Workflow
longsys江波龙:以创新存储技术引领AI时代变革

Core Insights - The rapid development of artificial intelligence (AI) technology has significantly increased the importance of enterprise storage as a key support for AI computing platforms [1] - Jiangbolong, a leading semiconductor storage company in China, is driving a technological revolution in enterprise storage with its innovative SOCAMM product [1] Group 1: Product Development and Performance - Jiangbolong has successfully collaborated with top clients to develop SOCAMM, showcasing its strong technical reserves and forward-looking market layout in next-generation server memory technology [3] - SOCAMM offers a performance leap with advanced LPDDR5X technology, providing over 2.5 times the bandwidth of traditional DDR5 RDIMM at the same capacity, effectively eliminating data transmission bottlenecks in AI training and inference [3] - The product features a 20% reduction in latency by integrating the storage controller with memory units, crucial for real-time data processing scenarios [3] Group 2: Energy Efficiency and Design Innovation - SOCAMM's low voltage characteristic (1.1V) results in power consumption being only one-third of standard DDR5 RDIMM, significantly reducing energy usage in data centers [3] - The compact modular design of SOCAMM is one-third the size of standard RDIMM, supporting high-density deployment and optimizing heat dissipation [5] - The innovative 4-N-4HDI stacking technology enhances hole density by over 10 times, providing a solid physical foundation for SOCAMM's 8×16bit multi-channel architecture [5] Group 3: Market Position and Future Outlook - Jiangbolong emphasizes deep collaboration with clients, offering customized storage solutions under the PTM model to meet diverse customer needs [5] - The company's enterprise storage products are known for high reliability, stability, and low latency, widely used in data centers, cloud computing, and AI [5] - Jiangbolong plans to continue focusing on the enterprise storage market, driving innovation to contribute to the advancement of AI computing platforms [5]