Graphics functionality != display output What about laptop GPUs, which don't necessarily output to the screen at all times. Sometimes they don't even have a capability to do so. If it's coprocessor working alongside the general processor for the primary purpose of accelerating graphics computing workloads, it seems appropriate to call it a GPU.

Edit: perhaps your point is that it doesn't make sense to call a device designed primarily to accelerate ML workloads or just general purpose vector calculations. In that case I'd agree that GPU isn't the right name.

If it looks like a duck and it walks like a duck, why is it not a duck? If you are using a DSP to process graphics, then at least in the context of your system it has become your graphics processor.

Plenty of GPUs don't have (or aren't used for their) display output. It's a GPU because of what it does: graphics processing. Not because of what connectivity it has.

CPUs and DSPs are not primarily designed for graphics work, therefore they don't count as GPUs. CPU are general-purpose, DSPs might be abused for graphics work.

The "G" in GPU doesn't imply that they have to render directly to a screen. In fact, professional graphics cards are commonly used for bulk rendering for animating videos.

Datacenter GPUs are mostly used for AI these days, but they can nevertheless do graphics work very well, and if they are used for generative AI or if their built-in super sampling capability is used, the distinction becomes rather blurry.