Infrastructure
Facility
High-performance GPU resources spread across lab workstations, on-premises servers, and national HPC clusters.
Compute Resources
9 Workstations · NVIDIA RTX 4090 & RTX 5090
Day-to-day development machines for prototyping, ablations, and fast iteration on DL and SNN workloads. Mix of RTX 4090 (24 GB) and the latest RTX 5090 GPUs.
4 × NVIDIA RTX Ada 6000 — 48 GB VRAM · 220 GB RAM
On-premises GPU server for memory-intensive training runs. 192 GB total GPU VRAM with a large 220 GB system RAM — suited for large-batch training, LLM fine-tuning, and hardware simulation.
NVIDIA A100 GPUs — National Supercomputing Mission
Access to CDAC's national-scale HPC infrastructure powered by A100 GPUs for large-scale model training and compute-heavy research experiments.
NVIDIA V100 GPUs — IIT Madras Shared HPC
Access to IIT Madras's Robert Bosch Centre for Data Science and AI shared HPC cluster for distributed training and collaborative large-scale experiments.
NVIDIA A100 GPUs — IIT Madras HPC Facility
Access to the PARAM Shakti supercomputing facility at IIT Madras, featuring NVIDIA A100 GPUs for large-scale deep learning, high-performance computing, and data-intensive AI research.
9
RTX 4090 / 5090 Workstations
for everyday research & prototyping
4× RTX Ada 6000
On-premises server for on-the-fly training, large models & experiments
220 GB
High-memory server for large-batch workloads
A100 & V100
National & IIT Madras HPC access for large-scale training