Started in April 2019, Beijing VirtAI Technology Co., Ltd. was founded by a number of high-tech talents who were former heads of the China Research Institutes of the Fortune 500 and had a common vision, focusing on providing global leading data-center class AI accelerator virtualization and resource management software and solutions for customers cross the world.
With innovative technology, VirtAI Tech's Orion AI computing platform helps customers build datacenter class AI accelerator resource pools, so that user applications can transparently share and use AI accelerators on any server in the data center without modification, which can not only help customers improve resource utilization and can greatly facilitate the deployment of user applications. A bunch of artificial intelligence, Internet and top public cloud companies have begun to use the Orion AI computing platform.
Adhering to the rigorous working spirit, taking "software to promote progress and innovation to achieve high efficiency" as its mission, VirtAI Tech is serving customers from China to the whole world with its partners, enriching the corporate culture, continually establishing a good corporate social image, and trying to become the China's leading local virtualization software company.
OrionX software is a system software that provides GPU resource pooling and GPU virtualization capabilities for AI applications in cloud or data centers and CUDA applications. Through the efficient communication mechanism, AI applications and CUDA applications can run on any physical machine in the cloud or data center, Container or VM without the need to mount a physical GPU. At the same time provide these applications with hardware computing power in the GPU resource pool.
Provides great flexibility for the deployment of AI applications and CUDA applications in the cloud and data centers. There is no limit on the location of the GPU server and the number of resources.
Orion GPU resources are allocated when the AI application and CUDA application start, and are automatically released when the application exits. Reduce GPU idle time and improve the shared GPU turnover rate.
Compatible with existing AI applications and CUDA applications, it still has the performance of using GPU acceleration.
Through the management and optimization of GPU resource pools, the utilization rate and throughput rate of GPUs in the entire cloud and data center are improved.
Through unified management of GPUs, GPU management complexity and cost are reduced.
B-809，No.8，Haidian North 2nd Street，Beijing 100080，China