The first Nvidia GeForce RTX 50 Series GPU is now available. CG channel hardware expert Jason Lewis sets initial benchmark test results for GeForce RTX 5090 and 5080 in CG apps including 3DS Max, Blender, Maya and Unreal Engine.
Last week we saw the launch of the GeForce RTX 5090 and GeForce RTX 5080, the first cards in the GeForce RTX 50 series, NVIDIA’s new consumer GPUs based on the Blackwell architecture.
Since then, there have been quite a few online reviews of new GPUs, and most of the time they focus on gaming and some on AI workloads. But how do they work in DCC and graphics workflows?
Well, I also started testing the GeForce RTX 5090 and 5080 using a few popular DCC applications.
Technical specifications
We tested new GPUs against comparable cards from the ADA Generation GeForce RTX 40 Series GPUs, GeForce RTX 4090, and GeForce RTX 4080 Super.
Here’s how the GeForce RTX 5090 and GeForce RTX 5080 stack on paper against previous generation counterparts:
Nvidia’s GeForce RTX 50 Series GPU and previous generation counterpart RTX 5090 RTX 4090 RTX 5080 RTX 4080
Super Architecture Blackwell Adarablace Blackwell Adarablace Couda Couda Core 21,760 16,384 10,752 10.240 Tensor Core* 680 512 336 320
RT Core*170 128 84 80 Base Clock (GHz) 2.01 2.23 2.30 2.29 Boost Clock (GHz) 2.41 2.52 2.62 2.55 Computing Performance
FP32 (TFLOPS)* 104.8 82.6 56.3 52.2 GPU memory
32 GB GDDR7
24GB
gddr6x
16 GB GDDR7
16 GB
GDDR6X TGP 575W 450W 360W 320W Release Date 2025 2022 2025 2024 Launch $1,999 $1,599 $999 $999 $999
*Data obtained from third-party websites.
Testing Process
Early testing used the same test rig and benchmark scene as the latest group tests for GeForce RTX 40 series GPUs, and was updated to the current version of the software.
Before you reach the actual test results, I would like to repeat what many of you already know. Hardware development almost always outweighs software development.
Professional graphics applications evolve very slowly compared to gaming and other consumer applications, and it takes time for CG software to reach the level of optimization that allows for real use of new GPU architectures.
Benchmarking and Performance for DCC Applications
As you can see, the initial test results are everywhere. In some tests, the GeForce RTX 5090 and 5080 are significantly faster than the counterparts of previous generations, but slightly faster in some and actually slower in some.
So, what does this mean? Most other early reviewers have concluded that both the NVIDIA GPU drivers and the software used require an update before assessing the true performance of the new GPU, and we have to take the same boat I think I’m riding it.
I will do more extensive testing of these new GPUs in the coming weeks and then provide a comprehensive review. By then I would like to take a more clear picture.
Early availability
One of the last things I want to talk about is the current availability of GeForce RTX 5090 and 5080.
Many reviewers, including myself, believe that last week it will be a GPU paper release. This is because the retail channel is currently barely available.
Some retailers estimate that it could take months before supply catches up to demand, and while gamers are naturally upset about this, it’s not that problematic for CG artists.
By the time new GPUs are fully delivered, the drivers should be mature, and graphics software should be able to take advantage of that potential better.
Many DCC software developers will release major new versions of the software during Spring, so it will be interesting to see how these new versions work on Blackwell GPUs.
Find the specifications for GeForce RTX 50 Series GPUs on Nvidia’s website
Follow our CG channels on Facebook, Instagram and X (formerly Twitter) to speak up about this story. In addition to being able to comment on the story, followers of social media accounts can also watch videos that don’t post to the site itself, such as creating the latest VFX movies, animations, gaming movies, and motion graphics projects.
