依據歐盟施行的個人資料保護法，我們致力於保護您的個人資料並提供您對個人資料的掌握。
按一下「全部接受」，代表您允許我們置放 Cookie 來提升您在本網站上的使用體驗、協助我們分析網站效能和使用狀況，以及讓我們投放相關聯的行銷內容。您可以在下方管理 Cookie 設定。 按一下「確認」即代表您同意採用目前的設定。
News Center
May. 13, 2024
GIGABYTE to Showcase at ISC High Performance 2024
May 13, 2024 - GIGABYTE Technology (TWSE: 2376) and Giga Computing, a subsidiary of GIGABYTE and an industry leader in enterprise solutions, announced today its participation at ISC 2024 (May 12-16). At the GIGABYTE booth #H20 is a new rack-scale solution, GIGA POD, which is a groundbreaking AI supercomputing infrastructure designed to accelerate innovations in enterprise AI.
GIGABYTE has played a pivotal role in providing technology leaders with a robust supercomputing infrastructure housed in powerful GIGABYTE accelerated compute platforms. These platforms, using x86 processors, house NVIDIA HGX™ H100 systems integrating eight NVIDIA H100 Tensor Core GPUs, offering unparalleled processing power for AI workloads.
GIGA POD is a turnkey service designed to streamline the creation of AI supercomputing clusters. With professional assistance, organizations can effortlessly create a cohesive unit of racks interconnected to form a powerful computing environment. This AI ecosystem platform thrives with parallel processing, facilitated by the blazing- fast communication enabled by NVIDIA NVLink™ technology.
Key features of GIGA POD include:
- Industry Connections: GIGABYTE maintains close collaborations with industry giants such as AMD, Intel, and NVIDIA. This ensures a swift response to customer requirements and timelines, guaranteeing the timely delivery of cutting-edge solutions.
- Depth in Portfolio: GIGABYTE boasts a comprehensive portfolio of servers, including GPU, compute, storage, and high-density solutions. With numerous SKUs tailored for diverse enterprise applications, GIGABYTE offers unmatched versatility to meet the unique needs of every customer.
- Scale Up or Out: GIGA POD is designed for scalability, allowing organizations to effortlessly scale computing resources as needed. Whether adding new nodes or processors, GIGA POD enables seamless integration, ensuring optimal performance and efficiency.
- High Performance: GIGABYTE's servers and rack designs are engineered for peak performance, from single GPU servers to entire clusters. Optional liquid cooling further enhances performance, ensuring maximum efficiency even for highly demanding workloads.
- Experienced: With a proven track record in deploying large GPU clusters, GIGABYTE brings a wealth of experience to every project. Its expert team is ready to guide customers through the deployment process, providing personalized timelines and support to meet their specific requirements.
- Proven Success Case: In 2021, Taiga Cloud, Europe's first and largest dedicated Generative AI Cloud Service Provider, partnered with GIGABYTE to construct its inaugural AI cluster, powered by NVIDIA HGX™ A100 system. By Q4 2023, Taiga Cloud and GIGABYTE led Europe in deploying NVIDIA HGX™ H100 clusters. This collaborative effort has elevated Taiga Cloud's capacity to deliver flexible, secure, and compliant clean energy computing power, driving generative AI innovations across Europe.
GIGABYTE’s new rack-scale service is not done here. The groundwork has been laid to also support and design GIGA POD’s infrastructure around the AMD Instinct™ MI300X Accelerator with AMD Infinity Fabric™. This baseboard platform hosts eight AMD MI300X GPUs and 1.5 TB of HBM3 memory, making it more than capable of handling demanding HPC and AI workloads, especially AI inference.
GIGABYTE showcases much more as hot innovations require a cool solution. To overcome the challenge of today’s hiking power consumption in data centers, advanced cooling designs are needed with the benefits of greater operational stability and improved PUE (Power Usage Effectiveness). GIGABYTE's air, liquid, and immersion cooling solutions are ready to deploy to fit various IT strategies and infrastructures.
GIGABYTE's Advanced Cooling Solutions leverages innovative designs and cutting-edge materials to redefine performance and efficiency in computing. By optimizing thermal management, these solutions enhance system stability and longevity, ensuring peak performance even under the most demanding workloads.
The ISC Student Cluster Competition 2024:
This year is the 13th edition of the competition. The goal is to give STEM university students the chance to show off their skills in HPC and machine learning applications. A team from National Tsing Hua University in Taiwan, led by Prof. Zhou Zhiyuan, will compete using servers and hardware provided by GIGABYTE. Students were given one G493-SB1 GPU server with ten NVIDIA H100 PCIe GPUs and three R183-S90 computing servers. The competition will be for students to use their software skills to improve the efficiency of the systems under certain power constraints to achieve the best performance. Based on past competitions, GIGABYTE expects them to performwell at ISC.
Stay Cool, Power Up: Experience Innovation in HPC and AI Solutions at GIGABYTE Booth #H20