We made deals with various providers worldwide, all having different SLAs and specs. Below are examples of offerings, based on Docker or Bare Metal.

Shared GPU instance (Docker)

There are various configurations. Below is the default.

• One docker-instance on shared host
• One dedicated AMD GPU, shared is possible.
• 100 mbit internet
• 5GB space
• Self-serviced

The shared GPU option is perfect for startups.

As the service is docker-based, it can easily scale up to hundreds of GPU-nodes.

Scalable GPU service

Perfect for GPU-only tasks and need to handle peaks.

• No CPU, AMD GPUs.
• Web-API based.
• Scales up to 2 PFLOPS
• Various SLAs available for diverse needs.

The web-API is easy to understand and your current code can easily be ported.

Six NVidia GTX GPUs

• High-end CPU
• 6 NVidia GPUs
• Hosted in Europe
• Various SLA options

High-end GTX GPUs for large-scale operations in the cloud.

For more details, get in contact with Stream HPC.