ROCm Inference Optimizer
8.2
A software suite leveraging AMD's ATOM and ATOMesh technologies to intelligently distribute and optimize LLM inference workloads across multiple GPUs, maximizing throughput and reducing latency, particularly for prefill and decode stages.
250h
mvp estimate
8.2
viability grade
8
views
technology stack
C#
PostgreSQL
Difficult
inspired by
AMD ships ATOM + ATOMesh for LLM inference optimization