What I can't remember is if IRIX OpenGL is always in a VSync-like mode when it double-buffers. I think "yes".
I also think VSync is always enabled.
I remember reading something about it on usenet a long time ago -- IIRC, someone asked why the FPS counter never goes above the monitor's refresh rate and the answer from an SGI employee was that it's because vsync is always enabled.
So, could I just clarify whether, if the timing in the program may not match to the refresh rate of display, the IRIX OpenGL automatically adjust the timing not to conflict with the display, that is, wait until the beam scan go through end of the frame?
Even if we want to present an image for 1 s (1000ms), the real presentation will be
a bit shorter or longer, strictly speaking? (16.7ms difference with 60Hz frame rate?)