Dynamic GPU Memory with Run:ai

Raz demoing Fractional GPU with Dynamic GPU Memory Allocations!
Fractional GPU allows multiple workloads to share a single GPU. With Dynamic GPU Memory Allocations, Run:ai brings to Kubernetes the concept of GPU Memory Request and GPU Memory Limit, similar to CPU Memory Request and Limit.
Workloads can now share the GPU memory of a single GPU in a dynamic way by bursting above their average consumption.
Read more in our blog
www.run.ai/blog/dynamic-gpu-m...