Customized Solutions for Diverse AI Training Requirements
Peak AI Performance
Visual Computing Excellence
Reliable & Manageable
It's a critical step in AI development that determines the model's ability to make accurate predictions and perform tasks effectively in the real world.Â
Training requires substantial computational power and sophisticated algorithms to refine the model’s 'understanding'. The result is an AI that can not only comprehend complex datasets but also evolve its knowledge base over time, providing the foundation for reliable and insightful AI inference applications..Â
VELOCITY N218G
Introducing the VELOCITY N218G, a pioneering solution in the realm of direct liquid cooling, meticulously engineered for the colossal demands of AI and HPC environments. This marvel of technology is a 2U 4-node rear access server system, embodying the pinnacle of performance and innovation. At the heart of the VELOCITY N218G lies the NVIDIA Grace Hopper Superchip, a testament to cutting-edge technology, providing an astounding 900GB/s NVLink-C2C Interconnect that sets a new benchmark for data processing and transfer speeds.
Designed with a CPU+GPU architecture, it ensures an unparalleled computing experience, supporting up to 480GB CPU LPDDR5X ECC memory per module, alongside an impressive up to 96GB GPU HBM3 per module, delivering unmatched memory capabilities for the most demanding tasks. Compatibility with NVIDIA BlueField-2 / BlueField-3 DPUs further enhances its versatility, making it a powerhouse for advanced networking and security features.
Connectivity is no less exceptional, with 8 x 10Gb/s BASE-T LAN ports powered by Intel® X550-AT2, 4 x dedicated management ports, a CMC port, and extensive storage and expansion options including 16 x 2.5" Gen4 NVMe hot-swappable bays, 8 x M.2 slots with PCIe Gen5 x4 interface, 4 x FHHL PCIe Gen5 x16 slots, and 4 x OCP 3.0 Gen5 x16 slots. This system is fortified with a triple 3000W (240V) 80 PLUS Titanium redundant power supply, ensuring reliability and efficiency.
The VELOCITY N218G, with its focus on the revolutionary NVIDIA Grace Hopper Superchip, represents the forefront of server technology, offering a direct liquid cooling solution that redefines performance standards for giant-scale AI and HPC deployments.
Higher Performance and Faster Memory—Massive Bandwidth for Compute Efficiency
The NVIDIA GH200 Grace Hopper™ Superchip is a breakthrough accelerated CPU designed from the ground up for giant-scale AI and high-performance computing (HPC) applications. The superchip delivers up to 10X higher performance for applications running terabytes of data, enabling scientists and researchers to reach unprecedented solutions for the world’s most complex problems.
VELOCITY R226A
Introducing the VELOCITY R226A, a state-of-the-art 2U server that represents the zenith of computing power and efficiency, optimized for high-density GPU acceleration and designed to meet the demands of modern data centers and AI workloads. This powerhouse supports 4 x AMD Instinctâ„¢ MI250 OAM GPU modules, fully harnessing the capabilities of AMD Infinity Fabric Links technology for seamless interconnectivity and unmatched data throughput.
At the core of the VELOCITY R226A are the AMD EPYCâ„¢ 7003 Series processors, including variants equipped with AMD 3D V-Cacheâ„¢ Technology, utilizing dual processor configuration with cutting-edge 7nm technology. This setup is designed to deliver exceptional computational performance and efficiency.
AMD Instinct MI300X Platform
The AMD Instinct MI300X Platform integrates 8 fully connected MI300X GPU OAM modules onto an industry-standard OCP design via 4th-Gen AMD Infinity Fabricâ„¢ links, delivering up to 1.5TB HBM3 capacity for low-latency AI processing. This ready-to-deploy platform can accelerate time-to-market and reduce development costs when adding MI300X accelerators into existing AI rack and server infrastructure.
