In theory yes, in practice that almost never happens.

The number of teams I've seen pursue golden-image tests on GPU pipelines and fail has been 100%(4/4).

What's the source of the image differences?

I probably shouldn't find this disturbing, but I do.

It's a host of things, generally I've seen:

  - Poorly defined APIs/types that allow vendors to diverge.
  - Legit driver bugs.
  - Vendors preferring optimizations to following the spec(including skipping memzero requirements).
  - Hardware that doesn't behave how it should(fp32 falling back to fp16 silently for instance).

Overall you don't really win any points in the market for correctness, instead performance and efficiency are highly valued.

Can you provide some rendered screenshots with a few fonts and sizes? (Ideally including a side-by-side comparison with Freetype?)