Supermicro ARS-221GL-NHIR 2NVMe
- Condition
- New
- Manufacturer
- Supermicro
- Form factor
- 2U
- RAM
- 960GB (LPDDR5x)
- GPU
- 2 x NVIDIA GH200 (Hopper)
- Accelerator manufacturer
- NVIDIA
- Accelerator model
- NVIDIA GH200
- Installed GPU/APU count
- 2 onboard GPUs
- Maximum GPU/APU count
- 2 onboard GPUs
- Memory per accelerator
- 144GB
- Total accelerator memory
- 288GB
- GPU memory type
- HBM3e
- Accelerator memory bandwidth
- Up to 5 TB/s per GPU (10 TB/s aggregate)
- Accelerator form factor
- Onboard GPUs integrated in two NVIDIA GH200 Grace Hopper Superchips
- Accelerator interconnect
- NVIDIA NVLink between the two GH200 Superchips
- Accelerator-to-CPU interface
- NVIDIA NVLink-C2C
- Supported compute data types
- FP64, FP32, TF32, BF16, FP16, FP8, INT8
- Accelerator virtualization
- NVIDIA MIG, GPU passthrough
- MIG support
- Up to 7 isolated GPU instances per GPU
- Software ecosystem
- NVIDIA CUDA; NVIDIA AI Enterprise-compatible
- Maximum system power
- 4kW
- Cooling
- Air cooling with up to 6 removable heavy-duty 60 mm fans
- Recommended workloads
- HPC, AI and deep learning training and inference, large language models and generative AI
- Dimensions
- 87 x 438.4 x 900 mm (H x W x D)
- Weight
- 40 kg
Need a technical consultation on choosing a server?
Submit a requestSupermicro ARS-221GL-NHIR 2NVMe is a 2U enterprise server storage configuration with two NVMe drives installed in the front E1.S storage section. It is designed for AI factories, research centers, cloud providers and enterprise data centers that need fast local storage for AI and HPC workloads.
Key features:
-
Storage basket and service access: The selected storage basket comprises two NVMe drives installed in the front E1.S storage section. Front placement keeps the drives accessible during planned maintenance without describing the fixed positions as hot-swap.
-
Installed accelerator subsystem: The installed accelerator subsystem contains 2 x NVIDIA GH200 Grace Hopper Superchip. It provides 288GB HBM3e GPU memory in aggregate. NVLink-C2C connects each CPU-GPU pair, while NVIDIA NVLink links the two GH200 Superchips.
-
Storage topology and data path: This two-drive configuration uses two of the platform’s three front fixed E1.S NVMe positions, leaving one position available. Two internal M.2 NVMe slots provide additional boot or service-storage options.
-
Enterprise compute platform: The installed compute subsystem combines two 72-core NVIDIA Grace CPUs with 960GB of ECC LPDDR5X memory. It supplies the CPU, memory and PCIe resources required to feed the accelerator subsystem and host production services.
-
Management and availability: BMC health monitoring and a dedicated 1GbE management port, together with 4 x 2000W redundant (2+2) 80 PLUS Titanium power supplies, support remote administration, component monitoring, planned service operations and resilient data-center operation.
Use cases:
-
AI model development: Uses the installed NVIDIA GH200 Grace Hopper Superchip subsystem for model development, training and fine-tuning workloads suited to the accelerator architecture.
-
Inference and production services: Provides parallel compute and accelerator memory for LLM, multimodal and enterprise inference services.
-
HPC, simulation and analytics: Accelerates scientific computing, numerical simulation, engineering, analytics and other parallel workloads supported by the installed accelerator subsystem.
-
Data-intensive pipelines: Combines the specified 2NVMe storage basket with high-bandwidth accelerator and network paths for datasets, checkpoints, staging and local scratch storage.
Supermicro ARS-221GL-NHIR 2NVMe is a precisely defined 2NVMe storage configuration for organizations that need this drive basket, installed NVIDIA GH200 Grace Hopper Superchip acceleration and dependable enterprise lifecycle management.
to Buy a Server
from Servermall
warranty
testing
We want to make sure that our equipment is fully reliable for you. That’s why all our equipment (even new) is subject to complete diagnostics, testing, and part replacement.
testing
We fully load our equipment with various utilities and software. Rest assured, it’ll work for a very long time.
Power supply
The components work at full capacity and are tested for several days. If we are not sure about something, we replace it.
of all equipment
Our regulations imply the same testing for both used and new equipment. Oh, and by the way, we complement the official manufacturer’s warranty with our own warranty.