…sparse attention, MoE execution, quantization, multi-token prediction, scheduling, graph capture, and distributed inference. In less than a month, AMD Instinct™ MI355X GPU performance improved by more than 100x on the same…
…W4A8 & W8A8 Quantization with AMD Quark — ROCm Blogs Quantize Kimi-K2.5 to W4A8 and W8A8 using AMD Quark and serve on MI325X with FlyDSL and AITER for further inference acceleration. May…
…May 26, 2026 QuickReduce FP4 Quantization and Benchmarking on MI355 — ROCm Blogs Learn how QuickReduce uses FP4 quantization to accelerate all-reduce communication and evaluate its performance on AMD Instinct MI355 GPUs…
…Benchmarking AI Coding Agents for GPU Kernel Optimization on AMD Instinct GPUs — ROCm Blogs Explore how AI coding agents compare on real GPU kernel optimization with AgentKernelArena, AMD's open benchmarking arena…
…We benchmarked twelve models across BF16 and Q8_0, and the prompt-processing numbers are not subtle. The short version (AMD EPYC™ 9755, 128 cores, single socket): Up to 4.5x the…
…debugging or deployment and orchestration on AMD Instinct GPUs. Performance Results Explore the latest performance benchmarks on AMD GPUs. AMD Infinity Hub Get advanced GPU software containers and deployment guides. Open-Source…
…We have two options to reproduce the benchmark results. While the Model Automation and Dashboarding (MAD) method automatically gathers comprehensive performance metrics under preconfigured benchmark settings, standalone benchmarking allows end users can…
…May 20, 2026 QuickReduce FP4 Quantization and Benchmarking on MI355 — ROCm Blogs Learn how QuickReduce uses FP4 quantization to accelerate all-reduce communication and evaluate its performance on AMD Instinct MI355 GPUs…
…July 12, 2026 QuickReduce INT3 Quantization and Benchmarking on MI355 — ROCm Blogs Learn how QuickReduce uses INT3 quantization to accelerate all-reduce communication and evaluate its performance and accuracy on AMD Instinct…
…EDSR Benchmark dataset . Each model card includes inference scripts, pre-compiled model caches, evaluation tools, and benchmark results. After installing the Ryzen AI SW and activating the conda environment, getting started takes…