Rent GPU
NVIDIA A100
The NVIDIA A100 GPU delivers powerful acceleration for AI, data analytics, and high-performance computing with multi-instance GPU (MIG) support.
Memory
80GB
GPU RAM
Bandwidth
2.0TB/s
Memory Bandwidth
Form factor
SXM & NVL
Architecture
Interconnect
600GB/s
NVLink Switch
Accelerated
Workload & Computing
A100 is designed for both training and inference, enabling flexible partitioning for multiple workloads while maintaining strong performance and efficiency. A training workload like BERT can be solved at scale in under a minute by 2,048 A100 GPUs, a world record for time to solution.
- Deep Learning Training
- Inference at Scale
- High-Performance Computing
Architecture Comparison
Select the form factor optimized for your specific workload
| Feature | A100 SXM | A100 PCIe |
|---|---|---|
| architecture | Ampere | Ampere |
| memory | 80GB HBM2e | 40GB to 80GB HBM2e |
| memory Bandwidth | 2.0 TB/s | 1.9 TB/s |
| recommended For | AI Training & HPC | Enterprise AI & Data Analytics |