(Yicai Global) Sept. 13 -- AliCloud, a cloud service platform affiliated with Alibaba Group Holding Ltd. [NYSE:BABA], has launched a new generation of its heterogeneous acceleration platform, capable of providing global multi-scenario capabilities for the artificial intelligence sector, online news outlet Huanqiu reported.
The release marks the debut of AliCloud's heterogeneous computing family, which comprises six heterogeneous instances including a GPU and a field-programmable gate array (FPGA) that can meet the computing demands of complex applications, including image rendering, high-performance computing, and AI.
This new platform is especially significant for the AI field, where it can halve deep learning costs and lower the threshold for AI computing.
"AliCloud hopes to provide the AI industry with the most powerful computing capabilities. We have opened up such services as video recognition, voice recognition, image recognitionand the ET Industrial Brain and Medical Brain scenario solutions," said Zhang Xiantao, head of heterogeneous computing at AliCloud. "Today, our new platform will further improve the ecology of the AI industry and meet the core business demands of more customers,"
Many services within Alibaba, such as the Pailitao image search app, have achieved excellent computing acceleration results through heterogeneous computing. These solutions are now being exported through AliCloud. Continuous innovation in AliCloud's infrastructure, distributed in 13 regional data centers around the world, can provide AI companies with high-performance, elastic-scaling, and service-sustainable acceleration engines.
Heterogeneous computing technology was first introduced to the world in the mid-1980s. The technology has become one of the most popular research fields in parallel/distributed computing thanks to its cost-effectiveness in obtaining high-performance computing capabilities, strong scalability, highly efficient use of computing resources, and tremendous development potential.