Giga Computing is dedicated to improving human life through technological advancements of computing. In response to the growing demands for artificial intelligence, high-performance computing, and cloud computing, we uphold an innovative spirit and continuously exploring new server application markets, particularly focusing on AI&HPC servers with GPU modular architecture, ongoing development of x86 and ARM64 architecture servers, and high-density servers that enhance computing power within limited spaces. Additionally, we are focusing on the autonomous development of direct liquid cooling (DLC) technology and immersion cooling technology. Our efforts in diversification and integration within the industry ecosystem have secured our position as a market leader.
High-efficiency and Energy-saving Technology
To address the increasing demand for data centers and the global trend toward carbon reduction, Giga Computing has developed server cooling solutions using DLC and immersion cooling.
The DLC solution involves immersing operating components like CPUs or GPUs in sealed pipelines filled with coolant. This approach eliminates the need for fans and heat sinks typically required in air-cooled servers. By efficiently transferring the heat generated from high-performance operations out of the server, this method allows for the installation of more processors in a confined space, helping customers achieve costs savings, reduced energy consumption and noise, and enhanced system performance in high-density IT deployments.
The immersion cooling solution is suitable for any enterprise application. With immersing servers in a cooling tank, where the heat generated by high-performance computing is conducted into a non-conductive cooling liquid. The cooling liquid distribution system then transfers the heat to either air or liquid cooling pipes, maintaining the servers' high-efficiency operation. The immersion cooling solution system does not require the use of pumps or sprinklers, significantly reducing equipment failure rates and maintenance needs. This approach not only lowers operating costs but also benefits the environment, with simultaneous upgrades in energy efficiency.
During the product lifecycle of a server, the phase with the highest carbon emissions is the usage stage, followed by the raw materials stage. After evaluating different cooling solutions during the usage stage, it has been confirmed that the immersion cooling solution offers the best energy-saving and carbon reduction effects, followed by the DLC solution.
|
|
|
|
|
|
Innovative Product Applications
As one of the earliest server suppliers to invest in the R&D of advanced cooling technology, Giga Computing has been continuously innovating in recent years. Beyond servers and liquid tank solutions designed for immersion cooling, the market will soon welcome our comprehensive contact liquid cooling computing solutions. These offering will not only feature dedicated servers that compatible with various cooling equipment brands but also include our proprietary cold plates, cooling circulation equipment, and management systems. These measures aim to provide customers with a convenient one-stop procurement and service solution, assisting them in achieving their sustainability objectives. At the same time, they help Giga Computing master key computing infrastructure technologies. This not only expands our business scope but also lays a solid foundation for the development of IT computing equipment for various future scenarios, accumulating valuable experience and a rich database.
Innovative Data Center Deployment and Design Technologies
|
Building on the existing technology leadership of L6 (barebones system) and L10 (fully assembled system), we have innovatively developed the L11 cabinet rack server array designed for data centers. With the rise of AI cloud computing, which differs from traditional mainstream cloud data centers, there is a need for multiple GPU AI servers to form foundational working groups, maximizing the data processing performance of AI. The current development deploys a structure where 8 GIGA PODs form a foundational working group. Each group is equipped with 32 G593 servers, totaling 256 NVIDIA H100 Tensor Core GPUs, achieving 1 exaflop (one quintillion floating-point operations per second) of FP8 floating-point performance. We will later offer diversified GIGA POD solutions, incorporating AMD INSTINCTTM GPU AI servers, to significantly enhance AI workload capabilities. |
GIGA POD AI Data Center | |
 |
Giga Computing is dedicated to improving human life through technological advancements of computing. In response to the growing demands for artificial intelligence, high-performance computing, and cloud computing, we uphold an innovative spirit and continuously exploring new server application markets, particularly focusing on AI&HPC servers with GPU modular architecture, ongoing development of x86 and ARM64 architecture servers, and high-density servers that enhance computing power within limited spaces. Additionally, we are focusing on the autonomous development of direct liquid cooling (DLC) technology and immersion cooling technology. Our efforts in diversification and integration within the industry ecosystem have secured our position as a market leader.
High-efficiency and Energy-saving Technology
To address the increasing demand for data centers and the global trend toward carbon reduction, Giga Computing has developed server cooling solutions using DLC and immersion cooling.
The DLC solution involves immersing operating components like CPUs or GPUs in sealed pipelines filled with coolant. This approach eliminates the need for fans and heat sinks typically required in air-cooled servers. By efficiently transferring the heat generated from high-performance operations out of the server, this method allows for the installation of more processors in a confined space, helping customers achieve costs savings, reduced energy consumption and noise, and enhanced system performance in high-density IT deployments.
The immersion cooling solution is suitable for any enterprise application. With immersing servers in a cooling tank, where the heat generated by high-performance computing is conducted into a non-conductive cooling liquid. The cooling liquid distribution system then transfers the heat to either air or liquid cooling pipes, maintaining the servers' high-efficiency operation. The immersion cooling solution system does not require the use of pumps or sprinklers, significantly reducing equipment failure rates and maintenance needs. This approach not only lowers operating costs but also benefits the environment, with simultaneous upgrades in energy efficiency.
During the product lifecycle of a server, the phase with the highest carbon emissions is the usage stage, followed by the raw materials stage. After evaluating different cooling solutions during the usage stage, it has been confirmed that the immersion cooling solution offers the best energy-saving and carbon reduction effects, followed by the DLC solution.
Innovative Product Applications
As one of the earliest server suppliers to invest in the R&D of advanced cooling technology, Giga Computing has been continuously innovating in recent years. Beyond servers and liquid tank solutions designed for immersion cooling, the market will soon welcome our comprehensive contact liquid cooling computing solutions. These offering will not only feature dedicated servers that compatible with various cooling equipment brands but also include our proprietary cold plates, cooling circulation equipment, and management systems. These measures aim to provide customers with a convenient one-stop procurement and service solution, assisting them in achieving their sustainability objectives. At the same time, they help Giga Computing master key computing infrastructure technologies. This not only expands our business scope but also lays a solid foundation for the development of IT computing equipment for various future scenarios, accumulating valuable experience and a rich database.
Innovative Data Center Deployment and Design Technologies
Building on the existing technology leadership of L6 (barebones system) and L10 (fully assembled system), we have innovatively developed the L11 cabinet rack server array designed for data centers. With the rise of AI cloud computing, which differs from traditional mainstream cloud data centers, there is a need for multiple GPU AI servers to form foundational working groups, maximizing the data processing performance of AI. The current development deploys a structure where 8 GIGA PODs form a foundational working group. Each group is equipped with 32 G593 servers, totaling 256 NVIDIA H100 Tensor Core GPUs, achieving 1 exaflop (one quintillion floating-point operations per second) of FP8 floating-point performance. We will later offer diversified GIGA POD solutions, incorporating AMD INSTINCTTM GPU AI servers, to significantly enhance AI workload capabilities.
| GIGA POD AI Data Center |
|