Gigabyte G894-SD3-AAX7 8NVMe
- Condition
- New
- Manufacturer
- GIGABYTE
- Form factor
- 8U
- Remote-Control-Modul
- GIGABYTE Server Management (GSM)
- CPU
- 2x Intel Xeon 6767P
- RAM
- 16x 128GB RDIMM 6400MHz (Support up to 32 DIMM xrtov DDR5 up to 8192 Gb)
- Power supply unit
- 8x 3000W Titanium
- HDD
- 2 x 1.92TB M.2/ 2x 3.84TB U.2 (up to 8 HDD 2.5'' NVMe)
- NIC card
- MCX755106AS-HEAT
- GPU
- B300 HGX
- Accelerator manufacturer
- NVIDIA
- Accelerator model
- NVIDIA B300 (HGX B300)
- Installed GPU/APU count
- 8
- Maximum GPU/APU count
- 8 SXM GPUs
- Memory per accelerator
- 288GB
- Total accelerator memory
- 2304GB
- GPU memory type
- HBM3E
- Accelerator memory bandwidth
- 8 TB/s per GPU
- Accelerator form factor
- SXM
- Accelerator interconnect
- Fifth-generation NVIDIA NVLink with NVSwitch; 1.8 TB/s per GPU
- Accelerator-to-CPU interface
- PCIe 5.0 through the NVIDIA HGX baseboard
- Supported compute data types
- FP64, FP32, TF32, BF16, FP16, FP8, FP6, FP4 and INT8
- Accelerator virtualization
- NVIDIA vGPU, SR-IOV and GPU passthrough
- MIG support
- Up to 7 isolated GPU instances per GPU
- Software ecosystem
- NVIDIA CUDA, NVIDIA AI Enterprise and NVIDIA Transformer Engine
- Maximum system power
- 36kW
- Cooling
- Air cooling; passive SXM modules with dedicated GPU-tray airflow and hot-swap fan modules
- Recommended workloads
- AI reasoning, large-scale generative AI training and inference, LLM development, HPC, simulation, analytics and data science
- Dimensions
- 351 x 447 x 923 mm (H x W x D)
Need a technical consultation on choosing a server?
Submit a requestGigabyte G894-SD3-AAX7 8NVMe is an 8U enterprise GPU system storage configuration with 8 front 2.5-inch PCIe Gen5 NVMe hot-swap positions. It is designed for AI factories, research centers, cloud providers and enterprise data centers that need fast local storage for accelerated workloads.
Key features:
-
Storage basket and service access: The chassis provides 8 front-accessible 2.5-inch PCIe Gen5 NVMe hot-swap positions. Supported drives can be serviced without opening the chassis, helping reduce planned maintenance time.
-
Accelerator subsystem: The installed accelerator subsystem contains 8 x NVIDIA B300 SXM (HGX B300), providing 2,304GB (8 x 288GB) of aggregate GPU memory. Fifth-generation NVIDIA NVLink with NVSwitch; 1.8 TB/s per GPU.
-
Storage topology and data path: The listed assembly includes two 3.84TB U.2 NVMe SSDs in the front basket and retains capacity for up to eight front 2.5-inch NVMe drives. 2 x 1.92TB M.2 SSDs installed; 2 x 3.84TB U.2 NVMe SSDs installed in the front basket. This layout supports fast dataset staging, checkpoints and local scratch storage without mixing the basket with a different drive protocol.
-
Enterprise compute platform: The compute subsystem is configured with 2 x Intel Xeon 6767P processors and 2,048GB DDR5 RDIMM (16 x 128GB, 6400 MT/s); 32 DIMM platform. It supplies CPU, memory and PCIe resources for the accelerator complex and production services.
-
Management and availability: ASPEED AST2600 BMC and dedicated management LAN; 8 x 3000W 80 PLUS Titanium power supplies in the listed assembly. These facilities support remote administration, health monitoring and resilient operation.
Use cases:
-
AI model development: Uses the installed NVIDIA B300 SXM (HGX B300) subsystem for training, fine-tuning and large-model development.
-
Inference and production AI: Provides parallel compute and accelerator memory for LLM, multimodal and enterprise inference services.
-
HPC, simulation and analytics: Accelerates scientific computing, numerical simulation, analytics and other parallel workloads supported by NVIDIA B300 SXM (HGX B300).
-
Data-intensive pipelines: Combines 8 front 2.5-inch NVMe positions with high-bandwidth GPU and network paths for datasets, checkpoints, staging and local scratch space.
Gigabyte G894-SD3-AAX7 8NVMe is a purpose-built 8U platform for organizations that need 8 front 2.5-inch PCIe Gen5 NVMe bays, 2,304GB (8 x 288GB) of installed GPU memory and enterprise-class compute and management in one deployable system.
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.