Interesting results when I tested frame limit.
Bonanza. 10,500 ft. over terrain. Live Weather ON.
My default is MSFS Frame Sync = OFF, nVidia Frame Limit = OFF.
GPU Temp = 66°C
GPU Load = 97%
GPU Power Draw = 392.4 W
GPU Power Consumption = 87.2% TDP
These settings result in 95 FPS, 10.58 overall latency / 8.8ms Main Thread latency, 9.7ms GPU latency.
Next screenshot is MSFS Frame Sync = ON (50% monitor refresh,) nVidia CP Frame Limit = OFF.
GPU Temp = 60.1°C
GPU Load = 55%
GPU Power Draw = 280.8 W
GPU Power Consumption = 62.4% TDP
These settings result in 60 FPS, 16.67ms overall latency / 7.8 ms Main Thread latency, 14.7ms GPU latency.
Lower FPS, but much higher latency. Not there’s no yellow spikes in the first pic, but solid yellow with a frame rate limit. Not sure if that matters, but it’s certainly notable.
This last screenshot shows MSFS Frame Sync = ON (50% monitor refresh,) nVidia Frame Limit = ON (60 FPS).
GPU Temp = 57.8°C
GPU Load = 56%
GPU Power Draw = 274.8 W
GPU Power Consumption = 61.1% TDP
These settings result in 60 FPS, 16.67ms overall latency / 8.0 ms Main Thread latency, 14.5ms GPU latency.
No significant change when limiting frame rate in nVidia Control Panel.
Conclusion: I don’t understand why frame latency goes up when GPU load goes down. Maybe reducing the GPU load forces the CPU to work harder. But then why is the GPU latency so much higher also?
I’ll play around in high stress areas, like flying street level in Tokyo, or landing at KLAX, to see if I get stutters. If not, then I’ll ignore the yellow spikes and be happy my GPU is running cooler and is less stressed.
Next experiment: I will up my my monitor’s refresh rate to 144Hz, and then try the max refresh rate of 165Hz and set MSFS and nVidia limits to 50% of that number.