root@hpc.gndec.ac.in :~$ ./welcome.sh

GNDEC HPC Cluster
NVIDIA H100  ·  Department of Computer Science & Engineering

High-Performance Computing infrastructure for deep learning, scientific simulation, and cutting-edge engineering research at Guru Nanak Dev Engineering College, Ludhiana.

NVIDIA H100 80GB CUDA 12.x Tensor Cores · 4th Gen NVLink · PCIe 5.0 HBM3 Memory 3,958 TFLOPS FP8
bash — hpc.gndec.ac.in — 80×24
user@hpc:~$ nvidia-smi --query-gpu=name,memory.total,driver_version --format=csv,noheader
NVIDIA H100 80GB HBM3,  81920 MiB,  545.23.08
user@hpc:~$ cat /etc/hpc-info
[ SYSTEM INFORMATION ]
InstitutionGuru Nanak Dev Engineering College, Ludhiana
DepartmentComputer Science & Engineering
Hostnamehpc.gndec.ac.in
GPUNVIDIA H100 SXM5 80GB HBM3
CUDA Cores16,896
Memory BW3.35 TB/s
PurposeAI · ML · Deep Learning · HPC Research
──────────────────────────────────────────────
All systems nominal.
3,958
TFLOPS FP8
80 GB
HBM3 VRAM
16,896
CUDA Cores
700 W
TDP
hardware

NVIDIA H100 — Technical Specifications

Architecture
Hopper
NVIDIA 4th Gen · SM 9.0
🔲
CUDA Cores
16,896
Compute capability 9.0
🧮
Tensor Cores
528 4th Gen
FP8 · FP16 · BF16 · INT8
🧠
GPU Memory
80 GB HBM3
3.35 TB/s bandwidth
🚀
FP8 Tensor
3,958 TFLOPS
With sparsity
📊
FP64
34 TFLOPS
Double precision
🔗
Interconnect
NVLink 4th Gen
900 GB/s bidirectional
💾
L2 Cache
50 MB
PCIe 5.0 · 128 GB/s
FP8 Tensor (w/ sparsity)3,958 TFLOPS
INT8 Tensor Core3,958 TOPS
BF16 / FP16 Tensor1,979 TFLOPS
FP32 CUDA67 TFLOPS
FP64 Double Precision34 TFLOPS
Memory Bandwidth3.35 TB/s
capabilities

Research & Applications

🤖
Deep Learning & AI
Train large neural networks — transformers, CNNs, GANs — at FP8 precision. Full support for PyTorch, TensorFlow, JAX, and ONNX.
🗣️
LLM Research
Fine-tune and run inference on large language models like LLaMA, Mistral, and custom transformers with 80 GB VRAM headroom.
📐
Computer Vision
Object detection, segmentation, 3D point clouds, video understanding, and autonomous systems research at scale.
🔬
Scientific Simulation
CFD, finite element analysis, climate modelling, and physics-based neural operators (PINNs) leveraging FP64 compute.
🧬
Computational Biology
Molecular dynamics, protein structure prediction (AlphaFold2), genome analysis, and bioinformatics pipelines.
📊
Data Science & Analytics
Large-scale data processing with RAPIDS cuDF, graph analytics with cuGraph, and accelerated XGBoost / cuML.
📡
Signal & Image Processing
Real-time signal processing, medical imaging reconstruction, remote sensing, and satellite data analysis.
🔐
Security & Cryptography
Adversarial ML research, GPU-accelerated cryptographic operations, anomaly detection, and network analytics.
institution

About GNDEC & CSE Department

// College Information
Name
Guru Nanak Dev Engineering College
Location
Ludhiana, Punjab, India
Type
Autonomous Technical Institute (Govt.)
Affiliation
I.K. Gujral Punjab Technical University
NAAC Grade
A++
// Department of CSE
Department
Computer Science & Engineering
Programs
B.Tech · M.Tech · Ph.D.
Focus Areas
AI · ML · HPC · Networks · Systems
HPC Domain
hpc.gndec.ac.in
GPU
NVIDIA H100 SXM5 80GB HBM3
getting started

Access & Usage

🔑
SSH Access
Connect with your GNDEC credentials.
ssh user@hpc.gndec.ac.in
📓
JupyterHub
Browser-based GPU notebooks for interactive research and coursework.
📦
SLURM Scheduler
Submit batch GPU jobs for large-scale non-interactive training runs.
🐳
Containers
Docker/Singularity images with PyTorch, TensorFlow, RAPIDS & CUDA.
quick-start guide
$ # 1. Connect
$ ssh yourname@hpc.gndec.ac.in
$ # 2. Load modules
$ module load cuda/12.3 python/3.11
$ # 3. Verify GPU
$ python3 -c "import torch; print(torch.cuda.get_device_name(0))"
NVIDIA H100 80GB HBM3
$ # 4. Submit a job
$ sbatch --gres=gpu:1 --partition=h100 train.sh
Submitted batch job 4271
Access Request
Faculty, research scholars, and students of GNDEC may request access by emailing cse@gndec.ac.in with your employee/roll number, department, and a brief description of your research use case.