国内知名GPU厂商——长沙景嘉微电子股份有限公司于2024年3月12日发布重大公告，宣布其自主研发的“景宏系列”AI 算力产品已研发成功，即将推向市场。这一创新成果标志着国产GPU在高性能计算领域的又一重要突破。
据公告介绍，景宏系列智算模块及整机产品专为AI训练、AI推理及科学计算等应用领域设计，具备强大的运算性能。产品支持INT8、FP16、FP32、FP64等多种混合精度运算，以满足不同计算需求。此外，景宏系列还引入了先进的多卡互联技术，可实现算力的灵活扩展，为高性能计算提供更为广阔的空间。
值得注意的是，景嘉微的这款新产品能够无缝适配国内外主流CPU、操作系统及服务器厂商，兼容当前主流的计算生态、深度学习框架和算法模型库，显著缩短了用户的适配验证周期，提升了应用效率。这一特性无疑为用户提供了更大的便利，同时也增强了产品在市场中的竞争力。
景嘉微此次的成功研发，不仅展现了公司在GPU领域的技术实力，也预示着国产高性能计算解决方案的崛起，对于推动我国信息技术产业的自主可控和创新发展具有重要意义。未来，景嘉微将如何在AI算力市场中崭露头角，值得业界期待。
英语如下：
**News Title:** “Domestic Breakthrough: Jingjia Micro Successfully Develops High-Performance ‘Jinghong Series’ AI Computing Products”
**Keywords:** Jingjia Micro, GPU, AI Computing Power
**News Content:**
Title: Jingjia Micro Unveils “Jinghong Series” AI Computing Products, Pioneering a New Chapter in Domestic High-Performance Intelligent Computing
On March 12, 2024, Jingjia Micro Electronics Co., Ltd., a renowned domestic GPU manufacturer based in Changsha, issued a significant announcement that its self-developed “Jinghong Series” AI computing products have been successfully developed and are即将to be launched in the market. This innovative achievement signifies another crucial breakthrough for domestic GPUs in high-performance computing.
According to the announcement, the Jinghong series of intelligent computing modules and complete machines are tailored for AI training, AI inference, and scientific computing applications, boasting impressive computational capabilities. The product supports a variety of mixed-precision operations, including INT8, FP16, FP32, and FP64, catering to diverse computational requirements. Furthermore, the series introduces advanced multi-card interconnect technology, enabling flexible expansion of computational power and broader prospects for high-performance computing.
Notably, Jingjia Micro’s new product seamlessly integrates with both domestic and international mainstream CPUs, operating systems, and server manufacturers. It is compatible with the current mainstream computing ecosystem, deep learning frameworks, and algorithm model libraries, significantly shortening users’ adaptation and verification cycles and enhancing application efficiency. This feature undoubtedly offers greater convenience to users and strengthens the product’s competitiveness in the market.
Jingjia Micro’s successful development not only demonstrates the company’s technical prowess in the GPU field but also foreshadows the rise of domestic high-performance computing solutions. It holds significant importance for promoting self-reliance and innovation in China’s information technology industry. The industry eagerly awaits to see how Jingjia Micro will make its mark in the AI computing power market in the future.
【来源】https://www.ithome.com/0/755/583.htm
Views: 1