
| GAMESS | Free ab initio quantum chemistry. |
| ORCA | Widely used for DFT and coupled-cluster calculations. |
| NAMD | Molecular dynamics; highly optimized for GPU. |
| GROMACS | Molecular dynamics; GPU-accelerated. |
| LAMMPS | Classical MD; many interatomic potential packages. |
| Quantum ESPRESSO | Plane-wave DFT for materials. |
| CP2K | Electronic structure and MD. |
| Package | Notes |
|---|
| BLAST | Sequence alignment. |
| BWA / Bowtie2 | Short-read alignment to reference genomes. |
| STAR | RNA-seq alignment. |
| SAMtools / BCFtools | SAM/BAM processing and variant calling. |
| GATK | Variant calling pipeline (requires Java). |
| Trimmomatic / Fastp | Read quality trimming. |
| SPAdes | Genome assembly. |
Other domain applications
| Package | Notes |
|---|
| MATLAB | Licensed for academic use; launched with matlab -nodisplay -nojvm in batch. |
| WRF | Weather Research & Forecasting model. |
| OpenFOAM | Open-source CFD; multiple versions. |
| LS-DYNA | Structural and crash analysis; licensed academic use on supported systems. |
Requesting new software
Email HPCHelp@csi.cuny.edu with:
- The software name and the version you need.
- A link to its install documentation or source.
- A one-line justification (what research it enables).
Staff prioritize packages that benefit multiple research groups. For personal Python / Julia / R packages, you can usually manage them yourself inside your home directory without waiting for a ticket.