NVIDIA A16
The NVIDIA A16 is a data center graphics card based on the Ampere architecture. Released on 2021-04-12 with a launch price of $2,500.00, it features 5120 CUDA cores and 64 GB of GDDR6 memory.
General info about A16
| Architecture |
Ampere (GA107) |
| Market Segment |
Data Center |
| Release Date |
2021-04-12 |
| Launch Price |
$2,500.00 |
| Manufacturing Process |
8nm |
Technical specs of A16
| CUDA Cores |
5120 |
| Base Clock Speed |
900 MHz |
| Boost Clock Speed |
1400 MHz |
| Transistor Count |
34.8B |
| VRAM Capacity |
64 GB GDDR6 |
| Memory Bus Width |
512 bits |
| Memory Bandwidth |
888.0 GB/s |
| TDP |
250 W |
Key technical parameters of the NVIDIA A16 include its 5120 CUDA cores, 900 MHz base clock, and 1400 MHz boost clock, delivering high performance for data center applications.
Benchmark performance of A16
| FP32 (float) |
15 TFLOPS |
| FP16 (half) |
60 TFLOPS |
The NVIDIA A16 achieves 15 TFLOPS in FP32 performance, making it a top contender in data center workloads.
Gaming performance of A16
The NVIDIA A16 is not designed for gaming.
Features and connectivity of A16
| Tensor_cores |
Gen 3 |
| Mig |
True |
| Ports |
PCIe Gen4 |
Physical dimensions of A16
| Length |
10.5in |
| Width |
4.4in |
| Height |
2-slot |
Rankings and efficiency of A16
| Ranking Position |
#18 |
| Popularity Ranking |
#22 |
| Cost Effectiveness |
1.9/5 |
| Power Efficiency |
1.8/5 |