Specifications
MPN
TCS-HPE-DL380A-G12
Model
HPE ProLiant Compute DL380a Gen12
Form Factor
4U rackmount
GPU Options
8x double-wide, RTX PRO 6000 Blackwell Server Edition
GPU Options (ext)
10x double-wide L40S / H100 NVL / H200
GPU Options (single-wide)
Up to 16x single-wide NVIDIA L4
CPU
2x Intel Xeon 6, up to 144 cores each
Memory
32x DDR5 DIMM slots, up to 4 TB
Storage
16x E3.S NVMe front bays
PCIe
PCIe Gen5 throughout, x16 GPU risers
Networking
OCP 3.0 mezzanine plus standard PCIe NIC slots
Cooling
Air cooled or Direct Liquid Cooling (DLC)
GPU Field Upgrade
Kit available to take a sub-8-GPU chassis to 8 double-wide GPUs
Management
HPE iLO 6 with Redfish
DL380a Gen12 GPU configurations
| GPU option | Form factor | Count | Typical workload |
|---|---|---|---|
| NVIDIA L4 | Single-wide | Up to 16 | High-density VDI, video transcode |
| NVIDIA L40S | Double-wide | Up to 10 | LLM inference, graphics + AI |
| NVIDIA H100 NVL | Double-wide | Up to 10 | Training, HPC |
| NVIDIA H200 (NVL) | Double-wide | Up to 10 | Long-context inference, fine-tuning |
| NVIDIA RTX PRO 6000 Blackwell Server Edition | Double-wide | Up to 8 | Omniverse, rendering, inference |
Use Case
The DL380a Gen12 is the mainstream-enterprise answer to GPU density: a 4U chassis that drops into a standard 19-inch rack without changing the facility, supporting up to eight double-wide data-centre GPUs on PCIe Gen5 x16 risers with no need for SXM baseboards or NVSwitch. Dual Intel Xeon 6 processors at up to 144 cores each and 32 DDR5 DIMM slots (4 TB) keep the host side from becoming the bottleneck, and 16 E3.S NVMe bays give high-throughput local scratch for dataset staging. It is also offered with direct liquid cooling for sites that have already deployed a CDU loop.
Need HPE ProLiant Compute DL380a Gen12?
Enterprise hardware — configured, tested, deployed. Contact us for pricing and availability.
Request Quote