I have an RTX 2080 and a 750 Ti with 2GB of VRAM. I'm curious if using the 750 Ti just for display output would help save VRAM while I render on the 2080. My concern is that the desktop, browsers, and apps like Discord can consume almost 1GB of VRAM. I also want to explore lossless scaling, though I'm worried the 750 Ti may not handle it well.
Power-wise, I'm okay since the 750 Ti only uses about 60W, and my B350 motherboard can support a second GPU at 2.0x4 speed, leaving the primary GPU and NVMe at full performance. I'd like to know if it's worth the effort to set this up, considering I might need to uninstall the current AMD drivers and sort out placement for the 2080 since it doesn't have rear I/O. Can I just install the latest drivers for both GPUs despite the 4-year gap in release?
3 Answers
Honestly, I wouldn't recommend it. The driver compatibility issues could create a lot of headaches and problems that are probably worse than any VRAM savings you'd achieve. It's more trouble than it’s worth!
The only real benefit I see with this kind of setup is using the second GPU for lossless scaling or frame generation. I’m thinking about testing an older GPU alongside my 3060 Ti for that purpose, but when it comes to VRAM savings, that doesn’t seem likely at all.
Nope, without Crossfire or SLI, the setup won’t work the way you’re thinking. Only one GPU can handle the display and GUI processes at a time. The second GPU could potentially handle non-display tasks, but you're likely stuck with all GUI operations going through the main GPU that’s running the monitor.
So, everything is rendered on the GPU connected to the monitor, right? Unless you specifically set a game to use the second GPU, it would remain inactive without direct commands. I’ve done something like this before with an AMD and NVIDIA card combo and noticed the GPU plugged in always had higher usage.
Is it mainly because of the age gap in the GPUs? I once ran an AMD 6700 XT alongside a 2080 without major issues, with both drivers installed.