- NVIDIA SLI and AMD CrossFire are dead, but dual GPUs can still be useful for personal computing.
- We’ve asked the experts where does it still make sense to use two GPUs.
- Read on and find out what kind of tasks are suited for dual GPU use in non-enterprise situations.
Dual GPUs configured in SLI and later in CrossFire mode was the height of PC gaming 10-15 years ago.
That has changed quickly, with both manufacturers ditching support for this technology that used two identical GPUs to render half of the image each, the combined the output before displaying on your monitor.
CONTENTS
NVIDIA SLI and AMD CrossFire dual GPU modes for gaming are completely dead
I’m sure some people are decrying the demise of dual GPU gaming, but the truth is that only a fraction of the users had the cash required for such a high-end configuration.
Not only that, but the edge-cases resulted in a ton of bugs and compatibility problems that make it hard for users to enjoy the benefits of SLI and CrossFire.
The fact VRAM was duplicated between GPUs, you could only use identical models didn’t help with the mass adoption of this technology. Oh, and the performance was not 2x, as you may expect.
If you don’t want the read to the end the conclusion of this article is that yes, you can use multiple GPUs at the same time, and they don’t even have to be from the same manufacturer, but that doesn’t mean you will get the benefits of running both GPUs in the same computer.
Most apps can’t use more then one GPU at a time, so you could just run two graphic intensive apps at the same time, assuming you are assigning each GPU to the correct app.
So, when does it make sense to invest in a dual GPU configuration today? This is a question we will try to answer in this article. We have gathered ideas from industry experts to highlight the advantages of multiple graphics cards in the same system.
High-Performance Video Editing and 3D Rendering
One personal computing task that would benefit significantly from using dual GPUs is high-performance video editing and 3D rendering. Tasks such as rendering complex video effects, real-time editing, and running high-quality simulations can place immense demands on a single GPU. By utilizing dual GPUs, the workload can be split, speeding up rendering times and enhancing real-time performance, which is especially beneficial for content creators working on large-scale projects.
Dual GPUs would provide the needed power to handle multitasking in video editing, allowing smoother workflows when working with multiple high-resolution streams or complex visual effects. Additionally, for industries like gaming or simulation modeling, dual GPUs could ensure better frame rates and rendering speeds, providing a far more immersive and efficient computing experience. This capability would be particularly impactful for professionals in creative fields who need both precision and speed.
Shehar Yar, CEO, Software House
Advanced Data Analysis and Machine Learning
For me personally, one of the most practical tasks that benefits from dual GPUs is running advanced data analysis and machine learning models. I’ve noticed that when training large-scale models, like those used for predictive analytics or recommendation systems, a single GPU often gets bogged down, slowing the process dramatically. With dual GPUs, the workload can be split efficiently, which I believe reduces training time by up to 40%. This helps deliver insights faster, which is crucial in a fast-paced business environment.
Anders Bill, Cofounder/CPO, Superfiliate
High-Performance Data Visualization
In my opinion, dual GPUs would be a huge advantage for anyone working on high-performance data visualization, especially when rendering large, complex datasets. Personally, I think industries like finance, healthcare, or scientific research, which deal with massive amounts of data, could see significant improvements by using dual GPUs.
For instance, when visualizing real-time stock market trends or large-scale patient data, the sheer volume can overwhelm a single GPU. I’ve found that having a second GPU dedicated to handling the rendering while the primary one processes the core data can speed up analysis by up to 50%. This dual-processing capability allows for smoother, quicker visualization and immediate interaction, which can make a big difference when working with time-sensitive data.
Matthew Goulart, Founder, Ignite Digital
GPU passthrough virtualization for near bare metal performance
This one is from me. Be warned!
If you have used a virtual machine before you have seen the biggest performance difference compared to bare metal aka non virtualized hardware: any graphics intensive app will slow down to a crawl, or becomes completely unusable.
If you have two or more GPUs it’s possible to completely attach a GPU not in use by the supervisor machine to a virtual machine by using GPU passthrough virtualization. For the virtual machine it will look like the GPU is attached directly, and the performance will be very close to the native performance.
This could allow you to run Linux on your main machine, then maybe launch a Windows virtual machine just for gaming. That would work very well.
Did we miss any important dual GPU use case? Let’s continue the discussion in the comments.
