Supermicro AS-2145GH-TNMR-LCC — 4x AMD Instinct MI300A APU (2U)
Converged CPU+GPU APUs with unified memory for HPC and AI in 2U.
We help you choose, configure, and deliver the right system — no obligation.
Configuration at a Glance
Tailored per engagement. Full technical overview below.
Configuration Options
Core specifications for this system. Every component is configurable to your workload — request a quote for a tailored build.
4x AMD Instinct MI300A APU, 128GB unified HBM3 each (512GB total)
96 'Zen 4' cores total (24 per APU)
Unified, fully coherent CPU+GPU HBM3
8x 2.5in hot-swap NVMe (SAS3/SATA3 optional)
Overview
The Supermicro AS-2145GH-TNMR-LCC fits four AMD Instinct MI300A APUs — combining Zen 4 CPU cores and CDNA 3 GPU compute on one coherent package — into a liquid-cooled 2U node with 512GB of unified HBM3. Nexus Compute specifies, configures, and tests these converged systems on ROCm, delivering them warranty-backed through authorized channels for memory-coherent HPC and AI codes.
Who This Solution Is For
Business Benefits
Unified coherent memory
Shared CPU-GPU HBM3 removes explicit host-device copies, simplifying code and cutting data-movement overhead.
Compute density in 2U
Four APUs deliver converged CPU and GPU performance in a compact, liquid-cooled rack node.
Tested HPC platform
Validated on ROCm with HPC libraries so simulation and AI codes run correctly on arrival.
Typical Business Use Cases
Coupled HPC simulation (CFD, weather, molecular dynamics)
Memory-coherent scientific AI workloads
Mixed CPU/GPU research pipelines
Space-constrained dense compute deployments
Industry Applications
Technical Overview
Built on the Supermicro AS-2145GH-TNMR-LCC 2U platform with four MI300A APUs in SH5 sockets, each fusing 24 Zen 4 cores and CDNA 3 GPU compute units with 128GB of coherent HBM3 (512GB unified per node). Liquid cooling and 4x 2700W Titanium power support sustained operation, with PCIe Gen5 expansion and AIOM networking.
| Accelerator | 4x AMD Instinct MI300A APU, 128GB unified HBM3 each (512GB total) |
| Integrated CPU | 96 'Zen 4' cores total (24 per APU) |
| Memory Architecture | Unified, fully coherent CPU+GPU HBM3 |
| Storage | 8x 2.5in hot-swap NVMe (SAS3/SATA3 optional) |
| Networking/Fabric | AIOM (OCP NIC 3.0); PCIe Gen5 x16 FHHL slots |
| Cooling | Liquid-cooled (LCC) |
| Form Factor | 2U rackmount |
| Power | 4x 2700W Titanium redundant |
Specifications are indicative and configured to each engagement. Request a quote for a configuration tailored to your requirements.
Warranty, Support & Fulfillment
Every system ships from an authorized channel, configured and tested, with the documentation enterprise buyers need — backed by warranty and a dedicated account team.
Enterprise Warranty
Full manufacturer warranty with optional on-site, next-business-day support and extended coverage.
Authorized Channel
Sourced through Tier-1 distribution and OEM partners — never grey market. Asset & warranty records included.
Lead Time & Deployment
48-hour quotes, then configured, burn-in tested, and delivered on a committed schedule.
Nationwide Fulfillment
Coordinated logistics, rack-and-stack, and delivery wherever your infrastructure lives.
Frequently Asked Questions
How does MI300A differ from MI300X?
MI300A is an APU that integrates Zen 4 CPU cores with CDNA 3 GPU compute and shared coherent memory, ideal for tightly coupled HPC. MI300X is a GPU-only accelerator with 192GB HBM3 aimed at large-model AI. We match the part to your workload.
What does unified memory mean for our code?
CPU and GPU share the same HBM3 address space, so there are no explicit host-to-device transfers. Coupled simulation and AI codes often run with simpler memory management and less overhead.
Is this suitable for both HPC and AI?
Yes. The MI300A targets converged HPC and AI; it excels at simulation and memory-coherent workloads. For pure large-model serving, our MI300X servers are the better fit.
Hardware Assistance
Configure the Supermicro AS-2145GH-TNMR-LCC — 4x AMD Instinct MI300A APU (2U) with Nexus Compute
Tell us your requirements and a hardware specialist will help you specify, configure, and quote the right system — typically within two business days. No obligation.