SK Telecom (SKT) has launched its AI-based B2B answer, ‘SKT Enterprise AI Cloud Supervisor,’ which optimises AI infrastructure sources and allows the built-in administration of AI growth environments. AI Cloud Supervisor is a business product developed from SKT’s experience in managing and working large-scale GPU sources.
Additionally Learn: SK Telecom to Monitor Fibre Community Standing With AI
AI Cloud Supervisor
AI Cloud Supervisor is a GPU cluster-specialised answer primarily based on an AI job scheduler, designed to boost GPU useful resource effectivity. By managing quite a few GPU sources as in the event that they had been a single laptop, the answer maximises GPU efficiency and reduces AI growth studying time. SKT famous that when a number of jobs are requested in a cloud atmosphere, the AI Job Scheduler effectively allocates restricted sources to make sure easy job execution.
Actual-time Monitoring and Useful resource Optimisation
As large-scale knowledge studying requires GPU sources, optimising these sources is crucial for AI cloud options. For instance, GPU utilization and efficiency are monitored in real-time for every undertaking, and unused GPUs are detected and reassigned to different tasks, SKT defined. Moreover, useful resource effectivity is improved by prioritising the allocation of cloud sources to high-priority tasks.
Past optimising GPU utilization, AI Cloud Supervisor gives a complete MLOps (Machine Studying Operations) atmosphere, supporting all levels of AI growth—from knowledge preprocessing and mannequin coaching to deployment and inference.
In accordance with SKT, the platform helps large-scale parallel studying, enhancing velocity by distributing knowledge throughout a number of computer systems. Its web-based interface (Internet UI) permits a number of builders to collaborate on tasks while not having particular software program installations.
Additionally Learn: NVIDIA AI Aerial: Merging Wi-fi Networks and Generative AI on a Unified Platform
Kim Myung-guk, head of SKT Cloud CO, stated, “With the discharge of ‘AI Cloud Supervisor,’ firms making ready to implement AI providers will be capable to maximise GPU useful resource efficiency and effectively develop AI providers” and “Sooner or later, we are going to stably provide AI knowledge facilities and GPU servers, and we will even present important options to handle them to steer the AI knowledge middle answer enterprise discipline.”
Giant-Scale GPU Cluster Collaboration
SK Telecom has just lately collaborated with Hana Monetary to determine a large-scale GPU cluster utilizing AI Cloud Supervisor, enabling Hana Monetary Group to fulfill rising AI expertise calls for, the official launch stated.
