# IARA

> supercomputer released in 2021

**Wikidata**: [Q134087112](https://www.wikidata.org/wiki/Q134087112)  
**Source**: https://4ort.xyz/entity/iara

## Summary
IARA is a supercomputer released in 2021, located in Campinas, Brazil. It is powered by AMD EPYC 7742 64C processors and NVIDIA A100 GPUs, connected via InfiniBand networking. IARA is used for research purposes and is manufactured by NVIDIA.

## Key Facts
- IARA was released in 2021 and is located in Campinas, Brazil
- It uses AMD EPYC 7742 64C processors running at 2.25GHz with 64 cores each
- The system features NVIDIA A100 SXM4 GPUs with 40GB memory each
- IARA has a peak performance of 4.13 petaflops and a sustained performance of 3.66 petaflops
- It contains 24,800 processor cores and uses InfiniBand for high-speed networking
- The supercomputer is ranked 272nd on the TOP500 list as of November 2024
- IARA is classified as a research supercomputer and is manufactured by NVIDIA
- It runs on the Q381 operating system

## FAQs
### Q: What is IARA?
A: IARA is a supercomputer released in 2021, located in Campinas, Brazil. It is used for research purposes and features AMD EPYC 7742 processors and NVIDIA A100 GPUs, delivering up to 4.13 petaflops of computing power.

### Q: Who built IARA and where is it located?
A: IARA was manufactured by NVIDIA and is located in Campinas, Brazil. It was released in 2021 as a research supercomputer.

### Q: What are IARA's main technical specifications?
A: IARA uses AMD EPYC 7742 64C processors at 2.25GHz, NVIDIA A100 SXM4 40GB GPUs, and InfiniBand networking. It has 24,800 processor cores and delivers 3.66 petaflops sustained performance.

## Why It Matters
IARA represents a significant computing resource for research in Brazil, providing high-performance computing capabilities to the scientific community. As a research supercomputer, it enables complex simulations, data analysis, and computational research that would be impossible on conventional computing systems. The system's combination of AMD processors and NVIDIA GPUs makes it well-suited for both traditional HPC workloads and emerging AI/ML applications. Its presence in Brazil helps reduce the digital divide in high-performance computing access between developed and developing nations, allowing Brazilian researchers to compete on a global scale without needing to rely on international computing resources. The system's performance level (ranking in the top 300 supercomputers worldwide) makes it a valuable national asset for advancing scientific discovery, engineering design, and data-intensive research across multiple disciplines.

## Notable For
- Ranked 272nd on the TOP500 list as of November 2024, demonstrating world-class performance
- Features cutting-edge AMD EPYC 7742 processors paired with NVIDIA A100 GPUs for optimal performance
- Provides 4.13 petaflops peak performance, enabling complex scientific and engineering computations
- Uses InfiniBand networking for high-speed inter-node communication essential for parallel computing
- Serves as a research computing resource in Brazil, supporting local scientific advancement

## Body
### Technical Architecture
IARA is built around AMD EPYC 7742 64C processors, each featuring 64 cores running at 2.25GHz. These processors are paired with NVIDIA A100 SXM4 GPUs, each equipped with 40GB of high-bandwidth memory. The system uses InfiniBand networking to connect nodes, providing the low-latency, high-bandwidth communication necessary for parallel computing workloads.

### Performance Metrics
The supercomputer delivers a peak performance of 4.13 petaflops (Rpeak) and a sustained performance of 3.66 petaflops (Rmax). It contains 24,800 processor cores, enabling massive parallel processing capabilities. The system's performance places it at 272nd on the TOP500 list as of November 2024, though it achieved higher rankings in previous years (111th in June 2021 and 125th in November 2021).

### System Classification
IARA is classified as a research supercomputer, indicating its primary purpose is to support scientific discovery and computational research rather than commercial or industrial applications. The system runs on the Q381 operating system, which is optimized for high-performance computing environments.

### Hardware Configuration
The system's hardware configuration represents a balanced approach to high-performance computing, combining AMD's server-class processors with NVIDIA's data center GPUs. This combination is particularly effective for workloads that benefit from both traditional CPU processing and GPU acceleration, such as scientific simulations, machine learning, and data analytics.

### Research Applications
As a research supercomputer, IARA supports a wide range of computational workloads including climate modeling, computational chemistry, physics simulations, bioinformatics, and artificial intelligence research. The system's architecture makes it particularly well-suited for both traditional HPC applications and emerging AI/ML workloads that require massive parallel processing capabilities.

## References

1. [Source](https://www.top500.org/site/50863/)
2. [Source](https://top500.org/system/179931/)
3. [TOP500](https://top500.org/system/179931/)