NTS HPC Building Blocks Built in the USA
HPC Compute & Fabric Building Blocks
MPI-ready CPU nodes, NDR InfiniBand or RoCE options, and scratch pairing—staged as cluster cells through NTS Cluster Solutions.
HPC Building Blocks
MPI nodes and fabric—cluster-ready SKUs
NTS HPC building blocks span compute nodes, login and management hosts, fabric options, and scratch pairing for CFD, weather, and research batch programs.
Every SKU is engineer-to-order with burn-in, MPI smoke tests, and procurement support across SEWP V, ITES-4H, and SLED vehicles—aligned with the Cluster Solutions hub for whole-cell design.
Browse HPC configurations below or ask NTS architects to scope nodes, fabric, scheduler, and scratch as one acceptance-tested program.
NTS documents rack elevation, PDU draw, thermal headroom, and driver baselines on every RFQ so facility teams are not discovering constraints at install.
Related storage, networking, and software lines can ride the same quote when your vehicle and SOW allow—keeping multi-vendor programs under one accountable integrator.
Product categories
Explore our server platforms
Rackmount, GPU, blade, storage, cluster, high-density, and Ampere — engineered for scalability and efficiency.
- MPI burn-in
- Scheduler GRES
- Fabric options
- Fat-tree layouts
- Loopback soak
- Rail optimization
NTS catalog
Available HPC configurations
Select a platform below to configure, request a quote, or add to your procurement package.
- Configure & quote
- CPU / memory / storage specs
- Contract-ready procurement
Configure & quote
CPU / memory / storage specs
Contract-ready procurement
NTS Elite NX-72GP-Liquid
Built on NVIDIA GB300 Grace™ Blackwell Ultra Superchips
72 NVIDIA Blackwell Ultra GPUs and 36 NVIDIA Grace™ CPUs
CPU and GPU connected by fifth-generation NVIDIA NVLink
Up to 17 TB LPDDR5X memory with Error-correction code (ECC)
Up to 20 TB HBM3E GPU memory with 576 TB/s bandwidth
Up to 37 TB of fast-access memory
Integrated NVIDIA ConnectX®-8 SuperNIC
Specifications shown are reference configurations—final BOM validated by NTS architects for your workload, facility power, and security baseline.
Frequently asked questions
CPU and GPU systems for simulation, modeling, and research computing - with interconnect and storage bandwidth treated as first-class design inputs.
- Not just TFLOPS sheets
- Facility gating early
Depends on existing fabric skill and latency targets. NTS sizes rails to your MPI/NCCL patterns and isolates storage traffic.
- Leaf-spine options
- Adaptive routing where supported
Many can - SEWP, ITES, and education cooperatives are common. Attachments include topology and acceptance tests.
- Milestone schedules
- Long-lead GPUs flagged
Sustained high-TDP GPU nodes and dense racks often need DLC for clock stability and facility kW. Air remains fine for lighter nodes.
- CDU planning
- Worksheets provided
