Giga Computing Announces New GPU Servers for Intel Xeon and NVIDIA Platforms
Giga Computing, a subsidiary of GIGABYTE and a leading provider of high-performance servers, server motherboards, and workstations, has unveiled two new GPU servers designed for the 4th Gen Intel Xeon Scalable CPU and NVIDIA HGX H100 4- and 8-GPU platforms. These servers, the GIGABYTE G363-SR0 and GIGABYTE G593-SD2, offer optimal performance and cooling efficiency, catering to processors and GPUs with high TDP requirements. They provide the highest GPU compute power available in their respective sizes.
In addition, Giga Computing will be qualifying the new NVIDIA L40S GPU, which is ideal for multi-purpose workloads such as AI, graphics, and video. They will also be qualifying the new workstation-oriented NVIDIA Ada Generation GPUs: the NVIDIA RTX A5000, NVIDIA RTX A4500, and NVIDIA RTX 4000.
G363-SR0: Direct Liquid Cooling
The G363-SR0 is specifically designed for direct liquid cooling (DLC) to efficiently remove heat from both Intel Xeon processors and the four NVIDIA GPU modules. With a 1:2 CPU-to-GPU ratio, this 3U server delivers exceptional compute performance for AI and HPC workloads. It features eight 2.5" bays for high-speed data access and six PCIe Gen 5 low-profile slots for advanced networking. The system is powered by a 2 + 1 redundant power supply configuration using 3000 W titanium power supplies.
DLC technology allows for faster performance and greater energy savings compared to traditional air-cooled systems. By transferring heat to facility water, which is then carried out of the data center, the associated cooling infrastructure can be reduced, resulting in lower costs and improved ROI.
G593-SD2: 8 GPU System
The G593-SD2 is designed for GPU-accelerated computing and is ideal for AI training and compute-heavy workloads. This 5U chassis features a split design, with the top 1U dedicated to dual Intel Xeon CPUs and the bottom 4U for NVIDIA GPUs. This design ensures top-tier performance without the need for advanced cooling technologies.
GIGABYTE is committed to meeting customer and partner needs, and based on feedback, made a small adjustment to the G593 series. The G593-SD2 model removes low-profile slots and replaces them with FHHL slots required for NVIDIA BlueField DPUs. This change allows the G593-SD2 to leverage multiple BlueField cards for accelerated AI and HPC data movement and enhanced security isolation.
Support for Accelerators
Giga Computing will support NVIDIA's new accelerators by qualifying its enterprise solutions. The NVIDIA L40S GPU, based on the NVIDIA Ada Lovelace architecture, offers breakthrough multi-workload acceleration for data centers. The new NVIDIA RTX Ada Generation GPU workstations target professional users who require high-performance ray tracing, AI-accelerated compute, and professional graphics rendering.
Giga Computing's commitment to powerful GPU solutions extends throughout the year, with more servers supporting liquid cooling via included kits. This allows customers to reduce energy costs and create greener data centers.
