AI in Sweden
- AI in Sweden: Server Configuration Overview
This article details the server configuration supporting Artificial Intelligence (AI) initiatives within Sweden. It is intended as a guide for new system administrators and developers working with these resources. This infrastructure is crucial for ongoing research and development in machine learning, natural language processing, and computer vision. Please refer to the System Administration Guide for general server management procedures.
Overview
The Swedish AI infrastructure is distributed across several key data centers, prioritizing redundancy, scalability, and energy efficiency. The core architecture leverages a hybrid cloud model, utilizing both on-premise hardware and cloud resources from providers like Amazon Web Services and Microsoft Azure. This allows for flexible resource allocation based on project needs and cost optimization. We adhere to the principles outlined in the Data Security Policy.
Hardware Specifications
The on-premise infrastructure is built around high-performance servers optimized for AI workloads. These servers primarily utilize GPUs for accelerated computing. The following table details the specifications for the primary server class:
| Component | Specification | |
|---|---|---|
| CPU | Dual Intel Xeon Platinum 8380 (40 cores/80 threads per CPU) | |
| RAM | 512 GB DDR4 ECC Registered RAM | |
| GPU | 8 x NVIDIA A100 80GB PCIe 4.0 | |
| Storage | 4 x 8TB NVMe SSD (RAID 0) for OS and temporary data | 16 x 18TB SAS HDD (RAID 6) for long-term storage |
| Network | Dual 100GbE Network Interface Cards (NICs) | |
| Power Supply | 3000W Redundant Power Supplies |
Additional servers are configured with different GPU types (e.g., NVIDIA RTX 3090, AMD Radeon Pro W6800) based on specific project requirements. See the GPU Allocation Policy for details on requesting GPU resources.
Software Stack
The servers run a customized Linux distribution based on Ubuntu Server 22.04 LTS. The core software stack includes:
- CUDA Toolkit: For GPU-accelerated computing. Version 11.8 is currently deployed.
- cuDNN: NVIDIA CUDA Deep Neural Network library. Version 8.6.0.
- TensorFlow: An open-source machine learning framework. Version 2.12.0.
- PyTorch: Another popular open-source machine learning framework. Version 2.0.1.
- Docker: For containerization and deployment of AI applications. Version 20.10.
- Kubernetes: For container orchestration. Version 1.26.
- NCCL: NVIDIA Collective Communications Library. Used for multi-GPU communication.
- MPI: Message Passing Interface. For distributed computing.
Detailed installation and configuration instructions for each software package are available in the Software Documentation Repository.
Network Topology
The network infrastructure is designed for high bandwidth and low latency. Servers are interconnected via a high-speed InfiniBand network. The network topology is a fat-tree architecture, providing multiple paths between any two servers. The following table summarizes the network configuration:
| Network Segment | IP Range | Subnet Mask | Gateway |
|---|---|---|---|
| Management Network | 192.168.1.0/24 | 255.255.255.0 | 192.168.1.1 |
| Data Network (InfiniBand) | 10.0.0.0/8 | 255.255.0.0 | 10.0.0.1 |
| Public Network | Various (dynamic) | N/A | N/A |
Firewall rules are configured according to the Network Security Policy to restrict access to sensitive resources. All network traffic is monitored using Nagios for intrusion detection and performance analysis.
Storage Infrastructure
Data storage is a critical component of the AI infrastructure. We utilize a combination of local SSDs for fast access to frequently used data and a centralized network file system (NFS) for long-term storage. The NFS server is a cluster of high-capacity storage servers running Ceph. The following table details the storage capacity and performance:
| Storage Type | Capacity | Performance (IOPS) | Redundancy |
|---|---|---|---|
| Local SSD | 32 TB per server | 500,000+ | RAID 0 |
| NFS (Ceph) | 5 PB | 100,000+ | Erasure Coding (EC) |
Data backups are performed daily and stored offsite according to the Backup and Disaster Recovery Plan. Access to storage resources is controlled through user authentication and authorization using LDAP.
Security Considerations
Security is paramount. All servers are hardened according to the Server Hardening Guide. Regular vulnerability scans are performed using OpenVAS. Access control is strictly enforced using role-based access control (RBAC). All data is encrypted at rest and in transit. We comply with the requirements of the Swedish Data Protection Authority. Please review the Incident Response Plan in case of security breaches.
Main Page Server Documentation AI Research Projects Data Center Locations Contact Support
Intel-Based Server Configurations
| Configuration | Specifications | Benchmark |
|---|---|---|
| Core i7-6700K/7700 Server | 64 GB DDR4, NVMe SSD 2 x 512 GB | CPU Benchmark: 8046 |
| Core i7-8700 Server | 64 GB DDR4, NVMe SSD 2x1 TB | CPU Benchmark: 13124 |
| Core i9-9900K Server | 128 GB DDR4, NVMe SSD 2 x 1 TB | CPU Benchmark: 49969 |
| Core i9-13900 Server (64GB) | 64 GB RAM, 2x2 TB NVMe SSD | |
| Core i9-13900 Server (128GB) | 128 GB RAM, 2x2 TB NVMe SSD | |
| Core i5-13500 Server (64GB) | 64 GB RAM, 2x500 GB NVMe SSD | |
| Core i5-13500 Server (128GB) | 128 GB RAM, 2x500 GB NVMe SSD | |
| Core i5-13500 Workstation | 64 GB DDR5 RAM, 2 NVMe SSD, NVIDIA RTX 4000 |
AMD-Based Server Configurations
| Configuration | Specifications | Benchmark |
|---|---|---|
| Ryzen 5 3600 Server | 64 GB RAM, 2x480 GB NVMe | CPU Benchmark: 17849 |
| Ryzen 7 7700 Server | 64 GB DDR5 RAM, 2x1 TB NVMe | CPU Benchmark: 35224 |
| Ryzen 9 5950X Server | 128 GB RAM, 2x4 TB NVMe | CPU Benchmark: 46045 |
| Ryzen 9 7950X Server | 128 GB DDR5 ECC, 2x2 TB NVMe | CPU Benchmark: 63561 |
| EPYC 7502P Server (128GB/1TB) | 128 GB RAM, 1 TB NVMe | CPU Benchmark: 48021 |
| EPYC 7502P Server (128GB/2TB) | 128 GB RAM, 2 TB NVMe | CPU Benchmark: 48021 |
| EPYC 7502P Server (128GB/4TB) | 128 GB RAM, 2x2 TB NVMe | CPU Benchmark: 48021 |
| EPYC 7502P Server (256GB/1TB) | 256 GB RAM, 1 TB NVMe | CPU Benchmark: 48021 |
| EPYC 7502P Server (256GB/4TB) | 256 GB RAM, 2x2 TB NVMe | CPU Benchmark: 48021 |
| EPYC 9454P Server | 256 GB RAM, 2x2 TB NVMe |
Order Your Dedicated Server
Configure and order your ideal server configuration
Need Assistance?
- Telegram: @powervps Servers at a discounted price
⚠️ *Note: All benchmark scores are approximate and may vary based on configuration. Server availability subject to stock.* ⚠️