AMD Instinct MI325X 8-GPU OAM Server (Lenovo ThinkSystem SR680a V3)
Brand: AMD | Category: GPUs
SKU: 7DHB | Part #: 7DHB | MPN: 7DHB
Contact for Pricing — Request a Quote
About the AMD Instinct MI325X 8-GPU OAM Server (Lenovo ThinkSystem SR680a V3)
The AMD Instinct MI325X 8-GPU OAM Server delivers 2048 GB (2 TB) of HBM3E memory with 6.0 TB/s bandwidth per GPU, making it a high-capacity accelerator platform engineered for demanding AI and HPC workloads. Built on the Lenovo ThinkSystem SR680a V3 8U rack server, this system combines eight AMD Instinct MI325X GPUs in OCP Accelerator Module form factor with AMD's CDNA 3 architecture to enable enterprise-grade AI training, large language model inference, and scientific computing at scale.
The system supports a comprehensive range of precision formats—FP64, FP32, FP16, BF16, FP8, INT8, and INT4—allowing flexibility across mixed-precision workloads. AMD ROCm open-source GPU compute stack and framework support for PyTorch, TensorFlow, JAX, HIP, and OpenCL provide broad software compatibility for development and production deployments. Connectivity is enabled via PCIe Gen 5 host interface and AMD Infinity Fabric GPU interconnect, while Lenovo XClarity Administrator delivers unified management across the platform. For IT procurement teams and AI infrastructure teams evaluating high-performance GPU servers, this configuration (part number 7DHB) combines AMD's architectural advantages with Lenovo's proven server reliability. Contact Omnixon Global to request a quotation and explore deployment options tailored to your environment.
Key Specifications
- Manufacturer: AMD
- GPU Model: AMD Instinct MI325X
- GPU Architecture: AMD CDNA 3
- GPUs per Server: 8 x OAM (OCP Accelerator Module)
- Memory per GPU: 256 GB HBM3E
- Total GPU Memory: 2048 GB (2 TB)
- Memory Bandwidth per GPU: 6.0 TB/s
- Supported Precision Formats: FP64, FP32, FP16, BF16, FP8, INT8, INT4
- GPU Interconnect: AMD Infinity Fabric
- Host Interface: PCIe Gen 5
- Part Number: 7DHB
Technical Specifications
| Brand | AMD |
| Category | GPUs |
| SKU | 7DHB |
| Part Number | 7DHB |
| Condition | New |
| GPU Model | AMD Instinct MI325X |
| Manufacturer Part Number (MPN) | 7DHB |
| Server Platform | Lenovo ThinkSystem SR680a V3 |
| GPU Architecture | AMD CDNA 3 |
| GPUs per Server | 8 x OAM (OCP Accelerator Module) |
| Memory per GPU | 256 GB HBM3E |
| Total GPU Memory (8-GPU System) | 2048 GB (2 TB) |
| Memory Bandwidth per GPU | 6.0 TB/s |
| Supported Precision Formats | FP64, FP32, FP16, BF16, FP8, INT8, INT4 |
| GPU Interconnect | AMD Infinity Fabric |
| Host Interface | PCIe Gen 5 |
| Software Platform | AMD ROCm (open-source GPU compute stack) |
| Framework Support | PyTorch, TensorFlow, JAX, HIP, OpenCL |
| Form Factor | OAM (OCP Accelerator Module) in 8U rack server |
| Target Workloads | AI training, LLM inference, HPC, scientific computing |
| Management Software | Lenovo XClarity Administrator |
| Product Generation | ThinkSystem V3 platform |
Frequently Asked Questions about AMD Instinct MI325X 8-GPU OAM Server (Lenovo ThinkSystem SR680a V3)
What server platforms accept the AMD Instinct MI325X 8-GPU OAM Server (Lenovo ThinkSystem SR680a V3)?
Reference servers include Dell PowerEdge XE9680 / XE9712, HPE Cray XD670, Lenovo ThinkSystem SR685a / SR675 V3, Supermicro AS-A21GE / SYS-821GE, Gigabyte G593 / G894, ASUS ESC. Share your target platform in the RFQ and we will confirm chassis-to-GPU compatibility and recommended NIC pairing.
How long is the lead time on AI GPUs?
Highly model-dependent. L40S / RTX-class: typically 3-6 weeks. H100/H200/B200 in SXM form factor: 12-16 weeks for whole-platform allocations. We quote genuine-channel ETAs only — no grey-market promises.
Do you supply matched networking (Quantum InfiniBand / Spectrum-X)?
Yes — Omnixon stocks the full NVIDIA networking lineup (Quantum-2 / Quantum-X InfiniBand, Spectrum-X Ethernet, ConnectX NICs, BlueField DPUs) so we can quote a complete training-cluster BOM, not just the GPUs.
Can you help with NVIDIA AI Enterprise licensing?
Yes. We hold genuine-channels for NVIDIA AI Enterprise software subscriptions. Add it to your RFQ and we quote node-aligned licensing along with the hardware.