スケーラブルシステムズ List of News Articles

Scalable Systems Corporation has released information about servers and workstations equipped with the latest GPU H200 NVL and next-generation Blackwell architecture. We are introducing the latest products that are optimal for AI, HPC, and advanced analysis applications. ▼For more details, click here https://www.sstc.co.jp/products/hp2c_server_workstation.html #H200NVL #Blackwell #AI #HPC #ScalableSystems

Scalable Systems' high-end workstation "KRONOS series" now supports the latest Blackwell generation GPUs, including the NVIDIA RTX PRO 6000 Blackwell Workstation Edition. The Blackwell architecture is a next-generation architecture for AI, HPC, and graphics processing that surpasses the traditional Ampere and Ada generations, achieving significant advancements in both computational performance and power efficiency. ■ Features - Standard equipped with a native 12VHPWR power connector (up to 600W) - High-speed bandwidth with PCI Express 5.0 ×16 - High-reliability power supply (1200W–1300W, Titanium/Gold certified) - High-performance water cooling system for stable temperature control of GPU/CPU - ECC-compatible memory and IPMI remote management, ideal for professional use In the rapidly advancing fields of AI development and generative AI, the KRONOS WS provides an optimal platform to maximize the performance of next-generation GPUs. For detailed product specifications and customization inquiries, please contact us.

At Scalable Systems, we have released a comparison document reflecting the latest information on the NVIDIA H200 NVL and RTX PRO Blackwell series. H200 NVL: Optimized for large-scale AI processing with HBM3e memory. Blackwell RTX PRO: The latest architecture for workstation applications. Comparison table: Get an overview of the key specs of H100 / H200 / RTX PRO (Blackwell) at a glance. Please refer to the comparison table on our page linked below under "Related Links."

IT solutions supported not only by technology but also by the power of "people." Scalable Systems values providing support that closely aligns with each customer's individual situation, going beyond merely offering products. By combining advanced technical skills with flexible and sincere responsiveness, we deliver peace of mind in all phases, from implementation to operation and maintenance. Being a "partner you can consult with immediately when in trouble" is the value we prioritize above all else. Our company has established a "one-stop system" that consistently handles everything from requirements definition, design and construction, to post-implementation operational support and troubleshooting, not just product supply. We possess the technical skills and execution capabilities to respond flexibly and swiftly to on-site challenges, and we have extensive experience in verifying and implementing both domestic and international products. We cherish our customers' voices, think through challenges together, and respond with technology. That is our approach. Scalable Systems aims to be a trusted technological partner for our customers, evolving continuously with both product strength and service strength as our dual pillars. For the latest company information, please refer to the "Company Information and Various Materials Download" in the related links below.

The NVIDIA H200 NVL is a GPU based on the NVIDIA Hopper architecture, designed for next-generation AI and HPC. It is the world's first to feature HBM3e memory, achieving faster and larger memory bandwidth compared to the previous generation H100 NVL, making it capable of handling unprecedented workloads in generative AI, natural language processing (LLM), scientific simulations, and data analysis. We have started offering GIGABYTE GPU servers equipped with the NVIDIA H200 NVL. GIGABYTE's G-series servers are characterized by a high-density and highly scalable design to maximize AI performance. Supported CPU platforms: - Latest Intel Xeon 6 series (6700P/6500P) - AMD EPYC 9004 series (high core count and many PCIe lanes available) Models equipped with the H200 NVL will be proposed with optimal server configurations based on your applications, environment, and budget. Currently, only representative configurations are listed on our webpage linked below, but we are always open to customization inquiries.

Next-Generation Storage Infrastructure for the AI Era In the rapidly evolving fields of AI/ML and HPC, the demand for massive data processing and real-time performance is increasing. The VDURA V5000 All Flash Appliance is a high-performance, high-reliability storage platform designed to meet these needs. ■ Accelerate AI Processing with Ultra-Fast I/O Performance With all storage based on flash (NVMe/SAS SSD), it achieves throughput and low latency that are tens of times greater compared to traditional HDD-based systems. ■ Flexible Scalable Architecture A scale-out design that allows for incremental expansion on a node-by-node basis. ■ High Availability and Redundant Configuration Minimizes the risk of data loss during failures through distributed redundancy (RAID-like) and data replication. ■ Optimized for AI/ML and HPC Supports RDMA and IB for high-speed integration with GPU clusters and AI inference platforms. ■ Intuitive Management Interface High-functionality operational management tools compatible with dedicated Web GUI and REST API. ■ Space and Power Saving for Cost Optimization Significantly reduces space and power compared to HDD-based systems.

This is a product line of rack servers, storage servers, and GPU servers equipped with Intel Xeon 6 processor family. You can choose from a wide range of product models tailored to your applications, budget, and configurations. Therefore, we cannot list all available product models on our website, but we will provide the optimal processor model and product model suited to your needs. The Intel Xeon 6 processor family offers performance and flexibility to handle diverse workloads, with the option to choose from two types of CPU microarchitectures: P-cores and E-cores. The Intel Xeon 6 6700P/6500P series features P-cores optimized for high computational workload processing, supporting up to 86 cores per socket and configurations of 1, 2, 4, or 8 sockets per system. In addition to traditional DDR5-6400, it also supports MRDIMM, achieving high bandwidth of up to 8800 MT/s. Furthermore, with support for PCIe 5.0 and up to 88 lanes, a single socket configuration can utilize up to 136 lanes, significantly accelerating I/O.