Infrastructure
GPU Computing Facility for Researchers — Department of Computer Science & Engineering

Overview

As part of its ongoing commitment to Infrastructure & Lab Development, the Department of Computer Science & Engineering has established a dedicated high-performance GPU computing facility to support faculty and student research in Artificial Intelligence, Machine Learning, Deep Learning, and Data Science. The facility is built around a state-of-the-art enterprise GPU server designed for demanding, compute-intensive workloads and is available for use by researchers across the department.

GPU Computing Server — ASUS ESC8000A-E12

Facility Highlights

High GPU CapacityEnterprise-grade 4U rackmount server platform supporting up to 8 GPUs for scalable AI/ML workloads.
Powerful Compute64-core / 128-thread AMD EPYC™ 9554 processor for high-throughput data pre-processing and parallel computation. ×2
Dedicated AI Accelerator48 GB GDDR6 NVIDIA RTX 6000 Ada GPU for training and inference of large deep learning models.
Fast Memory & Storage4 × 64 GB high-speed DDR5 RAMs paired with 2 TB NVMe SSD for rapid data access and model checkpointing.
Ample Data Storage20 TB enterprise-grade HDD plus expandable M.2 storage for large research datasets.
Reliable UptimeRedundant 3000W Titanium-rated power supplies ensure uninterrupted operation for long-running training jobs.

Technical Specifications

#ComponentDescription
1Server ChassisESC8000A-E12 — 2× Socket, 24× DIMM slots, 4× 3.5" SATA + 2× NVMe bays, 1000 mm ball-bearing rail, 2+2×3000W 80+ Titanium redundant power supply, 4U chassis supporting up to 8 GPUs; front 1× PCIe x8 slot, rear 1× PCIe x16 slot with additional 12+4P GPU power cable kit.
2ProcessorAMD EPYC™ 9554 — 64 Cores / 128 Threads, 3.1 GHz–3.75 GHz boost, 256 MB L3 Cache, 360 W TDP.
3Memory64 GB Registered DDR5, 4800 MHz / 5600 MHz.
4Primary Storage2 TB NVMe 4th Generation Enterprise M.2 SSD.
5Storage ExpansionHyper M.2 x16 Card v2 — 4× M.2 sockets.
6Bulk Storage20 TB 3.5" WD Ultrastar, 7200 RPM, SATA Enterprise HDD.
7GPUNVIDIA RTX 6000 Ada — 48 GB GDDR6 memory.

Research Applications

The facility is intended to support a broad range of research and academic activities, including:

Deep Learning & AITraining and fine-tuning of deep neural networks, CNNs, and transformer-based models.
Computer VisionLarge-scale image and video processing, object detection, and segmentation research.
Natural Language ProcessingLanguage model experimentation, text analytics, and generative AI research.
High-Performance ComputingLarge-scale simulations, numerical computation, and parallel algorithm development.
Bioinformatics & Data ScienceGenomics, medical imaging, and other GPU-accelerated data-intensive research.
Student ProjectsHands-on project work for PhD, M.Tech, and final-year B.Tech students under faculty supervision.

Eligibility & Access

Who Can Use It
  • Faculty members of the Department of Computer Science & Engineering.
  • Research scholars, PhD candidates, and PG/UG students working on faculty-supervised research or projects.
How to Request Access

Submit a request through the department's laboratory in-charge / HOD along with a brief note on the research requirement and expected resource usage.

Scheduling

Usage is scheduled and monitored to ensure fair and efficient sharing of GPU resources among all researchers.

Usage Policy

Users are expected to follow departmental data-handling, security, and lab-usage policies at all times.

Contact & Further Information

Dr. Dhananjay Kalbande

HoD, Department of Computer Science & Engineering

Empowering Research through High-Performance Computing