NVIDIA Supercharges DGX Spark With New 64GB AI Computer
According to the company newsroom, DGX Spark is designed to let developers and researchers run AI models and applications directly on a compact computer rather than relying on cloud services. The new version can run AI models with up to 100 billion parameters, according to NVIDIA.
The computer comes with NVIDIA's AI software and supports tools including Ollama, vLLM, and PyTorch, allowing users to develop and run AI applications locally.
“The new 64GB configuration gives developers, researchers and AI enthusiasts a new option for running capable AI workloads locally,” NVIDIA said.
Users can also connect two DGX Spark systems, combining their memory to 128GB and enabling them to run larger AI models and more demanding workloads.
NVIDIA said two connected 64GB systems delivered up to 1.7 times the performance of a single system in a test using the Qwen 3.8 27B model.
The company's Sync Cluster Assistant automatically configures the connection between two systems, allowing users to expand their computing capacity without changing their software setup.
NVIDIA is positioning DGX Spark for tasks including developing AI agents, running AI models, analyzing data, and creating AI applications. Its Sync Model Launcher, expected later in October, will allow users to run the Qwen3.8 27B model on one or two connected DGX Spark systems.
Legal Disclaimer:
MENAFN provides the
information “as is” without warranty of any kind. We do not accept any
responsibility or liability for the accuracy, content, images, videos,
licenses, completeness, legality, or reliability of the information
contained in this article. If you have any complaints or copyright issues
related to this article, kindly contact the provider above.

Comments
No comment