ARTIFICIAL INTELLIGENCE SERVERS & SOLUTIONS
AI is shaping the future.
Artificial intelligence (AI) applies advanced analysis and logic-based techniques, including machine learning and deep learning, to interpret events, support and automate decisions, and take actions.
Businesses around the world are responding vigorously to the new opportunities offered by AI. AI workloads include applications based on machine learning and deep learning, using unstructured data and information as the fuel to drive these applications.
AI workloads demand specific and high performing server infrastructure that many businesses are struggling to identify and build.
What makes a server an AI server?
A server is considered an AI server if it is specifically designed and optimized for the computational demands of AI workloads. These servers typically have specialized hardware and software configurations that are tailored to support the unique requirements of AI applications.
To be an AI server, it needs high-performance computing capabilities. AI workloads often require massive amounts of computing power to process and analyze large datasets. Powerful CPUs and GPUs are a necessity.
GPUs are indeed key components, but don’t underestimate the storage capability to accelerate your AI workloads! AI requires large amounts of memory and storage to handle the massive amounts of data involved.
Popular Artificial Intelligence workloads
- Face detection & recognition
- Text editors & autocorrect
- Search & recommendation algorithms
- Chatbots
- Digital assistants
Choosing 2CRSi for your AI needs
Based on your specific Artificial Intelligence workloads, we will shape a solution with the technology most adapted to your needs. 2CRSi’s in-house AI experts are available to design YOUR adapted solution.
An experienced, skilled, and dedicated team
- 2CRSi is staffed by engineers with an advanced degree and many years of industry experience in a domain such as mechanical engineering, thermal engineering, computer science, providing expertise in storage, network and compute power.
- We are committed to providing customer choice of hardware and software for HPC solutions.
- 2CRSi is the only company that provides 3 types of cooling from Air Cooled server, Direct liquid cooling server, Hybrid server, and Immersion server.
- Because 2CRSi offers a wide selection of solutions, we can be considered as a trusted advisor without trying to sell a one-size-fits-all approach to your problem. 2CRSi understands a broad spectrum of challenges and how to address them.
End To End Support Management
- AI Consulting (Audits / POC)
- 2CRSi R&D (Tailor-made Solutions / Qualification / Support)
- Planning & Installation Services, Factory integration
- Testing Quality Assurance
- Additional managed Services and support (Annual Maintenance Contract term)
Our AI Servers
GPU Servers
For running workload as powerful as AI, GPU servers are a must. See our products below:
Godì 1.8 SR-NV8
6U 19"
The excellence of AI Supercomputing Platform based
on Intel® Xeon® 4th generation and NVIDIA SXM5 with 8 H100 80GB
!
Godì 1.8 ER-NV8
7U 19"
The excellence of AI Supercomputing Platform based
on Intel® Xeon® 5th generation and NVIDIA SXM5 with 8 H200 141GB
!
Atlas 1.8GG
1U 21” hyper-dense AI server
Available in:
- The first server designed for dual-phase immersion cooling
- Up to 8 NVIDIA GPUs (H200, H100)
- 2 AMD CPUs (SP5 socket)
- 24 memory slots
Applications:
- HPC
- Rendering
- AI/ML
- GPU virtualization
OCtoPus 1.4
1OU 21” OCP AI server
Available in:
- OCP inspired
- Designed for single phase immersion cooling
- Up to 4 NVIDIA GPUs (H200, H100)
- 1 AMD CPUs (SP5 socket)
- 24 memory slots
Applications:
- HPC
- Rendering
- AI/ML
- GPU virtualization
Atlantis 1.4G0
1U 19” AI server
Available in:
- Designed for single phase immersion cooling
- Up to 4 NVIDIA GPUs (H200, H100)
- 1 AMD CPUs (SP5 socket)
- 24 memory slots
Applications:
- HPC
- Rendering
- AI/ML
- GPU virtualization
Our product designed for Edge AI
Compact Rugged Computer with NVIDIA® Jetson™
Edge AI being a booming field, 2CRSi and Tranquil IT designed and manufactured a complete solution in the form of a compact, rugged and embedded computer. To power this product, we chose the infamous NVIDIA® Jetson technology.
This module is specifically designed for running complete AI systems, offering an impressive 21 TOPS of accelerated computing power. It efficiently handles the processing requirements of parallel neural networks and seamlessly manages data from multiple high-resolution sensors all at once!
Our embedded computer made for Edge AI can run many IoT applications such as: image classification, object detection, segmentation, and voice recognition, to only cite a few.
Our AI Clients
TWO-I
Context: Video Content Analytics Software Solution in Security
Challenges:
- Hardware solution that carry the highest power possible
- Able to handle frequent updates
- High quality and available in a short period of time
- Respect of quality/value ratio
2CRSi Solution:
- High density solution with NVIDIA® RTX™ A5000 & 8000
Medical Research
Context: AI-Powered Medical Activity
Challenges:
- A multi-GPU solution for AI
- 4 to 8 GPUs in one 2U server
- Need for high-density
2CRSi Solution:
- 19'' ultra-dense 2U server with up to 8 high-end GPUs
- Cooled in air cooling
- Hosted in our data center
Ecommerce field
Context: Commercial Activity
Challenges:
- Optimized for images analysis (AI)
- A PC as compact as possible
- Industry standard with high range temperature
- High-end GPU from NVIDIA®
2CRSi Solution:
- A compact PC with optimized Air Flow
Our AI Project
AI-controlled cooling
2CRSi’s OCtoPus range is based on OCP principles. We design and manufacture servers as well as racks. Our goal is to use reinforcement learning algorithm to control the fans in real time: the speed they spin at when they need to diminish/augment (fans rpm) depending on the server actual workload.
Automatized fans can help to reduce the electricity cost. Our solution can dynamically adapt to prevent overheating for each server (CPU throttling) and, at the same time, optimally cool the rest of the rack, which makes it the best tool to efficiently cool components and offers sustainability.
Our research doesn’t stop at air cooling. We are also working to adapt our reinforcement learning algorithm to direct liquid cooling and immersion cooling systems.
