What is Nvidia DLDSR? And how to enable it
DSDR, FSR, RSR, DSR, AFMF2 how many more GPU initialisms do we need?
Share this article...
WePC is reader-supported. When you buy through links on our site, we may earn an affiliate commission. Prices subject to change. Learn more
DLDSR, or Deep Learning Dynamic Super Resolution, is Nvidia’s AI-powered Dynamic Super Resolution – an upgrade from DSR that uses spare GPU cycles to improve upscaling quality. This technology can be used in tandem with DLSS 3 to generate the most FPS possible, great for gamers struggling to play more-demanding titles.
The technology was introduced back in 2022 as part of the brand’s Game-ready driver update, providing an instant performance boost directly from driver-level implementation. But what is DLDSR? And how do you enable it in games?
Nvidia DLDSR – everything you need to know
NVIDIA developed the DLDSR technology as a method of Supersampling that aims to improve on the older DSR (Dynamic Super Resolution) technology. Thanks to its driver-level implementation, DLDSR should work with most games out of the box and doesn’t require developer support like DLSS.
DLDSR runs similarly to post-processing injectors at a driver level. It’s also said to increase FPS more than the standard DSR feature, as it makes use of a more advanced, artificial intelligence base algorithm.
DLDSR is only available on ADA GPUs, meaning you’ll have to own one of the latest RTX 4000 series cards to unlock this feature.
How does DLDSR work?
DLDSR uses AI to handle the upscaling and improve efficiency and performance over the standard DSR feature. Technically speaking, the two work similarly, DLDSR is just backed by AI. The feature works by taking an image and rendering it higher than your monitor’s resolution, then downscaling it back to the resolution you need.
The AI in DLDSR requires less data than the standard DSR to get the job done, which facilitates better performance because the GPU needs to spend less time rendering. Typically, DLDSR‘s 2.25X quality is comparable to DSR’s 4X, this means that where DSR needs to be at 4k to produce the same quality, DLDSR only needs 1620p.
On release, DLDSR was only available on the RTX 40 series GPUs, which had to do with Tensor cores, but compatibility has since been given to the 20 and 30 series GPUs too. These Tensor Cores are active when you enable any workload that requires AI. As of the time of writing, the 20, 30, and 40 series GPUs are the only Nvidia GPU series’ with Tensor cores.
When to use it?
This feature is best used when your GPU isn’t under too much stress. The more “spare” power your GPU has, the more the AI can dedicate to DLDSR. This also means that the better and more powerful your GPU is, the more spare rendering power you are likely to have available – yet another reason to upgrade. You can achieve this effect on older AAA titles, or playing E-sports titles, just be wary of the increased latency. You may also cap your framerates to limit your GPU utilization and create the same effect.
How to enable DLDSR
Step
Update drivers
Firstly, you want to make sure you have updated your graphics drivers to the latest version, or at least 511.23, when DLDSR was released. You are likely to get better performance with more recent driver updates.
Step
Open Nvidia control panel
Once the driver is updated, head over to the Nvidia control panel. You can either right-click the Nvidia icon in the tray or on the desktop and find it in the sub-menu.
Step
Enable DLDSR
In the control panel on the left-hand side under the 3D settings tab, select “Manage 3D settings”. From there, go under global settings to find the option for DSR – Factors and you can choose the option for DL scaling. You want to set this to 2.25X for the best balance between performance and quality.
GPUs that work best with DLDSR
We have compiled a short list of the best GPUs on which to run DLDSR. The better GPU you have, the better the performance will be across the board, so we’ve focused on the current generation of Nvidia GPU.
-
Best GPU for DLDSR
ASUS ROG Strix RTX 4090 OC
- Core Clock Speed: 2,640 MHz boost
- CUDA Cores: 16,384
- Memory Size: 24GB GDDR6X
- Dimensions: 357.6 x 149.3 x 70.1mm
- PSU Required: 1000W
- TDP: 450W
-
Runner UP
GIGABYTE GeForce RTX 4080 Eagle OC
- Core clock speed: 2205 MHz base, 2520 MHz boost
- CUDA count: 9,728
- Memory: 16GB GDDR6X
- Dimensions: 342 x 150 x 70 mm
- PSU required: 850W
- TBP: 320W
-
Budget pick
ASUS ROG Strix RTX 4060 OC
- Core clock speed: 1,830MHz base, 2,670MHz boost
- CUDA cores: 3,072
- Memory size: 8GB GDDR6
- Dimensions: 311.4 x 133.5 x 61.9mm
- PSU required: 550W
- TBP: 115W
Can you use DLDSR and DLSS together?
Yes, it’s encouraged if you want to get the best performance possible out of your GPU. It’s unlikely you’d need both on the 40 series GPUs, but it could help move things along if you’re still rocking one of the 20 series cards. If you don’t have an RTX card, then you can still use the standard DSR.
How is DLDSR different from DLSS?
As we mentioned before, DLDSR works at a driver level and does not need integration into games, because of this, it will work on almost any game. It also needs to be enabled in the GPU software itself, you won’t find a setting in your game for DLDSR.
DLSS, on the other hand, needs sewing into the game itself, as it works at the game engine level. DLSS is also based on presets, such as Quality, Balanced, and Performance, but DLDSR is not.
Both of these features produce different results but on the surface work similarly. Both of these options can be enabled and used together for some interesting results, though it is known to be a little glitchy at times.
