The NVIDIA A16 is an enterprise-class graphics accelerator, purpose-built for virtual desktop infrastructure (VDI) applications. It is actually made up of four discrete GPUs on a single board. Each of them is built on NVIDIA's Ampere architecture, and includes CUDA, RT, and Tensor cores along with 16GB of VRAM and dedicated video encoding and decoding engines. Across the whole card, that provides 64GB of memory - which can be split across anywhere from 4 to 64 concurrent users in VDI scenarios.
NVIDIA A16 cards are passively cooled, so they can only be used in servers specifically designed to provide plenty of airflow through the heatsinks on these cards. Since these are built for VDI, there are no direct video outputs.
NVIDIA A16 cards are passively cooled, so they can only be used in servers specifically designed to provide plenty of airflow through the heatsinks on these cards. Since these are built for VDI, there are no direct video outputs.
Model: NVIDIA NVA16TCGPUNC-KIT
Specifications
| Chipset Manufacturer | NVIDIA |
| Product Category | Data Center |
| Motherboard Connection | PCI Express 4.0 x16 |
Datacenter GPU for VDI that requires true server platforms.
