Skip to main content
Ctrl+K
AMD Logo
ROCm™ AI Ecosystem
  • Core SDK
  • AI Ecosystem
  • GPU Systems and Infrastructure
  • Toolkits
    • AMD Data Science
    • AMD Finance
    • AMD Life Science
    • AMD LLM Extension
    • AMD Simulation
  • Blogs
  • Developer Hub

ROCm AI Ecosystem

Infera 0.1

Getting started

  • Overview
  • Quick start
  • Installation and setup
  • Glossary

Components

  • Server and router
  • Engines
  • Operator
  • KV cache daemon

Features

  • Feature matrix
  • Compatibility matrix
  • Prefill-decode disaggregation
  • KV cache offload
  • KV-aware routing
  • Routing and transport

Serving

  • Deployment
  • Kubernetes

Examples

  • SGLang 1P2D — Kimi-K2.6
  • SGLang mixed — DeepSeek-V4-Pro
  • SGLang 1P1D — DeepSeek-V4-Pro

Reference

  • CLI reference
  • Environment variables
  • Benchmarking
  • Troubleshooting
  • Repository
  • Open issue

Index

Symbols | A | C | E | F | H | I | K | L | M | N | P | R | S | W | X

Symbols

  • 1P1D / 2P1D / 1P2D

A

  • AINIC

C

  • Connector / Transport

E

  • etcd

F

  • fp8 / mxfp4

H

  • hipFile / AIS

I

  • Infera

K

  • KV cache
  • kvd

L

  • L1 / L2 / L3 / L4

M

  • MI355X
  • MoRI / MoRIIO

N

  • NATS

P

  • PD / disaggregation
  • Prefill / Decode
  • Prefix

R

  • RCCL
  • RDMA / RoCEv2
  • Retention
  • ROCm
  • Router
  • Routing policy

S

  • Server
  • Shared arena (CopyFree)

W

  • Worker / Engine

X

  • xGMI

  • Terms and Conditions
  • ROCm Licenses and Disclaimers
  • Privacy
  • Trademarks
  • Supply Chain Transparency
  • Fair and Open Competition
  • UK Tax Strategy
  • Cookie Policy
  • Cookie Settings
© 2026 Advanced Micro Devices, Inc