The NGC-Ready program was developed to validate servers to run AI workloads from training to inference. To qualify, NGC-Ready servers are required to have passed an extensive suite of tests that verify their ability to deliver high performance when running NGC containers. NGC-Ready system validation includes tests of:
- Single and multi-GPU DL training using TensorFlow, PyTorch, and NVIDIA DeepStream Transfer Learning Toolkit.
- High volume, low latency inference using NVIDIA® TensorRT™, TensorRT Inference Server, and DeepStream.
- Data science using RAPIDS and XGBoost.
- Application development using the CUDA® Toolkit.
NGC-Ready for Edge servers have, in addition to passing the NGC-Ready tests, demonstrated their ability to support the NVIDIA EGX platform that uses both TPM for hardware-based key management and IPMI for remote systems management industry standards.