AIPG: GPU-as-a-Service with Liqid

The AI Proving Ground GPU-as-a-Service environment is a fully automated solution that enables WWT engineers to build physical server environments with different server, CPU, GPU, and Operating System options.  Dedicated specialized builds happen within minutes without the need to physically touch the servers thanks to our Liqid Composable Disaggregated Infrastructure solution and RackN Digital Rebar Platform. Customers can choose the following options for their specialized, dedicated server.

• Server Partners
  • Dell
  • HPE
• CPU (Central Processing Unit)
  • Intel
  • AMD
• GPU (Graphical Processing Unit)
  • NVIDIA – A100, A30, L40
  • Intel - Flex 140, Flex 170, Max 1100
  • AMD - MI210
• Operating Systems
  • RHEL 8, RHEL 9, and Ubuntu 22.04

Customers can quickly leverage a server or servers with different configurations without needing to fully build and integrate different accelerators into the servers by hand just to evaluate different AI Models from a training or inferencing standpoint.

The AI Proving Ground GPU-as-a-Service lab is designed to revolutionize the way customers approach the sizing, configuration, and deployment of AI architectures within their data centers. By offering a fully automated, customizable server environment setup, this lab aims to:

Simplify the Evaluation Process: Enable customers to effortlessly test various AI models for training and inferencing by providing quick access to servers equipped with a wide range of CPUs, GPUs, and operating systems. This streamlined process removes the complexities of manual hardware integration.

Foster Innovation through Flexibility: With a diverse array of server partners (Dell and HPE), CPUs (Intel and AMD), GPUs (NVIDIA, Intel, and AMD), and operating systems (RHEL 8, RHEL 9, and Ubuntu 22.04), participants have the freedom to experiment with multiple configurations. This flexibility supports the discovery of optimal setups for specific AI use cases.

Accelerate Deployment with Automation: Utilize the cutting-edge Liqid Composable Disaggregated Infrastructure solution and RackN Digital Rebar Platform to automatically build physical server environments within minutes. This capability drastically reduces setup times and eliminates the need for physical interaction with the servers.

Enhance Understanding of AI Infrastructure: Provide customers with hands-on experience in configuring and deploying AI environments. This practical knowledge helps in making informed decisions about the infrastructure required to support their AI projects, from concept through to deployment.

Drive AI Projects Forward: By offering a quick and efficient way to test different hardware and software combinations, the lab supports the rapid development and scaling of AI applications. Customers can focus on innovation and achieving their project goals, rather than being bogged down by the logistics of infrastructure setup.

This GPU-as-a-Service environment is more than just a lab; it's a catalyst for AI advancement, enabling customers to push the boundaries of what's possible in AI deployment and application.

• Orchestration Platform: Liqid Composable Disaggregated Infrastructure with RackN Digital Rebar Integration.
• Servers/CPU's: Dell R660s, HPE DL-325s, Intel CPUs, AMD CPUs
• GPU's: Nvidia A100s, A30s and L40s | Intel Flex 140s, Flex 170s and Max 100s | AMD MI210s
• Operating Systems: RHEL 8, RHEL 9 and Ubuntu 22.04