NVIDIA AIBs Highlight Strong Stable Diffusion Capabilities of GeForce RTX 4090 To Entice Customers, Takes Mere Seconds To Draw AI Images

NVIDIA's high-end GPUs such as GeForce RTX 4090 are now being marketed for features besides gaming such as Stable Diffusion AI image generation.

Stable Diffusion Might Be The Latest Reason To Tempt Customers Into Buying High-End NVIDIA  RTX 40 GPUs, 4090 Takes 30 Seconds To Generate AI Images

Ever since the popularization of AI-generated content, not just companies but consumers have also seen a rise in demand for gaming GPUs, more particularly, NVIDIA's recent RTX 40 options. As such, a range of AIBs are now taking advantage of this hype by marketing the stable diffusion capabilities of their latest gaming GPUs.

AIBs in the Asia Pacific market region, especially in China, are betting heavily on Stable Diffusion AI content generation. A media press release by Colorful China, published by MyDrivers, shows how a clever marketing campaign has been laid out that shows potential buyers the AI generation power of the GeForce RTX 4090 GPU.

It is said that the card can take mere seconds to output high-res AI images. It is also stated that almost 99% of the most popular Stable Diffusion applications & workloads are optimized around NVIDIA's CUDA cores. The NVIDIA RTX 4090 packs over 16 thousand of these cores along with a tensor core for faster AI operations. Another recommendation is given in terms of VRAM as having cards with more than 10 GB is a must otherwise the operation will stop working. That's where that 24 GB of VRAM on the RTX 4090 comes in. Even the RTX 4080 and 4070 Ti pack more than 10 GB of video memory.

Here we use the highly popular model "Counterfeit-V2.5" for image generation. After inputting the prompt, set the basic resolution to "800*500" to enable the face repair option, and then turn on the high-resolution repair option to directly The resolution is increased to 1600*1000, the number of sampling steps is 20, and the prompt word guide coefficient is adjusted to 10 to generate an image.

Colorful via MyDrivers

A more high-quality AI render can take around 34.5 seconds (30 seconds average) and that's on a single NVIDIA GeForce RTX 4090 graphics card. But that's not all, even Pro cards such as the NVIDIA RTX 6000 Ada which retails for $6800 US was recently tested by Taiwanese tech outlet, CoolPC, which also serves as a retailer.

NVIDIA GeForce RTX 4090 vs RTX 6000 Ada In Stable Diffusion AI GPU Performance (Image Credits: CoolPC TW):

2 of 9

In the review, one of the main focuses was on Stable Diffusion AI image generation and the RTX 4090 was able to generate 20 images in around 8 minutes while the RTX 6000 Ada took 9 minutes to compete for the same generation tasks. The 1-minute difference may not be a lot but the price difference of $5200 US will give many a reason to choose the GeForce RTX 4090 over the RTX 6000 Ada. With that said, the RTX 6000 Ada has its own advantages in the workstation and pro segment but if you are looking for a card with lots of Stable Diffusion AI prowess, then RTX 4090 seems to be the safest bet.

The reviewer did show that he was mostly interested in generating AI Waifus which also seems to be the case in Colorful's marketing as seen below:

Now the NVIDIA GeForce RTX 40 GPUs are first and foremost gaming products but just like we saw with crypto (which NVIDIA says provides nothing of benefit to society), Stable Diffusion AI and the hype surrounding it will definitely mark a rise in demand for the said graphics cards. I mean Stable Diffusion AI can be productive and there's a lot one can do besides just drawing their favorite anime waifu. There are lots of Stable Diffusion packages that are available at the moment such as this one which is designed for NVIDIA GPUs.

Written by Hassan Mujtaba