DRACO 128-Node H200 HPC Cluster

Overview

The DRACO 128-Node H200 HPC Cluster is a turnkey 128-node HPC cluster pairing massive CPU-forward compute with large-scale GPU acceleration — 24576 AMD EPYC cores across 128 nodes plus 512× H200 accelerators (72,192 GB HBM3e HBM3e, NVLink in-node), wired with a non-blocking InfiniBand fabric. It runs traditional FP64/MPI HPC and large-model AI on one cluster, on-premises, in INR, on a GST invoice.

Engineered for organisations whose workload spans serious simulation and serious AI, it is delivered racked, cabled, scheduled and validated as a single system — RDP sizes the nodes, fabric, parallel storage, scheduler and cooling, with one warranty and one support contract.

Key highlights

• 24576 EPYC cores across 128 nodes — large-scale CPU throughput for FP64, MPI and memory-bound HPC.
• 512× H200 · 72,192 GB HBM3e — large-scale GPU acceleration for model training and high-throughput inference alongside HPC.
• Non-blocking InfiniBand NDR fabric — full-bisection bandwidth for tightly-coupled MPI jobs and distributed training.
• 128 TB DDR5 ECC + 8 PB NVMe parallel NVMe — large memory and a high-throughput parallel filesystem at cluster scale.
• Scheduler-ready, turnkey — delivered with Slurm/Kubernetes, validated; one SKU, one warranty.
• On-prem data sovereignty — codes, data and models stay in-house; DPDP-friendly, air-gappable.
• Make-in-India OEM — predictable INR pricing, GST tax invoice (HSN 8471), pan-India onsite support, GeM-procurable.
• Scale path — grow node count toward supercomputer scale, or specialise into GPU-dense rack-scale systems.

AI workload fit (what it actually runs — honestly)

• HPC + simulation (primary): large FP64/MPI workloads — CFD, FEA, molecular dynamics, EDA, risk and Monte-Carlo — across 24576 cores.
• AI training & inference: train and fine-tune large models across the NVLink-connected H200 GPUs and serve them at high throughput.
• RAG & agentic: production AI services co-located with HPC for engineering and research teams.
• Engineering note: each node’s H200 GPUs are NVLink-connected for efficient in-node tensor parallelism, and nodes scale over a non-blocking InfiniBand spine — this cluster trains and serves large models alongside FP64 HPC. The CPU cores carry the simulation/MPI side; the GPUs carry the AI side. We validate real MPI and AI scaling on your codes rather than quoting a peak number.

AI workload positioning

This sits at the HPC + AI-training stage at cluster scale: a system for organisations running both large simulation and large-model AI. With 24576 CPU cores and 512 H200 GPUs on a non-blocking fabric, it is sized to sustain real HPC throughput plus model training — the owned alternative to renting hyperscale HPC.

Industry use cases

• Manufacturing & automotive — large-scale CFD/FEA plus AI-assisted design.
• Energy & engineering — reservoir, structural and process simulation at scale.
• BFSI — large risk, Monte-Carlo and quantitative analytics with AI.
• EDA & semiconductors — chip design verification at scale.
• Government & research — a shared institutional HPC + AI cluster.
• Life sciences — large molecular dynamics and bioinformatics pipelines.

Performance — and how to be sure

We don’t publish inflated peak numbers. The honest picture: 24576 EPYC cores and 512× H200 on a non-blocking fabric are sized for large-scale converged HPC + AI. Want certainty? Request a free benchmark of your HPC codes and AI models, including an MPI and distributed-training scaling test, on a representative configuration before you buy; we’ll send back real performance and scaling efficiency.

Series & upgrade path

• DRACO (flagship cluster tier) — this.
• Node ladder: up to 128+ nodes; accelerator ladder: L40S (inference/viz) → H200 (training).
• When to step up: for the largest AI training, move to GPU-dense rack-scale systems or a supercomputer — talk to an architect.

On-prem vs cloud — the TCO case

For sustained large-scale HPC, owning beats renting decisively: a continuously-running cloud HPC cluster of this size dominates an engineering budget, and on-prem removes egress and keeps codes and data in-house. RDP pricing is fixed in INR with a GST input-credit-eligible invoice — ask for a 3-year TCO comparison.

Software & day-one readiness

Ships pre-integrated and validated: NVIDIA driver, CUDA, cuDNN, NCCL, MPI (OpenMPI/MPICH), the InfiniBand stack, Slurm and/or Kubernetes, with common HPC toolchains, PyTorch and vLLM / Triton on a supported Linux LTS. Optional managed cluster operations and an HPC/MLOps platform.

Power, cooling & rack integration

A 128-node liquid-cooled cluster with substantial power density — RDP scopes facility power, cooling, the InfiniBand spine and floor/rack requirements in the design. (Exact power, BTU, flow and rack-depth figures confirmed on the build sheet.) Full out-of-band management across the cluster.

Deployment, warranty & support

• Made to order, integrated, racked, cabled, cooled and burned-in in India; project timeline confirmed at quote.
• Delivered as one system: nodes, leaf/spine switches, cabling, PDUs, parallel storage, and the full HPC/AI software stack.
• Onsite warranty + AMC with pan-India coverage, cluster-level SLAs and an RMA/escalation path (exact terms confirmed at quote).

Why RDP

14 years of Make-in-India infrastructure and 300,000+ devices shipped. Indian OEM, INR pricing, GST tax invoice (HSN 8471), pan-India onsite engineers, GeM availability, and DPDP / sovereign-AI-ready deployment.

Buy with confidence

This is a turnkey large-scale HPC + AI cluster, made to order — talk to an RDP solution architect, size the nodes, accelerators and fabric for your codes, get a 3-year TCO, and benchmark your own workloads before you commit. Request a quote to begin.

Specifications