The Denvr Cloud is powered by the best.
technology overview blurb
The Denvr AI Cloud Technology
Run advanced applications efficiently, reliably, and quickly.
Industry Leading Advanced GPU Compute Infrastructure
​
Supplement oversubscribed or underpowered compute resources and reduce training & optimization times with Denvr. Seamlessly scale across multiple GPUs on one or more instances. Scalable multi-GPU support is based on NVLink A100 for 600 GB/s inter-node bandwidth. Multi-node distributed training is based on InfiniBand HDR 100 Gbps inter-node bandwidth, using Horovod or other frameworks.
Supercharged Stacks | All Major Frameworks and Drivers
​
Simplify your ML jobs. Denvr VM stacks come with all major ML frameworks and toolsets preinstalled. Tools such as PyTorch, TensorFlow, Scikit-learn, Keras, Caffe, including major deep learning frameworks Apache MXNet and NVIDIA Cuda (cuDNN).
What ever your preference, we likely support it.
Jupyter Notebooks As a Service​
​
Run your Jupyter notebooks cheaper and faster on Denvr using NVIDIA A100 MIG (multi-instance GPU) for powerful and cost-effective computational notebooks, development environments and data science.
With full support for Python and other popular languages.
GPU-Accelerated Application Catalog
​
Users can deploy containers from the NVIDIA NGCâ„¢ hub which provides pre-built and tested AI, machine learning, and HPC frameworks. The catalog provides a range of options that meet the needs of data scientists, developers, and researchers with various levels of AI expertise.
​
NGC containers can be run on-prem, in the cloud, and edge or using hybrid and multi-cloud deployments.
Lower Your Costs With Total Control & Flexibility​
​
Denvr's cloud computing services are 40% to 80% lower cost than other HPCaaS providers, with the industry-leading computing performance.
​
Drastically reduce your costs and gain the full benefits and the flexibility of cloud computing. With Denvr, you no longer have to pay the high costs of on-demand computing when your workloads increase.