News

Internet Infrastructure

GMO GPU Cloud Adds Monitoring Dashboard Functionality Utilizing Grafana
~Visualize and Optimize GPU Server Resource Usage~

GMO Internet, Inc., a member of the GMO Internet Group (President and CEO: Tadashi Ito, hereinafter “GMO Internet”), has launched a monitoring dashboard feature for the “GMO GPU Cloud,” a GPU cloud service for generative AI.
This new feature leverages Grafana(*1), an open-source data visualization platform, to provide intuitive visualizations of GPU server resource usage and Slurm(*2) job execution status. By monitoring usage patterns of multiple jobs(*3) and workloads, users can optimize their GPU resource utilization more efficiently.

This feature is available at no additional cost (*4) with the “Dedicated Plan” of the GMO GPU Cloud.

(*1) An open-source interactive data visualization platform.
(*2) A type of job scheduler mainly used in supercomputers and large-scale research computing.
(*3) A job is a collection of instructions including what program to run, what data to use, how long to run, and how much memory to allocate.
(*4) Not available with the Shared Plan.

Background of the Monitoring Dashboard Featurediv>

In operations where jobs are submitted repeatedly on a daily basis, tracking the execution status of individual jobs or analyzing usage data over the long term can be challenging.
To address these issues and provide an intuitive environment for visualizing and analyzing usage, GMO Internet has developed this monitoring dashboard feature using Grafana.

Use Cases for the Monitoring Dashboard

The dashboard allows users to view the following information on a single screen:

・Number of running jobs in each partition
・Number of pending jobs in each partition
・Number of running jobs per user
・Number of pending jobs per user

Example 1:
Visualize the time zones during which jobs are executed. If usage is concentrated in certain periods, redistribute execution times to balance the load.

Example 2:
Monitor GPU utilization rates within a specific timeframe and adjust parameters to optimize resource efficiency.

【About GMO GPU Cloud】
(URL:https://gpucloud.gmo/

The GMO GPU Cloud is a high-performance GPU cloud service equipped with NVIDIA H200 GPUs, high-speed storage, and Japan’s first commercial deployment of the ultra-fast “NVIDIA Spectrum-X” network. (*5)In multi-node configurations, it delivers the fastest performance among commercial clouds in Japan.
In the TOP500 global supercomputer rankings announced in November 2024, it ranked 37th globally and 6th in Japan, marking the highest performance for a commercial cloud service in the country.

Its exceptional performance and stability in multi-node environments make it ideal for advanced AI development, including training of large language models (LLMs) and development of multimodal AI.
Additionally, GMO Internet plans to launch a cloud service using the next-generation NVIDIA Blackwell Ultra GPU with NVIDIA HGX B300 (hereafter “NVIDIA B300”) within 2025.

Through this service, GMO Internet aims to provide optimized infrastructure and flexible, customizable compute environments tailored to customers' workloads, thereby shortening development cycles, reducing costs, and promoting the growth of Japan’s AI industry.

(*5) As of November 22, 2024, based on internal research.
(Reference: GMO GPU Cloud ranked in TOP500 global supercomputers for the first time)https://www.gmo.jp/news/article/9266/

【Future Outlook】

GMO Internet will continue contributing to technological innovation in the rapidly advancing fields of AI and robotics through its AI infrastructure strategy centered around the GMO GPU Cloud.
By continuously providing the latest AI compute infrastructure and building flexible cloud environments tailored to user needs, GMO Internet strives to become an indispensable cloud service in Japan’s AI industry, fostering AI innovation from Japan.

■過去の参考リリース

2024年4月19日GMO GPU Cloud Launches Japan’s Fastest Generative AI GPU Cloud Service with NVIDIA H200 Tensor Core GPUs
June 11, 2024GMO GPU Cloud Becomes the First Cloud Provider in Japan to Adopt NVIDIA Spectrum-X for Generative AI
August 29, 2024GPU Cloud Usage Survey: Domestic Usage at Only 5.4%, Nearly 90% Rely on Overseas Services
September 26, 2024Performance of NVIDIA H200 GPU Environment Successfully Verified
November 13, 2024GMO GPU Cloud Showcases Infrastructure and Security for the AI and Robotics Era at the NVIDIA AI Summit
November 19, 2024GMO GPU Cloud Makes Its First Entry into the Global TOP500 Supercomputer Rankings
November 22, 2024GMO GPU Cloud Makes Its First Entry into the Global TOP500 Supercomputer Rankings
February 21, 2025GMO GPU Cloud Adds Multi-Instance GPU (MIG) Feature Powered by NVIDIA Technology
May 7, 2025GMO GPU Cloud Officially Selected as the AI Infrastructure for AIRoA’s Next-Generation Robot Development
May 12, 2025GMO GPU Cloud Chosen for Turing’s Multimodal Generative AI Platform for Autonomous Driving
May 14, 2025GMO GPU Cloud Secures Additional Investment for Expansion
June 11, 2025GMO GPU Cloud Ranks 34th Globally and 1st in Japan for Power Efficiency in Green500
July 1, 2025GMO Internet, Macnica, and NVIDIA Partner to Accelerate Generative AI Development with High-Speed GPU Cloud
August 4, 2025GMO GPU Cloud Adopts NVIDIA Blackwell Ultra GPU
October 2, 2025GMO GPU Cloud Begins Technical Trials for Next-Gen Decentralized AI Infrastructure Using IOWN

Inquiries

This website contains machine translations and may not always be accurate. Please note that it may differ from the original Japanese text.