AI Blogs - Page 2#
Bring Claude Code On‑Prem with AMD Instinct GPUs
Start running Claude Code securely with a self-hosted SGLang LLM on AMD Instinct MI355X GPUs.
Production-Ready MXFP4 Online Rotation with Fused Kernels on AMD Instinct™ MI355X
Learn how fused Gluon (Triton) kernels cut MXFP4 online-rotation overhead to near-zero on AMD Instinct MI355X, making it production-ready.
Quark Support for HuggingFace Diffusers and SVDQuant
Learn how to quantize, save, and reload diffusion models in Quark using its new SVDQuant and HuggingFace Diffusers support.
AUP Learning Cloud: Streamlining AI Education on AMD
An all-in-one ROCm JupyterHub platform that deploys GPU-ready AI teaching environments on AMD hardware with one installer and open-source teaching labs.
VSA: Accelerating Video Diffusion Inference with Sparse Attention on AMD GPUs
Accelerate video diffusion inference with VSA sparse attention: up to 3.31x attention kernel-time speedup on AMD Instinct MI308X GPUs
Reverse-Engineering hipBLASLt TensileLite Kernels: From Solution Name to a Tuning Config
Pin the pool's best kernel into a TensileLite tuning config by decoding its solution name, so an expanded re-tune can only match or beat it.
Introducing AMD CDNA™ 5 and the AMD Helios™ Rackscale Solution
Introducing AMD CDNA 5, the AMD Instinct MI455X GPU, and the AMD Helios rackscale solution: an open, integrated platform for rack-scale AI.
Closing the GPU Cluster Validation Gap: A Kubernetes-Native Approach with CVF
Learn how to validate AMD GPU clusters end-to-end with CVF: hardware acceptance, mesh bandwidth, RDMA, and RCCL testing in one pipeline.
AMD GPU Operator v1.5.0: DRA Support, Automated GPU Node Recovery, and Expanded Kubernetes Infrastructure Control
Discover how AMD GPU Operator v1.5.0 improves GPU scheduling, automates node recovery, and expands Kubernetes control.
Enabling Language-specific Reasoning in Multilingual Models with Reinforcement Learning
Learn how to train multilingual reasoning models with reinforcement learning and extend context windows on AMD Instinct GPUs.
Introducing Instella-MoE: A State-of-the-Art Fully Open Mixture-of-Experts Language Model
Explore Instella-MoE-16B-A3B, AMD’s fully open 16B MoE LLM with 2.8B active params per token, trained on AMD Instinct™ MI300 & MI325 GPUs.
Serve Kimi-K2.5-MXFP4 on MI355X with ATOM
Serve Kimi-K2.5-MXFP4 on MI355X with ATOM and gfx950 block-scaled FP4 kernels for optimized LLM inference.