Bulgarian Academy of Sciences

AI & High-Performance Computing Infrastructure for Molecular Modelling and Scientific Research

The Institute of Biophysics and Biomedical Engineering (IBPhBME) at the Bulgarian Academy of Sciences partnered with Persy to design and implement a high-performance computing (HPC) and GPU-accelerated infrastructure supporting molecular modelling, large-scale data processing and AI-driven scientific workloads.

Overview

The project required a scalable architecture capable of handling complex simulations, large datasets and GPU acceleration while ensuring long-term stability and reliability within a research environment.

The Challenge

• Processing complex molecular simulations and scientific datasets
• Supporting GPU-accelerated AI workloads
• Centralized storage and infrastructure orchestration
• Secure and high-speed networking
• Cost efficiency within a European research framework

Persy’s Approach

Instead of delivering standalone hardware components, Persy engineered a fully integrated HPC ecosystem — from architecture design to deployment and operational readiness.

• End-to-end infrastructure architecture
• Compute, storage and networking integration
• GPU acceleration strategy
• Deployment, testing and commissioning
• Operational training and handover

Solution Architecture

Compute Layer

• HPC cluster of eight Intel Xeon-based servers
• Professional NVIDIA GPU acceleration
• Remote workstation capabilities

AI Acceleration

• Dedicated GPU server with four NVIDIA accelerators
• Designed for high-performance AI workloads

Storage & Management

• Two enterprise storage systems with 96 TB capacity
• Centralized cluster storage
• Monitoring and web-based management server

High-Speed Networking

• Enterprise switch with 10GbE and 100GbE connectivity
• Low-latency interconnect between compute and storage

Infrastructure Reliability

• UPS power protection
• Climate-controlled server environment
• Optimized operational stability

Implementation Highlights

• Rack integration and deployment
• Configuration and performance validation
• Production rollout
• Administrator training

Results & Impact

• Accelerated molecular simulations
• Faster AI experimentation
• Centralized infrastructure control
• Optimized cost-to-performance ratio
• Reliable environment for long-running workloads

Project Value

≈ 76,980 EUR (excl. VAT)

Why Persy

• HPC & AI architecture expertise
• GPU infrastructure integration
• High-speed networking and storage
• Lifecycle deployment and support
• Operational readiness and training

Persy doesn’t just deliver hardware — Persy builds platforms for innovation.