I found that crazy thing about AMD (I didn't want to clutter further the other thread) Which possibly in my ignorance of how post-process shaders work, could also put us on the fast track to nicely solve latest #1795 quibble. The other was underlining that even though Microsoft didn't lay out a bed of roses for GL, that could still benefit of the discovery with the interop trick. It should be easily testable on W10 (without the need of special hardware) with PresentMon then. Or perhaps he's thinking too much linux-simplistically and it's not (see here). Indeed, I don't know for sure, and gregory may as well be right. My assumption was that could be adding (non-insignificant) extra input latency to the image. As you can see in the scheme above, this incurs in an extra buffer copy. Second: 95% of times pcsx2 is ran through DWM.
PCSX2 INPUT DELAY WINDOWS
I guess I prefer github chattiness?Īnyway, first of all, an OS: Windows tag would feel appropriate. Forums might have been better probably, but.
Instead this was more consultative rather than anything. Let's say that it would be easier to report the issue as such if it was one in the first place. Last but not least, I changed PFD_DOUBLEBUFFER to PFD_DOUBLEBUFFER_DONTCARE and W-T-F, borderless fullscreen is not exclusive anymore on AMD ()ĮDIT: seems we are all lovingly waiting from you 😘
PCSX2 INPUT DELAY DRIVER
Sample 2 Seems like nvidia/intel/amd all have hacks in their driver to trigger flip/exclusive modeĪside of helping with avoiding BitBlt, could this also use as a stop-gap measure to use DX shaders with GL renderer #1795? Well, surprise: say hello to NV_ DX _interop extension, which leads us back to the previous case. What about poor GL, you'll ask me then? Microsoft is too selfish to reserve it a fast track too (even though for some reason this not-so-seemingly-dumb guy claims otherwise) Though, I guess conditionals for DXGI_SWAP_EFFECT_FLIP_SEQUENTIAL or D3DSWAPEFFECT_FLIPEX not to crash on different OSs could do miracles :p The availability situation of the various swap chains. Now, putting aside stuttering, I'm wondering if those techniques couldn't be used. Follows the initially v-sync focused only brain-storm session post.
0 kommentar(er)
