# Opera Iceland KEF-1 SuperPOD

> Iceland supercomputer

**Wikidata**: [Q132798937](https://www.wikidata.org/wiki/Q132798937)  
**Source**: https://4ort.xyz/entity/opera-iceland-kef-1-superpod

## Summary
Opera Iceland KEF-1 SuperPOD is an Iceland-based TOP500-ranked supercomputer built by Nvidia that entered service in 2024 and delivers 10.53 petaflops (Rmax) of performance. The system is a DGX H100 SuperPOD installation located in Keflavík and is powered by 36,208 processor cores, NVIDIA H100 80 GB GPUs, and Intel Xeon Platinum 8480C CPUs.

## Key Facts
- **Performance**: 10.53 petaflops Rmax, 17.02 petaflops Rpeak (TOP500, June 2025)
- **TOP500 Rank**: #135 (June 2025), #111 (Nov 2024), #88 (June 2024)
- **Location**: Keflavík, Iceland; country of origin: Iceland
- **Inception**: 2024 (TOP500 entry date)
- **Manufacturer**: Nvidia; runs Nvidia DGX OS 5
- **Compute Hardware**: Intel Xeon Platinum 8480C 56C @ 3.8 GHz CPUs, NVIDIA H100 80 GB GPUs
- **Interconnect**: InfiniBand NDR400
- **Compiler**: NVIDIA CUDA Compiler (NVCC) 12.4
- **TOP500 System ID**: 180266
- **Total Processor Cores**: 36,208

## FAQs
### Q: What is the current TOP500 ranking of Opera Iceland KEF-1 SuperPOD?
A: As of June 2025, the system is ranked #135 on the TOP500 list, down from #88 in June 2024.

### Q: Where is Opera Iceland KEF-1 SuperPOD physically located?
A: The supercomputer is installed in Keflavík, Iceland.

### Q: Who built Opera Iceland KEF-1 SuperPOD?
A: Nvidia manufactured the system; it is a DGX H100 SuperPOD configuration.

### Q: What operating system does it use?
A: The machine runs NVIDIA DGX OS version 5.

### Q: How much performance does it deliver?
A: Measured (Rmax) performance is 10.53 petaflops; theoretical (Rpeak) performance is 17.02 petaflops.

## Why It Matters
Opera Iceland KEF-1 SuperPOD represents Iceland’s entry into the global supercomputing arena, providing a strategic Northern European location for high-performance AI and HPC workloads. Its Nvidia DGX H100 architecture offers state-of-the-art GPU acceleration, making it attractive for large-scale deep-learning training and inference tasks that benefit from abundant renewable energy in Iceland. The system’s TOP500 presence signals growing international investment in Icelandic data-center infrastructure, leveraging the country’s cool climate and geothermal power to operate energy-intensive supercomputers more sustainably. For researchers and companies requiring EU-adjacent, low-carbon compute resources, KEF-1 SuperPOD delivers both cutting-edge performance and environmental advantages.

## Notable For
- One of the few TOP500 supercomputers sited in Iceland
- Employs the latest Nvidia H100 GPUs and InfiniBand NDR400 networking
- Achieved a TOP500 peak ranking of #88 in June 2024
- Operates on Nvidia’s turnkey DGX SuperPOD platform for rapid deployment
- Utilizes Iceland’s renewable energy grid for reduced carbon footprint

## Body
### System Architecture
Opera Iceland KEF-1 SuperPOD is built on Nvidia’s DGX H100 SuperPOD reference design. Each DGX node pairs two Intel Xeon Platinum 8480C processors (56 cores each, 3.8 GHz) with eight NVIDIA H100 80 GB GPUs. Nodes are interconnected via InfiniBand NDR400, yielding 400 Gb/s bandwidth for low-latency, high-throughput collective operations critical to large AI models.

### Performance Trajectory
The system debuted on the TOP500 list in June 2024 at position #88 with 10.53 petaflops Rmax. By November 2024 it slipped to #111, and by June 2025 it ranked #135, reflecting the rapid pace of new, larger installations rather than any degradation in capability.

### Software Stack
KEF-1 runs NVIDIA DGX OS 5, a Linux-based distribution optimized for GPU workloads. Developers compile CUDA kernels with NVCC 12.4, matching the hardware generation for maximum compatibility and performance.

### Site Details
Located in Keflavík (TOP500 site ID 51006), the installation benefits from Iceland’s temperate climate, reducing cooling costs, and from the island’s 100 % renewable electricity mix, aligning with sustainability mandates increasingly common in European research and commercial projects.

## References

1. [Source](https://www.top500.org/system/180266/)
2. [Source](https://www.top500.org/site/51006/)
3. [TOP500](https://www.top500.org/system/180266)