Translation. Region: Russian Federal
Source: People's Republic of China in Russian – People's Republic of China in Russian –
An important disclaimer is at the bottom of this article.
Source: People's Republic of China – State Council News
BEIJING, Aug. 7 (Xinhua) — Chinese tech giant Huawei has announced the full open source code of its CANN (computing architecture for neural networks) software toolkit for its Ascend chips, to support users in independently unlocking the chips' deep potential and developing customized solutions.
CANN serves as a bridge between Ascend's high-level artificial intelligence (AI) training systems and chips, allowing its users to access computing power without having to deal with chip-level complexities.
The announcement was made at the Ascend computing industry summit held in Beijing on Tuesday, where representatives from leading AI enterprises, universities, research institutes and Huawei partners launched an open-source CANN ecosystem initiative to consolidate industry efforts, explore emerging AI capabilities and jointly build a thriving Ascend chip ecosystem.
Huawei's rotating chairman Xu Zhijun said computing power remains at the core of the company's AI strategy.
Almost every major company that makes high-performance cards to provide computing power has a similar architecture. However, unlike the closed-source CUDA ecosystem developed by American chipmaker Nvidia for almost two decades, Huawei's CANN has taken a completely different path by going open source.
Huawei told Xinhua that after seven years of development, CANN has made key breakthroughs in computing optimization, communication efficiency, and memory management. It is now able to provide computing power support throughout the entire process of training and deploying AI models.
The release of the open source code for CANN is a recent step by Huawei to disseminate its own technologies to the public. The company has previously published the source code for its HarmonyOS operating system and MindSpore artificial intelligence platform. -0-
Please note: This information is raw content obtained directly from the source of the information. It is an accurate report of what the source claims and does not necessarily reflect the position of MIL-OSI or its clients.
.