I found this interesting spreadsheet about input lag for dolphin (gamecube emulator) using different combinations of hardware:
https://docs.google.com/spreadsheets/d/1pz8j5...edit#gid=0
This guy's tests remind me of the input lag tests on mice that was going on around here a couple years ago. He's using photodiodes to detect a switch in the screen from black to white that is initiated by a button press on a USB gamepad. So the test procedure looks legitimate and tests everything between the controller press and what's seen on the screen, which would include USB speed, GPU, and monitor if I understand correctly.
The columns of interest are C and D. These are using the same version of software and the same 120 Hz CRT, but with only changing the computer components. Assuming the CPU+motherboard doesn't change anything (which it may because of USB processing maybe?) the only real difference is nVidia gtx 960 to AMD 5770.
Notice there is a 20 ms increase in input lag caused by GPU change from nV to AMD.
Quote: "it seems quite likely to me that the AMD 5770 card from 2009 is causing much higher latency than the GTX 960 from 2015. This was very surprising to me, I had expected that modern graphics cards should be very similar in processing time."
Now this is relevant to me as I recently upgraded from a GTX 460 to 380x and play quake on a 120 Hz CRT. I have said to myself lately that my setup does not feel as snappy as it did a year ago, but with these things it can be so hard to tell without real data. Maybe this only affects the 5770 this guy uses, and only for the digital->analog conversion [nevermind, the problem is there with digital output as well, though not as bad]. I imagine the DAC is an afterthought these days, and I assume posting this on any other forum people would probably just say who cares about analog output (<3 120Hz CRT).
But my question is has anyone seen any other tests of input lag for different GPUs? And even better would be analog vs digital for the two brands.
TL;DR - AMD card found having higher processing time for analog output compared to nVidia leading to higher input lag, can anyone confirm
https://docs.google.com/spreadsheets/d/1pz8j5...edit#gid=0
This guy's tests remind me of the input lag tests on mice that was going on around here a couple years ago. He's using photodiodes to detect a switch in the screen from black to white that is initiated by a button press on a USB gamepad. So the test procedure looks legitimate and tests everything between the controller press and what's seen on the screen, which would include USB speed, GPU, and monitor if I understand correctly.
The columns of interest are C and D. These are using the same version of software and the same 120 Hz CRT, but with only changing the computer components. Assuming the CPU+motherboard doesn't change anything (which it may because of USB processing maybe?) the only real difference is nVidia gtx 960 to AMD 5770.
Notice there is a 20 ms increase in input lag caused by GPU change from nV to AMD.
Quote: "it seems quite likely to me that the AMD 5770 card from 2009 is causing much higher latency than the GTX 960 from 2015. This was very surprising to me, I had expected that modern graphics cards should be very similar in processing time."
Now this is relevant to me as I recently upgraded from a GTX 460 to 380x and play quake on a 120 Hz CRT. I have said to myself lately that my setup does not feel as snappy as it did a year ago, but with these things it can be so hard to tell without real data. Maybe this only affects the 5770 this guy uses, and only for the digital->analog conversion [nevermind, the problem is there with digital output as well, though not as bad]. I imagine the DAC is an afterthought these days, and I assume posting this on any other forum people would probably just say who cares about analog output (<3 120Hz CRT).
But my question is has anyone seen any other tests of input lag for different GPUs? And even better would be analog vs digital for the two brands.
TL;DR - AMD card found having higher processing time for analog output compared to nVidia leading to higher input lag, can anyone confirm
Edited by Regent at 12:26 CDT, 21 March 2017 - 8577 Hits