nodes/iGavroche-rocm-ninodes

Files

T

Nino cc646ed4b4 Add quick installation guide

- Simple 3-step installation process
- Clear verification steps
- Expected performance improvements
- Quick start guide for users

2025-10-03 09:37:36 -04:00

1.4 KiB

Raw Permalink Blame History

Quick Installation Guide

🚀 Easy Installation (3 steps)

Step 1: Clone the Repository

cd ComfyUI/custom_nodes
git clone https://github.com/nino/rocm-nino.git ComfyUI-ROCM-Optimized-VAE

Step 2: Install the Plugin

cd ComfyUI-ROCM-Optimized-VAE
python install.py

Step 3: Restart ComfyUI

Close ComfyUI completely
Start ComfyUI again
Look for "ROCM VAE Decode" and "ROCM KSampler" nodes

✅ Verification

Check Node Categories: Look for "latent/rocm_optimized" and "sampling/rocm_optimized" in the node menu
Test Performance: Use the Performance Monitor nodes to analyze your setup
Load Example: Try the included example_workflow.json

🎯 Quick Start

Replace your standard "VAE Decode" with "ROCM VAE Decode"
Replace your standard "KSampler" with "ROCM KSampler"
Use default settings (optimized for gfx1151)
Enable all ROCm optimizations

📊 Expected Results

20-40% faster end-to-end generation
25-35% better VRAM usage
Reduced OOM errors with better memory management
Maintained or improved output quality

🆘 Need Help?

Issues: GitHub Issues
Documentation: Full README
Performance: Use the Performance Monitor nodes for recommendations

RocM-Nino v1.0.0 - ROCM Optimized Nodes for ComfyUI