Quick Start Guide


Welcome to GridUnesp

This guide will help you get started quickly on the cluster.

1. Create an Account

  1. Go to: https://ncc.unesp.br/registration/

  2. Fill out the registration form

  3. Wait for approval (usually 1-2 business days)

  4. You will receive an e-mail with your credentials

See also

Complete details: Registration

2. First Login

Linux/Mac:

ssh your_username@access.grid.unesp.br

Windows:

Use PuTTY or WinSCP (see Accessing the Cluster)

Security Verification:

On the first connection, verify the SSH key fingerprints:

RSA:     SHA256:X3iCb13fWj7u2Tvp/MCCpn0brfSNS5Ie6ehm6lsvPSQ
ECDSA:   SHA256:WVFokXOLnuH9+2e7xxRU2gp7XJqxwuE6H8bbUMqTCXo
ED25519: SHA256:+HEvFMmo0EA6ipPMJSaj5+Q6IabebRN+nRD0nxdYrKQ

Type yes to confirm if you obtained a match with one of these keys.

3. Your First Job

Create a test file:

cd ~
nano first_job.sh

File content:

Example script (first_job.sh)
#!/bin/bash
#SBATCH -J test
#SBATCH -n 1
#SBATCH -t 5:00

# job-nanny requires INPUT and OUTPUT
export INPUT=""
export OUTPUT="result.txt"

# Simple command
job-nanny bash -c "date > result.txt; echo 'Hello GridUnesp!' >> result.txt"

Submit the job:

sbatch first_job.sh

Check the status:

squeue -u $USER

View the result:

cat result.txt
cat slurm-*.out

4. Essential Commands

Job Management:

sbatch script.sh          # Submit a job
squeue -u $USER           # View your jobs
scancel JOBID             # Cancel a job
scontrol show job JOBID   # Job details

Software Modules:

module avail              # List available modules
module load name          # Load a module
module list               # View loaded modules
module unload name        # Unload a module

File Transfer:

# From your machine to GridUnesp
scp file.dat username@access.grid.unesp.br:~/

# From GridUnesp to your machine
scp username@access.grid.unesp.br:~/result.dat ./

5. Cluster Resources

CPU nodes: 56 nodes (node001-node056)

  • 28 cores per node (Intel Xeon E5-2680 v4)

  • 128 GB RAM per node

GPU node: 1 node (gpunode001)

  • 48 CPU cores (AMD EPYC 9224)

  • 4 NVIDIA L40S GPUs (48 GB each)

Storage:

  • /home/ - Your home directory

  • /tmp/ - Fast local storage (per node)

  • /store/ - Space for large files, shared across multiple nodes

Network:

  • 40 Gbps between nodes

6. Next Steps

Important

NEVER:

  • Run heavy tasks on the access server

  • Forget to define INPUT and OUTPUT for job-nanny

  • Store data only on GridUnesp (there is no backup system!)

ALWAYS:

  • Use job-nanny for all tasks

  • Keep backups of important data externally

  • Load the necessary modules in your script

  • Monitor your job queue: squeue -u $USER

Need help?

Good luck! 🚀