On August 26, the Omni AI Computing Power Endside Cloud Collaborative Innovation Forum was held at the Shenzhen International Convention and Exhibition Center. At the forum, the group standard “General Technical Specification for End Side Artificial Intelligence Computing Power Workstations” proposed by Omni Electronics and organized by the Shenzhen Artificial Intelligence Industry Association was officially released. This is the first domestic group standard for end-side AI computing power workstations. The standard will be implemented on September 1. According to information, the current end-side AI computing power equipment industry has long had pain points such as inconsistent indicators, vague product ratings, and lack of evaluation basis, which limits the formation of technical consensus between large local models and AI intelligent hardware carriers. The standard released this time strictly complies with GB/T 1.1 preparation requirements. It sets up three levels of configuration for end-side artificial intelligence computing power workstations, covering technical requirements such as inference performance, AI agent operating environment, network disconnection availability, information security, compatibility, and environmental reliability, and is equipped with complete testing and inspection methods. Specifically, the standard clarifies hard indicators such as computing power, memory, storage, and number of supported model parameters, and regulates inference performance testing methods such as initial token latency and throughput, as well as test models, inference engines, and test parameters. Industry insiders believe that the standard provides a unified technical standard for end-side AI computing power workstations, and will provide a technical basis for product testing, model selection and procurement, and quality evaluation, and help standardize and develop the industry with high quality.

Zhitongcaijing · 4d ago
On August 26, the Omni AI Computing Power Endside Cloud Collaborative Innovation Forum was held at the Shenzhen International Convention and Exhibition Center. At the forum, the group standard “General Technical Specification for End Side Artificial Intelligence Computing Power Workstations” proposed by Omni Electronics and organized by the Shenzhen Artificial Intelligence Industry Association was officially released. This is the first domestic group standard for end-side AI computing power workstations. The standard will be implemented on September 1. According to information, the current end-side AI computing power equipment industry has long had pain points such as inconsistent indicators, vague product ratings, and lack of evaluation basis, which limits the formation of technical consensus between large local models and AI intelligent hardware carriers. The standard released this time strictly complies with GB/T 1.1 preparation requirements. It sets up three levels of configuration for end-side artificial intelligence computing power workstations, covering technical requirements such as inference performance, AI agent operating environment, network disconnection availability, information security, compatibility, and environmental reliability, and is equipped with complete testing and inspection methods. Specifically, the standard clarifies hard indicators such as computing power, memory, storage, and number of supported model parameters, and regulates inference performance testing methods such as initial token latency and throughput, as well as test models, inference engines, and test parameters. Industry insiders believe that the standard provides a unified technical standard for end-side AI computing power workstations, and will provide a technical basis for product testing, model selection and procurement, and quality evaluation, and help standardize and develop the industry with high quality.
Disclaimer:Webull uses external vendor Google Translation Service for news translations where we endeavour to ensure these are correct, however, we recommend that you please double-check this information accordingly. Webull is not responsible for translation errors or issues.