Workflow
High - Performance Computing (HPC)
icon
Search documents
Hyper-responsive load management system for hyperscalers: MARA and TAE Power Solutions partner for first-of-its-kind grid efficiency platform
Globenewswire· 2025-06-25 12:05
Core Insights - MARA and TAE Power Solutions have announced a strategic collaboration to develop a high-frequency, real-time responsive load management system aimed at meeting the energy demands of hyperscale data centers and cryptocurrency mining operations [1][6] - The collaboration will leverage TAE Power Solutions' proprietary technology to create a 10MW clean energy storage network that provides microsecond-responsive load balancing [2][5] - The TAE Power Solutions platform is based on fusion energy research and integrates advanced hardware, real-time control, and machine learning algorithms for efficient energy management [3][5] Company Overview - MARA is focused on deploying digital energy technologies to enhance global energy systems and convert excess energy into digital capital [7][8] - TAE Power Solutions aims to provide clean energy solutions and has developed technologies for efficient energy storage and management, originally designed for TAE's fusion research [8] Technology and Innovation - The TAE Power Solutions D-Series BESS is designed for large commercial and industrial applications, functioning as a grid-responsive energy module that can be integrated into existing facilities [4] - The system continuously monitors battery performance metrics on a microsecond scale, enabling precision load balancing and fast frequency response essential for grid stability [5] Market Implications - The collaboration addresses the increasing energy demands of advanced technologies like AI and high-performance computing, ensuring that power systems can support growth without straining local electrical grids [6] - Prototypes of the new system are expected to be deployed by late summer 2025, with larger-scale commercialization anticipated to begin in early 2026 [6]
一家公司,单挑网络芯片三巨头
半导体行业观察· 2025-06-04 01:09
Core Viewpoint - Cornelis Networks has launched the CN5000 series, a 400Gbps Omni-Path technology aimed at enhancing AI and HPC performance, with plans for future UltraEthernet integration [1][19]. Group 1: Product Overview - The CN5000 series is designed for AI and HPC applications, supporting deployments of up to 500,000 endpoints and providing advanced lossless data transmission and congestion avoidance features [19][20]. - The CN5000 utilizes Omni-Path architecture, which includes credit-based flow control and dynamic fine-grained adaptive routing to ensure high performance and reliability [20][22]. - The series includes both air-cooled and liquid-cooled switch options, addressing the cooling needs of large AI companies [5][22]. Group 2: Performance Metrics - The CN5000 demonstrates superior performance in HPC workloads, achieving up to 30% faster execution in computational fluid dynamics and climate modeling compared to InfiniBand NDR [22]. - For AI applications, the CN5000's collective communication speed is six times faster than RoCE, enhancing the efficiency of large language model training [19][22]. Group 3: Future Developments - The upcoming CN6000 series, set for release in 2026, will integrate Omni-Path with Ethernet support, doubling bandwidth capabilities [16][21]. - The CN7000 series, expected in 2027, will combine UltraEthernet standards with Omni-Path architecture, targeting the most demanding AI and HPC environments [17][21]. Group 4: Market Positioning - Cornelis Networks aims to provide a cost-effective alternative to NVIDIA's InfiniBand, capitalizing on the growing demand for high-performance networking solutions in AI and HPC sectors [18][19].
AEM Expands Access to Production-Proven SLT and Burn-In Ecosystem for Advanced Computing Customers
Globenewswire· 2025-05-19 23:00
Core Insights - AEM Holdings Ltd. collaborates with Intel Foundry to enhance access to System-Level Test (SLT) and Burn-In capabilities for advanced computing devices, aiming to accelerate time-to-market for fabless customers [1][2][5] Group 1: Collaboration Details - The partnership leverages decades of collaboration to create a mature SLT and Burn-In ecosystem, essential for high-performance computing (HPC) and artificial intelligence (AI) applications [1][2] - AEM provides device-specific configurable test units, advanced handlers, and software support, complementing Intel Foundry's services such as factory automation and test program development [3][9] Group 2: Customer Benefits - Key benefits for customers include enhanced test coverage, reduced capital expenditure, and faster time-to-market, supported by a scalable platform and local engineering teams [5][6][9] - The collaboration aims to meet rising performance demands and facilitate chiplet-based architectures, ensuring reliability for AI and HPC applications [5][6] Group 3: Company Overview - AEM Holdings Ltd. is a global leader in test innovation with a presence across Asia, Europe, and the United States, offering comprehensive semiconductor and electronics test solutions [7][8]