The NVIDIA A100 PCIE Single Card is a high-performance computing accelerator designed for data centers and other demanding workloads. It is a PCIe Gen4 card that supports up to 80 GB of HBM2 memory and delivers up to 6x the performance of its predecessor, the NVIDIA V100.
Product Description
The NVIDIA A100 PCIE Single Card is designed to accelerate a wide range of workloads, including artificial intelligence, data analytics, and high-performance computing. It features the NVIDIA A100 GPU, which includes third-generation Tensor Cores and multi-instance GPU technology, allowing it to handle multiple workloads simultaneously.
Product Parameters
- Architecture: NVIDIA Ampere
- CUDA Cores: 6,912
- Tensor Cores: 432
- Memory: 40/80 GB HBM2
- Memory Bandwidth: 1.6 TB/s
- Form Factor: PCIe Gen4
- Thermal Solution: Passive
Specifications
- Peak Single Precision (FP32) Performance: 19.5 TFLOPS
- Peak Double Precision (FP64) Performance: 9.7 TFLOPS
- Peak Tensor Operation Performance: 624 TFLOPS (mixed-precision)
- Power Consumption: 250 W – 400 W
- Interface: PCIe Gen4 x16
- Dimensions: 10.5″ x 4.4″
- Operating System Support: Windows, Linux
