ComfyUI is an open-source, graphical node-based interface for building generative AI workflows using diffusion models. With over 100,000 stars on GitHub, it has become one of the most widely adopted tools for text-to-image, text-to-video, and image-to-3D generation. You can build workflows by connecting nodes in a drag-and-drop visual interface (no coding required). A large community contributes custom nodes and workflow templates, making ComfyUI a versatile front-end for models ranging from 12B to 27B parameters. For background and setup across AMD platforms, see the earlier ROCm blogs Running ComfyUI on AMD Instinct, Getting Started with ComfyUI on AMD Radeon™ RX 9000 Series GPUs, and Running ComfyUI in Windows with ROCm on WSL.
The models behind these workflows (Diffusion Transformers or DiTs, rectified flow transformers, and multi-view diffusion networks) are both compute- and memory-intensive, making GPU acceleration critical for practical use. This blog benchmarks three production-relevant ComfyUI workflows on AMD Instinct MI355X and NVIDIA B200 and shows that MI355X is up to 1.439x faster [1], enabled by PyTorch Attention optimizations for the CDNA4 (gfx950) architecture shipped with ROCm 7.2.0. You can follow along to set up your environment, run ComfyUI, and reproduce these results.