Lenovo Nvidia A2 Ampere 16GB GDDR6 Graphic Card

Lenovo NVIDIA A2 Ampere 16GB GDDR6 AI GPU Accelerator Graphic Card | PG179 | 03GX831 | 900-2G179-2720-001

The Lenovo NVIDIA A2 Ampere GPU Accelerator is a compact, power-efficient Tensor Core GPU engineered for AI inference, edge computing, and virtual desktop infrastructure (VDI) deployments. Built on NVIDIA’s Ampere architecture , the A2 delivers exceptional performance-per-watt, making it ideal for dense server environments and low-power enterprise systems.

Identified by Lenovo and NVIDIA part numbers PG179 , 03GX831 , and 900-2G179-2720-001 , this professional-grade accelerator provides 16GB of ECC-protected GDDR6 memory and advanced Tensor Core acceleration without requiring external power connectors.

Key Specifications

GPU Architecture: NVIDIA Ampere

CUDA Cores: 1,280

Tensor Cores: 40 (3rd Generation)

RT Cores: 10 (2nd Generation)

GPU Memory: 16GB GDDR6 with ECC

Memory Interface: 128-bit

Memory Bandwidth: 200 GB/s

Performance Highlights

Peak FP32 Performance: Up to 4.5 TFLOPS

TF32 Tensor Performance: Up to 9 TFLOPS (18 TFLOPS with sparsity)

FP16 Tensor Performance: Up to 18 TFLOPS (36 TFLOPS with sparsity)

INT8 Performance: Up to 36 TOPS (72 TOPS with sparsity)

INT4 Performance: Up to 72 TOPS (144 TOPS with sparsity)

Physical & Power Details

Form Factor: Single-slot, low-profile PCIe (full-height bracket supported)

Interface: PCIe 4.0 x8

Interconnect Bandwidth: 32 GB/s

Cooling: Passive heatsink (requires server airflow)

Power Consumption: Configurable 40W–60W TDP

Power Connector: None required (slot-powered)