Just recorded an idle frame (no code, no objects, Vsync On, real fullscreen). The idle frametimes are ok, but could still be better, it`s idle in the end. Same results with different idle setups with Vsync off, windowed mode. Seems Fusion gets the real issues when there's some kind of load to do, but it's not a basic idle state problem. Next thing I'll record is probably the sample app from the other posts and check that.
Well, looking closely and comparing to Axiom Verge which does the job right I would like to withdraw my statement that the idle runtime is ok. I've recoreded the microstutter example from Vol too, which shows about the same frame times as an idle frame. Compared to Axiom Verge (bottom image) which hits the needed 16.7 ms almost every time, Fusion rather seems to oscillate around the 16.7 line without almost ever hitting it. The frame times rather seem to accumulate with a standard deviation plus/minus the desired 16,7 ms. The average numbers per second (FPS) are ok this way, but the per-frame output is not and causes the microstuttering issue.
Have you considered setting up a remote desktop session with someone who experience the issue to try and debug it on their end? Obviously both parties would need to be willing but if some user experiencing it wish to do so, I think it would be the easiest way to do so. I'd volunteer, but I don't experience (or just not notice) the stuttering in examples or actual games and this computer is too weak to run any remote desktop app anyway.
Originally Posted by Simon
The lag from network latency would make microstutter even harder to spot or isolate.
I also think we should steer away from this “I have to see/experience it to believe it” thing. We obviously already demonstrated different options to measure and quantify. If someone can present a non-stuttering Fusion baseline (a non stuttering idle frame) on his setup this would raise the question if all of this is actually setup dependent. If not, it IS a Fusion runtime issue.
@Salamanderpants there is a pattern here of you not reading/misunderstanding what people are saying, and then attacking them in some way.
Originally Posted by salamanderpants
You say "yet Yves has changed the status of my bug report to "feedback" ". You made that comment, complaining he'd done so, 2 days ago. This is the "feedback" Yves gave 12 days ago
Can someone explain to me why changing the Window's Theme wouldn't be relevant to the issue? I feel like since mentioning it, there hasn't been anyone who found interest in that observation I made.
If changing my theme to Windows Classic removes the stuttering, wouldn't that be an important indication as to what might be causing it? So far, no one else has tried to validate this observation on their end. It makes me feel a bit frustrated.
Also, If others aren't experiencing any stuttering, perhaps they should check their theme and use a custom one that might be more demanding? I also have another monitor hooked up to my laptop, so perhaps that adds to the Theme processing. Perhaps those who aren't experiencing the stuttering could hook a monitor or two and use a custom theme to see if it introduces stuttering. Maybe even place some documents on each screen to ensure there is processing going on each monitor.
sorry dude, i did try again but still cudnt figure out how to set windows 10 to classic :(
@Prominent: Just recoreded Vol's mictrosttuer example on classic theme. Makes absolutely no difference.
The windows theme used may affect performance in games that run windowed fullscreen, but I doubt it should affect setups that are on real fullscreen, as windows (in theory) should not be involved there in handling the display output. As I mentioned in an earlier post: Using windowed vs real fullscreen makes no difference on my machine, when it comes to the recorded frametimes. However, real fullscreen somehow feels a lot better.
@Shin: Get it from there and copy it to "C:\Windows\Resources\Ease of Access Themes". Then you can select it next to the other default skins. Pretty nice theme, actually. Think I'll keep it.
cheers julian :)
& no, it made no difference, the microstutter is still there :(