On the surface it seems like a very strange assertion.
Even if there were subtle GDI rendering differences -- which I doubt -- it is hard to believe that a printout would let them positively identify Intel HD Graphics 630 specifically, as opposed to say HD Graphics 610 or 615 which are slightly slower clocked versions of the same GPU released at the same time and which almost definitely use the same drivers and GDI rendering system.
But if the information comes from elsewhere, as in already having a suspect and knowing what computer they used -- they should reasonably have given more information from the same source -- CPU model, etc.
I can't imagine a possible way to leak only GPU info -- unless GDI or the Intel drivers has some secret mechanism for intentionally rendering some sort of subtle identification code onto printer output.
You’re assuming extreme precision and competence. I’m guessing they have an incomplete database of GPU samples and a fuzzy result came back like “86% match to Intel 630,” and that’s what they published.
Even if there were subtle GDI rendering differences -- which I doubt -- it is hard to believe that a printout would let them positively identify Intel HD Graphics 630 specifically, as opposed to say HD Graphics 610 or 615 which are slightly slower clocked versions of the same GPU released at the same time and which almost definitely use the same drivers and GDI rendering system.
But if the information comes from elsewhere, as in already having a suspect and knowing what computer they used -- they should reasonably have given more information from the same source -- CPU model, etc.
I can't imagine a possible way to leak only GPU info -- unless GDI or the Intel drivers has some secret mechanism for intentionally rendering some sort of subtle identification code onto printer output.
It's just really hard to understand.