Quick Specs
- Type: Entry-Level AI Inference GPU
- GPU Architecture: NVIDIA Ampere (GA107)
- CUDA Cores: 1280
- Tensor Cores: 40 (3rd Generation)
- RT Cores: 10 (2nd Generation)
- Memory: 16 GB GDDR6
- Memory Interface: 128-bit
- Memory Bandwidth: 200 GB/s
- Interface: PCIe Gen4 x8
- Form Factor: Single Slot, Half-Height, Half-Length (HHHL)
- Cooling: Passive (requires system airflow)
- Power Consumption: Configurable TDP (40–60W)
- Base Clock: 1440 MHz
- Boost Clock: 1770 MHz
- ECC Support: Enabled by default
- vGPU Support: NVIDIA vPC, vApps, vWS, AI Enterprise, vCS
Part Number:
- H98K0
The Dell NVIDIA A2 16GB GPU (H98K0) is engineered for entry-level AI inference workloads, offering a balance of performance and efficiency. Built on the NVIDIA Ampere architecture, it features 1280 CUDA cores, 40 third-generation Tensor Cores, and 10 second-generation RT Cores. With 16 GB of GDDR6 memory and a memory bandwidth of 200 GB/s, the A2 delivers robust performance for tasks such as intelligent video analytics, natural language processing, and other AI-driven applications.
Designed for versatile deployment, the A2's low-profile, single-slot HHHL form factor and passive cooling solution make it suitable for a wide range of systems, including edge servers and space-constrained environments. Its configurable TDP ranging from 40W to 60W allows for optimized power consumption based on workload requirements. Additionally, the A2 supports NVIDIA's suite of virtual GPU software, enabling efficient resource utilization in virtualized environments.