Hi, we are building an QHD OLED interface board. Unfortunately, the display does only a self-refresh from the onchip memory which can be written to from outside. The display provides a vertical sync signal, so we know the refresh phase. I assume it is impossible for a graphic adaptor to be fed by this vsync (where the display plays the master's role), so I was thinking if at least fine tuning of the video output is possible - e.g. I could work out an algorithm which will use variable vertical blank time, if we could specify it per frame. Is that possible? My target is to minimize the latency and avoid tearing - the ideal case is when the memory write happens right after the display's refresh (read pointer). Any thoughts? The application is VR (actually dual QHD screens per HMD), and judging by the latest turn of AMD to provide best solution for VR, I am curious if we can do something with that latency issue.
I did get an engineering team to take a look at your question. It looks like we're talking about a custom engineering solution, and that is beyond what we can do for questions in the forums.