Ray Core has come a long way since its first major release back in 2020. Today, we are investing in Ray Core scalability and reliability, while also evolving Ray for an accelerator-native world. Hear about some of the improvements we’ve made since last year’s summit, including streaming generators and weekly Ray releases. Then, we’ll discuss Ray Compiled Graphs, a new backend for Ray Core that makes it simple to build high-performance applications on clusters of accelerators. We’ll do a deep dive of the key Compiled Graphs features, including faster Ray task execution and native support for GPU-GPU communication via NCCL. We’ll also give a brief overview of some of the main use cases in multi-GPU LLM inference and distributed training.
Ray Summit 2024
In-Person Agenda
READY TO REGISTER?
Come connect with the global community of thinkers and disruptors who are building and deploying the next generation of AI and ML applications.
Join the Conversation
Hashtag it
#RaySummitDon't wait for the conference to get the convo going. Join the Ray community now – ask a question in the forums, open a pull request or simply share why you’re excited. Create some buzz!