Software tools & optimizations#
Discover the latest blogs about ROCm software tools, libraries, and performance optimizations to help you get the most out of your AMD hardware.
Two Waves, One Stage Apart: Warp-Pipelining Triton GEMMs on AMD Instinct
Use warp-pipelining to overlap matrix work and data preparation in compute-bound Triton GEMMs on AMD Instinct GPUs.
Group-GEMM in XLA: Making Mixture-of-Experts Fast and Memory-Efficient
Learn how to make JAX/MaxText MoE training more efficient with two new OpenXLA Group-GEMM backends for ragged-dot: ROCm hipBLASLt and generic Triton.
FP8 GEMM Optimization on AMD RDNA™ 4 Architecture
Learn how to port an FP16 GEMM kernel to FP8 on AMD RDNA 4 and use the smaller element size to unlock larger tiles and deeper K slices.
Scaling 3D Gaussian Splatting with Distributed Compute: Why One GPU Is Never Enough
Learn how to scale 3D Gaussian splatting across GPUs with distributed training, RCCL, and real-world optimization tips for AMD Instinct GPUs
MoE Training Optimization with Primus
Learn how to speed up MoE training on AMD Instinct GPUs with Primus: a fused MoE megakernel, FP8/MXFP8 expert GEMMs, and training-at-scale tuning.
hipCIM 26.06: GPU-Accelerated Whole-Slide Image I/O on AMD Instinct™ GPUs
In this blog you will see how hipCIM 26.06 reads whole-slide images on AMD Instinct™ GPUs, and how fast and faithfully it decodes them.
Bringing Ultra-Fast, Interactive LLM Inference to AMD Instinct™ MI355X GPUs
Accelerate single-request LLM inference on AMD Instinct MI355X with megakernels, speculative decoding, and MXFP4 for 1.9-3.3x faster decode.
FP16 GEMM Optimization on AMD RDNA™ 3 and RDNA™ 4 Architectures
Learn how to build an FP16 GEMM kernel with WMMA on AMD RDNA 3 and RDNA 4, one optimization at a time, and approach hipBLASLt performance.
AMD Instinct MI455X vs MI355X: A Technical Look at the Advancing AI 2026 Inference Numbers
Explore the MI455X vs MI355X inference numbers from Advancing AI 2026, and learn how each benchmark was measured and what it really means.
DI Series: Serving Kimi-K3 MXFP4 Wide-EP Disaggregated on AMD Instinct MI300X and MI325X
Learn how to serve a 1.4 TiB Kimi-K3 MXFP4 mixture-of-experts model disaggregated across AMD Instinct MI300X and MI325X GPUs.
Automating Performance Bottleneck Identification with the TraceLens Agent
TraceLens Agent is an agentic workflow that examines GPU traces to deliver a prioritized optimization report, detailing root causes and fixes.
UltraQuant on AMD Instinct: More Efficient Agentic Serving for Qwen3.8-MXFP4
A native 4-bit MXFP4 KV cache that speeds up Qwen3.8 decode over 8-bit KV on AMD Instinct MI355X, with no accuracy loss.