I was recently given a piece of advice by an AI !!

When I was complaining to an AI about how I couldn't make AI videos on my own PC—whining about how my graphics card didn't have enough memory or how a 32GB VRAM GPU costs a fortune—the AI hit me with a straight punch.

The AI, which usually acts like a sycophant, told me a harsh truth for once.

To put the AI's argument simply,

“Before you ask for things you don't have, your PC's CPU is incredibly old, isn't it? It's dragging down every single process, so you should solve that first!!”,

and just like that, I was confronted with a harsh reality.

Come to think of it, the PC I'm using now was built several years ago, back when AI drawing and AI video creation didn't even exist, so I had built it with power efficiency in mind.

So, I steeled myself and, while trying to manage my limited budget, decided to replace the PC's CPU, motherboard, and power supply.

Admittedly, the CPU and motherboard are also used, but they should be far faster than what I'm currently using.

I actually asked the AI specifically how much faster it would be.

My current CPU is a CORE i5 9400F. The one I'm installing is a CORE i5 13400F.

The CORE i5 9400F was released in 2019.
The CORE i5 13400F was released in 2023.

It's a transition from the Core i5-9400F to the Core i5-13400F. To conclude, you will experience a dramatic evolution (a massive improvement) in both feel and performance for AI image generation (like Stable Diffusion) and AI video creation.

While the graphics card (GPU) is the star of the show for AI processing itself, the reality is that whether the GPU can perform at its best and how quickly the pre- and post-generation processing is completed depends entirely on the CPU's performance.

I will explain in detail how much it will improve, step by step.

1. A jump in overall specs

First, let's look at how much the basic performance as a CPU changes.

The 9400F had 6 cores of equal strength, but the 13400F uses a hybrid design with high-performance P-cores (6) and power-efficient E-cores (4) that excel at background tasks, resulting in a massive increase in overall power.

2. Specific improvement points in AI image generation

When generating images with Stable Diffusion or similar tools, you will benefit in the following areas:

• The initial response after pressing the 'Generate' button (prompt parsing and model loading) becomes lightning fast. Thanks to the 13400F's large cache and high memory bandwidth, the 'preparation time' before sending data to the GPU is reduced from several seconds to an instant.

• Stabilization and acceleration of high-resolution processing (Hires.fix and upscaling) Some parts of the process for beautifully enlarging AI images (Upscaler) and the steps for noise processing on generated image data make heavy use of CPU multi-thread performance. Since this is processed all at once with the power of 16 threads, the waiting time for high-resolution processing is significantly reduced.

• Comfort in image management and metadata loading Operations such as previewing a large number of generated images in folders or loading past generation logs (PNG Info) will become snappy and smooth without any stuttering.

3. Dramatic improvements in AI video creation (this is the biggest benefit)

In AI video creation (including not just generative AI, but also AI features in video editing software), you can feel the evolution of the CPU most strongly.

• Control of video generation AI (local environment) AI video is a super heavy-duty process that involves "generating and connecting dozens of images per second." With the 9400F (6 threads), the CPU usage would stick at 100% just by continuing to send instructions to the GPU, making the entire PC feel frozen and sluggish. With the 13400F (16 threads), there is enough headroom to let the GPU work to its limit while still being able to do other tasks in the background.

• AI features in video editing software (Premiere Pro, DaVinci Resolve, etc.) Features recently added to video editing software such as "AI automatic transcription," "AI subject cutout (smart mask)," and "AI noise removal" utilize not only the GPU but also the full power of the CPU. By switching to the 13400F, the time required for these AI analyses is reduced to less than half.

• Acceleration of final video export (encoding) When outputting the completed AI video, the multi-core performance of the CPU has a direct impact. Export time is nearly 2 to 3 times faster compared to the 9400F era.

💡 Checkpoints for achieving maximum effect

The Core i5-13400F is an excellent mid-range CPU, but to make it show its true value, it is perfect if you just check the following two points.

1. 32GB or more of memory (RAM) is recommended AI images and videos consume a lot of memory. Even if the CPU is new, if the memory is 16GB, that will become a bottleneck, so it is best to operate in a 32GB (DDR4-3200 or DDR5) environment if possible.

2. Balance with the graphics card (GPU) If the GPU (GeForce, etc.) you are currently using is from that time, switching to the 13400F will completely eliminate the "bottleneck state where the CPU holds back the GPU and prevents it from reaching its full performance." Even if you replace the GPU with a model with more VRAM (such as an RTX 3060 12GB or RTX 40 series) in the future, the 13400F will be able to handle that power sufficiently.

The "7 generations of evolution" from 2019 to 2026 is tremendous. Please enjoy how every task, from OS behavior to AI processing and exporting, becomes a whole new level of lightness!

That is what they say.