Given monitor-GPU sync technologies require support from both GPU and monitor sides, would a GPU benchmark result be accurate if measured while the monitor is physically turned off?
If the monitor is physically turned off, I’m curious how the GPU knows when to sync with the monitor to send the next frame.
A benchmark would require Vsync and Gsync to be turned off anyway.
It may run with the monitor off and record a reliable result, but what’s exactly the point?
I’m not asking about legacy v-sync. With g-sync technology, the GPU only sends a frame when the monitor is ready to frame it to your display. If the monitor is physically turned off, I’m curious how the GPU knows when to send the next frame. VRR and Freesync are similar to G-sync.
Doesn’t the monitor only display a frame when the CPU sends one, rather?
Google G-sync.
Digressing. OP.
What does “digressing. OP.” mean?
If you mean that explaining what g-sync is is off-topic from the original post, I believe you’re wrong – in fact it explains why the answer to the question is “it probably doesn’t matter if the monitor is on”.
In case it’s not clear why that’s connected: g-sync/VRR means that the CPU/GPU renders frames as fast as it can and pumps them out to the display as soon as they’re available.
This means the performance of the monitor doesn’t really matter; it’d be about the same as rendering with neither v-sync or g-sync on.
I guess you don’t have a g-sync, freesync, or VRR monitor? Otherwise, you’d understand the OP.
You’re wrong, and you’re wrong in this entire thread. Good-bye, have a nice day.
No, I agree with Vibstronium. And I do have such a monitor FWIW.
With V-Sync, the GPU renders frames to fit a fixed schedule, e.g. once every 1/60 second for 60 Hz. If it could render twice as fast, it will slow down and wait, causing your framerate to drop and giving invalid benchmarks results (lower than what your computer is capable of).
So benchmarks with V-Sync on are a bad idea.
With G-Sync/VRR, the GPU renders frames as fast as it can, and sends them to the display. The display draws them as soon as it can; instantly if enough time has passed since the last frame; there is no scheduled window.
If the GPU can render faster than the monitor can draw, your settings decide what happens, at least for NVIDIA cards. Either it caps at the screen’s max refresh rate as with V-Sync (bad for benchmarks), or you get tearing and a higher framerate (good for benchmarks).
So in short, I would expect that if you have it uncapped once reaching the monitor’s refresh rate, the results should be the same as having no sync and not limiting your benchmarks.
I don’t see the point of having the monitor off, though. Not only is it unlikely to matter, but it’s not a realistic scenario. Why not benchmark the way you will actually play – with the monitor on?
Let’s say you have a 120 hz g-sync monitor and a 144 hz g-sync monitor with g-sync enabled, clearly these two monitors will perform differently in high fram rate games (ie, you’ll get up to 120 fps on the 120 hz monitor and 144 fps on the 144 hz monitor). So my question is, if the monitor is off, what does does the gpu sync with and what performance do you get?
My guess though is that the GPU will turn off g-sync completely when the monitor is off and send frames as they are available up to the refresh rate setting instead of trying to sync.