Stable Diffusion Requirements 2024 (Best CPU, GPU, & RAM)

While there are many online tools and websites to run Stable Diffusion, installing it locally on your computer is the best way to use it. 

That’s because you don’t have any limitations on Stable Diffusion when using it locally. You can create as many images as you want, use various models, generate NSFW artwork, etc. 

But this brings up an important question: 

What are the requirements to run Stable Diffusion? 

Stable Diffusion requires a modern Intel or AMD processor with at least 16GB of RAM, an Nvidia RTX 3060 GPU with atleast 6GB of VRAM, and atleast 10GB of storage space. 

These Stable Diffusion requirements pretty much lie in the middle and with these specifications, you’ll be able to run it comfortably. However, you can also run Stable Diffusion with low system specifications. 

In this article, I’ll go in-depth about the software and hardware requirements for Stable Diffusion and how to run it perfectly without running into memory errors. 

Stable Diffusion Requirements

There are many factors in play when it comes to running Stable Diffusion but the most important factor is the GPU. 

Stable Diffusion primarily uses the GPU to run which makes it the most important component on your computer. Yes, your processor and RAM matter as well but not as much as having a powerful GPU. 

In the table below, I’ve broken down the Stable Diffusion requirements into three categories: 

To Get The Job Done: The specifications here will be able to run Stable Diffusion but it won’t be a smooth experience as image generation won’t be very fast and you may run into memory errors often. 

For Running It Effectively: These specifications will be able to run Stable Diffusion effectively and image generation will be fast as well.  

For Optimal Performance: These are the specifications that will run Stable Diffusion optimally and image generation will be insanely fast. 

From the above table, you’ll get some idea of what you’d need to run Stable Diffusion. 

Recommended: How To Run Stable Diffusion On Android

Stable Diffusion Requirements: PC Hardware To Buy

You already know by now that a powerful GPU is the most important factor for running Stable Diffusion. 

But let’s talk about the other hardware requirements and what you need to consider if you’re setting up a new PC build for Stable Diffusion. 

CPU: Since Stable Diffusion uses GPU to run, you don’t need a very powerful CPU. However, a 6-core processor is a decent option to go for if you’re building a new PC for Stable Diffusion. 

GPU: When it comes to GPU, an Nvidia RTX graphic card is the best option to choose as RTX cards come with CUDA cores which help a lot. On top of this, leaning towards more VRAM should be a priority as the more VRAM your GPU has, the faster the image generation process will be. 

RAM: A 16GB RAM is sufficient for running Stable Diffusion especially if you plan on multi-tasking while generating images. 8GB can get the job done but you won’t be able to smoothly run your browser or other applications while Stable Diffusion is running. 

Storage: Stable Diffusion requires a minimum of 10GB of storage space but I recommend having more free space on your drive, preferably around 50-75GB. That’s because you’ll most probably use many other checkpoint and Lora models when generating images and these models can be 3-6GB in size. 

Thermals: I’d also like to mention that your PC or laptop could get very hot while running Stable Diffusion so make sure you have good thermals or cooling for your computer. If you’re using a laptop like I do, use a cooling pad to avoid overheating. 

SDXL Requirements

It’s also important to note that as further developments are made into Stable Diffusion, the system requirements for Stable Diffusion might increase even further. 

We’ve already seen this when SDXL was launched which is their latest and much refined model. 

With the launch of SDXL, many users who were able to run Stable Diffusion smoothly struggled to generate images faster using the new SDXL model. 

That’s because the new SDXL model is trained on 1024×1024 size images and is more detailed. On top of this, the model also requires a refiner model along with it. 

This increased image generation times considerably for users on lower-end GPUs essentially making people rethink the optimal Stable Diffusion requirements.

To run SDXL effectively, you’d need atleast 8GB of VRAM on your graphic card as image generation will be faster. 

4GB and 6GB of VRAM can do the job but it isn’t enough to quickly generate 1024×1024 images. 

For reference, I have an RTX 3050 GPU with 4GB VRAM on my laptop and it takes 3-5 minutes in my workflow to generate a 1024×1024 image and then upscale it to 4K using SDXL models. 

Previously, I was able to generate images in less than a minute or two using Stable Diffusion 1.5 models. 

Does Stable Diffusion Run On AMD GPUs?

Stable Diffusion can run on AMD GPUs but not officially. There are certain workarounds you’ll have to go through in order to run Stable Diffusion on an AMD graphic card. 

This Reddit thread mentions one of the workarounds you can try for running Stable Diffusion on AMD GPUs. 

However, I would recommend using an Nvidia GPU as there’s not only official support for it to run Stable Diffusion but you’ll also get far better performance with Nvidia cards. 

Does Stable Diffusion Run On Mac? 

You can run Stable Diffusion on Mac devices such as Macbooks or iMac. There are certain forks of Stable Diffusion such as InvokeAI which run well on Macs. 

DiffusionBee is another software that lets you run Stable Diffusion on your Mac with a very easy and straightforward process. You can check out our detailed guide which shares multiple methods of running Stable Diffusion on Mac.

However, the Stable Diffusion requirements for Mac are completely different.

When it comes to running Stable Diffusion on Mac, you’d need more RAM instead of GPU memory especially if you’re using an M2 Mac. 

That’s because M2 devices have a unified memory and due to the shared nature of the memory, even a 16GB RAM would not be sufficient enough for running Stable Diffusion. 

You’d need at least 32GB RAM or 64GB ideally in order to properly run Stable Diffusion without any lag or slow speed. 

Recommended: How To Speed Up Stable Diffusion

Can You Run Stable Diffusion Without A GPU? 

You can run Stable Diffusion without a GPU but it’s counterproductive and not the best way as it’ll be slow and not very efficient.

Moreover, you’d need to have a much more powerful CPU and more RAM for running Stable Diffusion using your CPU. 

Also, you’d need a specific fork of Stable Diffusion that is designed to use the CPU instead of the GPU for running. One such fork of Stable Diffusion is OpenVino which is a good option and it works with both Intel and AMD CPUs. 

Recommend: ComfyUI Guide: A Stable Diffusion Interface That’s Insanely Fast

Conclusion 

Running Stable Diffusion requires a well-equipped machine as evident from the requirements shared in this article. 

If your current hardware meets these requirements, you can start using Stable Diffusion right away but if it doesn’t, you’d need to either upgrade your hardware or get a new computer or laptop. 

Hopefully, this article will help you understand more about running Stable Diffusion properly and will give you some clarity on the hardware you’d need to purchase if you plan on a new build. 

If you have any further questions regarding Stable Diffusion requirements, feel free to drop them in the comments below.