KEY WORKLOADS
ADVANCED AI TRAINING PLATFORMS FOR NEXT-LEVEL MODEL DEVELOPMENT AND DEPLOYMENT
AI Training and Development
Deep Learning Workloads
Machine Learning Operations (MLOps)
Neural Network Optimization
WANT TO KNOW MORE?
Customized Solutions for Diverse AI Training Requirements
Peak AI Performance
Visual Computing Excellence
Reliable & Manageable
It's a critical step in AI development that determines the model's ability to make accurate predictions and perform tasks effectively in the real world.Â
Training requires substantial computational power and sophisticated algorithms to refine the model’s 'understanding'. The result is an AI that can not only comprehend complex datasets but also evolve its knowledge base over time, providing the foundation for reliable and insightful AI inference applications..Â
VELOCITY N218G
Introducing the VELOCITY N218G, a pioneering solution in the realm of direct liquid cooling, meticulously engineered for the colossal demands of AI and HPC environments. This marvel of technology is a 2U 4-node rear access server system, embodying the pinnacle of performance and innovation. At the heart of the VELOCITY N218G lies the NVIDIA Grace Hopper Superchip, a testament to cutting-edge technology, providing an astounding 900GB/s NVLink-C2C Interconnect that sets a new benchmark for data processing and transfer speeds.
Designed with a CPU+GPU architecture, it ensures an unparalleled computing experience, supporting up to 480GB CPU LPDDR5X ECC memory per module, alongside an impressive up to 96GB GPU HBM3 per module, delivering unmatched memory capabilities for the most demanding tasks. Compatibility with NVIDIA BlueField-2 / BlueField-3 DPUs further enhances its versatility, making it a powerhouse for advanced networking and security features.
Connectivity is no less exceptional, with 8 x 10Gb/s BASE-T LAN ports powered by Intel® X550-AT2, 4 x dedicated management ports, a CMC port, and extensive storage and expansion options including 16 x 2.5" Gen4 NVMe hot-swappable bays, 8 x M.2 slots with PCIe Gen5 x4 interface, 4 x FHHL PCIe Gen5 x16 slots, and 4 x OCP 3.0 Gen5 x16 slots. This system is fortified with a triple 3000W (240V) 80 PLUS Titanium redundant power supply, ensuring reliability and efficiency.
The VELOCITY N218G, with its focus on the revolutionary NVIDIA Grace Hopper Superchip, represents the forefront of server technology, offering a direct liquid cooling solution that redefines performance standards for giant-scale AI and HPC deployments.
Higher Performance and Faster Memory—Massive Bandwidth for Compute Efficiency
The NVIDIA GH200 Grace Hopper™ Superchip is a breakthrough accelerated CPU designed from the ground up for giant-scale AI and high-performance computing (HPC) applications. The superchip delivers up to 10X higher performance for applications running terabytes of data, enabling scientists and researchers to reach unprecedented solutions for the world’s most complex problems.
VELOCITY R226A
Introducing the VELOCITY R226A, a state-of-the-art 2U server that represents the zenith of computing power and efficiency, optimized for high-density GPU acceleration and designed to meet the demands of modern data centers and AI workloads. This powerhouse supports 4 x AMD Instinctâ„¢ MI250 OAM GPU modules, fully harnessing the capabilities of AMD Infinity Fabric Links technology for seamless interconnectivity and unmatched data throughput.
At the core of the VELOCITY R226A are the AMD EPYCâ„¢ 7003 Series processors, including variants equipped with AMD 3D V-Cacheâ„¢ Technology, utilizing dual processor configuration with cutting-edge 7nm technology. This setup is designed to deliver exceptional computational performance and efficiency.
AMD Instinct MI300X Platform
The AMD Instinct MI300X Platform integrates 8 fully connected MI300X GPU OAM modules onto an industry-standard OCP design via 4th-Gen AMD Infinity Fabricâ„¢ links, delivering up to 1.5TB HBM3 capacity for low-latency AI processing. This ready-to-deploy platform can accelerate time-to-market and reduce development costs when adding MI300X accelerators into existing AI rack and server infrastructure.