Revolutionary fluid PCIe/CXL devices
H3 solutions allow you to dynamically compose or decompose PCIe/CXL devices to create your own composable IT infrastructure that meets your diverse application needs in a dynamic business environment.
H3 solutions allow you to dynamically compose or decompose PCIe/CXL devices to create your own composable IT infrastructure that meets your diverse application needs in a dynamic business environment.
The H3 Center management UI or API automatically discovers devices that are added, moved, or removed from the system. It also checks accelerators/memory utilization and performance and then performs continuous real-time analysis for better resource utilization.
With the right provisioning of resources, capital expenditure can be reduced with your initial purchase, other hardware purchases, and upgrades.
Future-proof PCIe Gen5/CXL solutions to protect your investment in the next 3-5 years and provide 64 GB bandwidth and sub-microsecond latency to boost your applications. PCIe Gen5/CXL is also backward-compatible with PCIe Gen4 for your existing PCIe devices.
GPU-accelerated memory/storage solution built of the PCIe Gen6 fabric and Gen6 SSDs, delivering up to 288 million IOPS at 512 bytes for intensive AI workloads such as Vector database or Graph Neural Network (GNN)
The PCIe Gen5 GPU Solution transforms generative AI with independent GPU deployment and a composable architecture, delivering cost-effective, high-performance solutions for data-intensive tasks. Breakthrough GPU functionalities enhance system dynamics, offering advanced memory management and real-time insights for optimal performance and flexibility.
CXL 2.0 Memory Pooling Solution that provides up to 6TB pooled memory, supporting memory expansion, pooling, and sharing for greater flexibility, system compatibility, and better data transition efficiency
Falcon 4205, a PCIe Gen4 expansion enclosure that provides up to 8 host connections for 4 GPUs and an additional PCIe device, grants dynamic provisioning capability for greater flexibility, device compatibility, and better system RAS.
Falcon 4210, a PCIe Gen4 expansion enclosure that provides up to 16 host connections and 5U space for diverse PCIe devices, grants dynamic provisioning capability for greater flexibility, device compatibility, and better system RAS.
With composable AI, GPUs can be assigned or reassigned to any connected servers to rapidly meet the needs of variable AI applications or different AI development stages. Users can provision GPUs from a pool of GPU resources using the management GUI or API. The result: higher GPU utilization, reduced infrastructure costs, and the flexibility to adapt instantly as AI workloads evolve.
Groundbreaking one-stop station from design to rendering
If you want to apply for any product display, please write a form and we will contact you after receiving the message.