Which GPU is recommended for Aximmetry?
To utilize all functionalities of Aximmetry we recommend using NVidia GPUs, the latest and most powerful of them currently being the RTX 4090.
Are all NVidia GPUs supported?
All modern NVidia cards are supported.
Regarding older NVidia cards, to use Aximmetry you need at least a card built on the Maxwell architecture (NVidia GTX 750 and above).
Can Aximmetry run on AMD GPU?
Yes, but AMD is only partially compatible with Aximmetry. The main reason for that is that to encode/decode H.264/.264 we use an NVidia GPU-based codec, which means you won't be able to record into MP4 or directly stream to social media.
Does Aximmetry benefit from Multi-GPU setups rendering the same image?
As Aximmetry is designed with a Single Strong GPU in mind to render a single frame, we suggest creating your scene based on the performance of the strongest GPU in your machine.
Can Aximmetry utilize Multi-GPU setups in any way?
Yes, Aximmetry SE can utilize multiple GPUs installed in a single machine. The user can choose between two modes:
- In Independent mode:
- Aximmetry can render multiple different outputs at the same time with multiple separate GPUs in either single-machine or multi-machine mode.
- This mode is used when we want to render two or more completely different contents. Each content is rendered on its own GPU thus distributing the load.
- In this mode, we can select a maximum of 4 GPUs. They don't have to be identical as each card renders independently from the other cards.
- You can route the render of GPU1 to be outputted via the output of GPU 2 for example, however, It’s highly advisable that each GPU renders the content of its own connected monitor to avoid unnecessary data traffic.
- In Linked mode:
- Aximmetry can share the rendering of individual frames between GPUs. (So every other Frame is rendered by another GPU.)
- This mode is used to maximize the rendering performance of our machine. It makes rendering a single output faster.
- It’s only usable when we direct outputs to I/O cards
- It works by rendering two or more consecutive frames at the same time parallel on the multiple GPUs thus multiplying the usable GPU performance per frame. The frames rendered this way are lined up and sent to the I/O card sequentially.
- The weakest card will determine the overall performance of the system.
What to take care of in case of a Multi-GPU setup?
- Aximmetry loads the data to both GPUs, so both GPUs have to have enough memory.
- Preview windows are always rendered with the primary GPU using its performance.
- In Linked mode, it is advised to use the same type* of GPU.
Same Type: For example, an RTX 40XX or 30XX from different manufacturers or even the same manufacturer will have different performance figures. So, even though they share the same GPU die \ chip, they have different cooling and clocks that will make them perform differently. This applies to all GPUs. So, make sure that the GPUs are from the same manufacturer and the same model.
Does Aximmetry DE with Unreal Engine benefit from a Multi-GPU setup?
At the moment not even Unreal Engine can fully utilize the capabilities of an SLI connection and Multi-GPU to render a single frame.
In case multiple Unreal Engine outputs are needed at the same time, then Aximmetry must be used with a Multi-Machine setup. In this case, each machine and its GPU are rendering an output separately. This way, one's scene can be rendered from different angles simultaneously.
More on the multi-machine setup here.
NVidia Quadro & Quadro Sync
Aximmetry works well with desktop GPUs. Using Quadro cards does not give any advantage over them. If you want to use NVidia Quadro cards, make sure to choose the equivalent of the recommended RTX cards.
However, if you would like to sync GPU outputs, you can use NVidia Quadro Sync, which only works with NVidia Quadro GPUs.
More on Quadro Sync here.