Astral
The Astral subcluster of SciClone contains one front-end (astral/as00) with 32 CPU cores and one node (as01) with 64 Platinum Intel cores and 8x Nvidia A30 GPUs with 24 GB of GPU memory. This cluster should be used for mainly small (1-16) cpu jobs and/or GPU jobs.
Hardware
|
Front-end astral/as00 |
GPU node as01 |
||
|---|---|---|---|
| Model | Dell PowerEdge R750 | Lenovo ThinkSystem SR670 V2 | |
| Processor(s) |
2×24 core Intel Xeon Gold 6336Y |
2x32 core Intel Xeon Platinum 8362 |
|
| Clock speed |
2.4 GHz |
2.80 GHz |
|
| GPU |
-- |
8x Nvidia A30 (24GB) | |
| Memory | 256 GB | 512 GB | |
|
Network interfaces |
Application |
HDR IB (as00-ib) |
HDR IB (as01-ib) |
| System |
10 GbE |
||
| OS | Rocky Linux 9.2 | ||
Slurm
The SLURM batch system is used on astral/as01 to run jobs. The maximum walltime for all jobs on this cluster is 72hrs. The front-end, astral, has 32 CPU cores available which will be the default node if you require only CPUs. The GPU node, as01, may be specidied via your command-line options:
To run an interactive job on CPUs only:
salloc -n 4 -t 1:00:00 # will get you 4 CPU cores
To request one or more GPUs:
salloc -n 4 -t 1:00:00 --gpus=2 # will get you 4 CPU cores and 2 GPUs
The --gpus=N can be added to your srun command (to run a job directly from the command line) or in your batch script, i.e.:
#SBATCH --gpus=2
Currently, there are no reserved slots for debugging. Please send email to [hpc-help] if you need help obtaining resources on this subcluster.
User Environment
To login, use SSH from any host on the William & Mary or VIMS networks and connect to astral.sciclone.wm.edu with your HPC username (usually the same as your WMuserid) and W&M password.
Your home directory on Astral is the same as everywhere else on SciClone, and all of the usual filesystems (/sciclone/homeXX, /sciclone/dataXX, /sciclone/scrXX, /local/scr, etc.) are available.
SciClone uses Environment Modules (a.k.a Modules) to automatically configure the user's shell environment across multiple computing platforms, as well as to organize the dozens of different software packages which are available on the system. We support tcsh as the primary shell environment for user accounts and applications.
The file which controls startup modules for astral is .cshrc.kuro. A copy of this file can be found in /usr/local/etc/templates on kuro.sciclone.wm.edu.
Preferred filesystems
The astral front-end server has a small 63 GB /local/scr partition which can be used for small files. The as01 GPU node has a 14TB NVME local scratch disk which should be utilized for local file i/o if possible.
The preferred global file system for all work on astral is /sciclone/scr10/$USER is a good alternative (NFS, but connected to the same InfiniBand switch).
Compiler flags
All W&M HPC clusters have Intel and GNU compilers available. There are somewhat frequent updates so please check the modules that are available on each cluster. Here are suggested flags for Ice-Lake cpus:
| Intel | C | icc -O3 -xHost -fma -align -finline-functions |
|---|---|---|
| C++ | icpc -std=c11 -O3 -xHost -fma -align -finline-functions | |
| Fortran | ifort -O3 -xHost -fma -align array64byte -finline-functions | |
| GNU | C | gcc -march=native -O3 -mfma -malign-data=cacheline -finline-functions |
| C++ | g++ -std=c11 -march=native -O3 -mfma -malign-data=cacheline -finline-functions | |
| Fortran | gfortran -march=native -O3 -mfma -malign-data=cacheline -finline-functions |
MPI
All W&M HPC clusters have both Intel MPI and OpenMPI installed. Some older clusters also have Mvapich2. All parallel jobs (shared memory, MPI or hybrid) need to be run through the batch system. The standard way to do this is to use the batch script#!/bin/tcsh
#SBATCH --job-name=test #SBATCH --nodes=1 --ntasks-per-node=8 #SBATCH -t 30:00
#SBATCH --gpus=1srun ./a.out >& LOGPlease see our Slurm page for more help.
