NVIDIA introduces four-node DGX Spark HPC Developer Kit at SC26
Original title: NVIDIA DGX Spark Developer Kit Turns a Desk Into a Supercomputing Lab
The switchless ring cluster combines 512GB of memory and 20TB of NVMe and can host models up to 700B parameters.
At SC26 in Chicago, NVIDIA introduced the DGX Spark HPC Developer Kit, a four-node cluster of DGX Spark systems for researchers, academics and developers. Each GB10 Grace Blackwell node links directly to two neighbors over ConnectX-7, with no network switch. NVIDIA says teams can run multi-node scientific workloads and, with NVFP4, models up to 700B parameters such as DeepSeek V4.1. The kit is on display and for sale; an HPC SDK playbook is coming soon.
- Four nodes provide 512GB of combined memory and 20TB of NVMe storage.
- Demonstrated workloads include LAMMPS with Kokkos and NAMD.
- Partners include Acer, ASUS, Dell Technologies, HP, Lenovo and MSI.
- Kvantify reports 70-qubit, 975-gate tensor-network simulations on one DGX Spark.
- Vendors
- NVIDIA
From the original article
4 min read at blogs.nvidia.com
At SC26, an annual supercomputing conference taking place this week in Chicago, NVIDIA is introducing the DGX Spark HPC Developer Kit for researchers, academics and developers. The kit is a four-node cluster of NVIDIA DGX Spark systems, each connected directly to its two neighbors in a ring, with an NVIDIA HPC SDK installation playbook coming […]
By Allen Bourgoyne · blogs.nvidia.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.