Menu
Menu
The NVIDIA® A100 Tensor Core GPU delivers unprecedented acceleration at every scale for AI, data analytics, and HPC to tackle the world’s toughest computing challenges. As the engine of the NVIDIA® data center platform, A100 can efficiently scale up to thousands of GPUs or, using new Multi-Instance GPU (MIG) technology, can be partitioned into seven isolated GPU instances to accelerate workloads of all sizes. A100’s third-generation Tensor Core technology now accelerates more levels of precision for diverse workloads, speeding time to insight as well as time to market.
| Architecture | Ampere |
| Process Size | 7nm |
| Transistors | TSMC |
| Die Size | 54 Billion |
| CUDA Cores | 826 mm2 |
| Streaming Multiprocessors | 6912 |
| Tensor Cores | 108 |
| Multi-Instance GPU (MIG) Support | Gen 3 |
| FP64 | 432 |
| FP64 Tensor Core | Yes, up to seven instances per GPU |
| FP32 | 9.7 TFLOPS |
| TF32 Tensor Core | 19.5 TFLOPS |
| BFLOAT16 Tensor Core | 19.5 TFLOPS |
| FP16 Tensor Core | 156 TFLOPS |
| INT8 Tensor Core | 312 TFLOPS* |
| INT4 Tensor Core | 312 TFLOPS |
| NVLink | 624 TFLOPS* |
| NVLink Interconnect | 312 TFLOPS |
| GPU Memory | 624 TFLOPS* |
| Memory Interface | 624 TOPS |
| Memory Bandwidth | 1248 TOPS* |
| System Interface | 1248 TOPS |
| Thermal Solution | 2496 TOPS* |
| vGPU Support | 2-Way Low Profile, 2-Slot |
| Secure and Measured Boot Hardware Root of Trust | 600 GB/s Bidirectional |
| NEBS Ready | 80GB HBM2e ECC on by Default |
| Power Connector | 5120-bit |
| Maximum Power Consumption | 1555 GB/s |
| PCIe 4.0 x16 | |
| Passive | |
| NVIDIA® Virtual Compute Server with MIG support | |
| CEC 1712 | |
| Level 3 | |
| 8-pin CPU | |
| 300 W |
Privacy & Cookie Policy: By accessing this website you agree to accept cookies. Read the Privacy Policy.