Blur Busters Forums

You want to be running in 144hz as a max refresh, but cap at ~120fps or something with rivatuner or engine level (if there is one).

If you run in 120hz mode, you will get vsync type input lag when you hit 120 fps

fenderjaguar wrote:You want to be running in 144hz as a max refresh, but cap at ~120fps or something with rivatuner or engine level (if there is one).

If you run in 120hz mode, you will get vsync type input lag when you hit 120 fps

I'd like to see the test that proves this... Has someone proven that vsync is enabled at max refresh rate?

From my personal experience I have not seen any input lag at 120fps/120hz with g-sync enabled and no frame limiter.

Turboman750 wrote:Has someone proven that vsync is enabled at max refresh rate?

Well, it's not vsync, it's still g-sync. It's just that there is the same input lag that you would get when vsync was enabled, because a frame is buffered, even with g-sync, when you hit the max refresh. It's the same with vsync, you can enable vsync and then cap slightly under the refresh and have less input lag. You can even cap at the refresh (cap at 60fps on 60hz vsync) and have slightly less input lag than if you were not capping at all and just effectively letting vsync cap the frame rate.

And the input lag at 120fps 120hz is less than 60 fps 60hz, but it's still there (8ms vs 16ms).

I'd like to see this input-lag through a test if anybody can conduct such a test.

The only issue I have with frame limiters like Rivatuner is the gpu down-clocks itself to run at those frames... So when I go from low visual activity to high visual activity, there is a frame drop and then the gpu has to crank itself back up to make up for that frame drop. I can notice stutter when these things happen - and from my experience it feels worse than your claim of "input lag".

I would like to see chief conduct this test without frame limiters to show the input-lag that you're claiming exists.

Turboman750 wrote:I'd like to see this input-lag through a test if anybody can conduct such a test.

The only issue I have with frame limiters like Rivatuner is the gpu down-clocks itself to run at those frames... So when I go from low visual activity to high visual activity, there is a frame drop and then the gpu has to crank itself back up to make up for that frame drop. I can notice stutter when these things happen - and from my experience it feels worse than your claim of "input lag".

I would like to see chief conduct this test without frame limiters to show the input-lag that you're claiming exists.

The Chief has already done the test when he got hold of a gsync module:

http://www.blurbusters.com/gsync/preview2/

Turboman750 wrote:The only issue I have with frame limiters like Rivatuner is the gpu down-clocks itself to run at those frames... So when I go from low visual activity to high visual activity, there is a frame drop and then the gpu has to crank itself back up to make up for that frame drop. I can notice stutter when these things happen - and from my experience it feels worse than your claim of "input lag".

I'm sorry, but I'm really sure that doesn't happen. Though, If you want, you can easily disable gpu boost 2.0, so that the GPU is constantly running at either the base frequency or the boost frequency.

Imo, GPUboost and heavily capped frame rate (ie 60 fps) gives higher minimum frame rate, since there is more thermal headroom for the boost frequency, should you need it.

Also, the frame rate is still capped at 143 fps, even with no frame rate cap. And the Core will still throttle itself, if it only takes say 50% GPU usage to provide you with 143 fps at any given point in the game. So capping it slightly lower isn't going to give significantly more throttling.

fenderjaguar wrote:

Turboman750 wrote:The only issue I have with frame limiters like Rivatuner is the gpu down-clocks itself to run at those frames... So when I go from low visual activity to high visual activity, there is a frame drop and then the gpu has to crank itself back up to make up for that frame drop. I can notice stutter when these things happen - and from my experience it feels worse than your claim of "input lag".

I'm sorry, but I'm really sure that doesn't happen. Though, If you want, you can easily disable gpu boost 2.0, so that the GPU is constantly running at either the base frequency or the boost frequency.

Imo, GPUboost and heavily capped frame rate (ie 60 fps) gives higher minimum frame rate, since there is more thermal headroom for the boost frequency, should you need it.

Also, the frame rate is still capped at 143 fps, even with no frame rate cap. And the Core will still throttle itself, if it only takes say 50% GPU usage to provide you with 143 fps at any given point in the game. So capping it slightly lower isn't going to give significantly more throttling.

I suppose you're right. I hope Nvidia fixes the input lag at max frame rate in the future so we don't have to worry about 3rd party apps to limit frame rate.

Turboman750 wrote:

fenderjaguar wrote:

Turboman750 wrote:The only issue I have with frame limiters like Rivatuner is the gpu down-clocks itself to run at those frames... So when I go from low visual activity to high visual activity, there is a frame drop and then the gpu has to crank itself back up to make up for that frame drop. I can notice stutter when these things happen - and from my experience it feels worse than your claim of "input lag".

I'm sorry, but I'm really sure that doesn't happen. Though, If you want, you can easily disable gpu boost 2.0, so that the GPU is constantly running at either the base frequency or the boost frequency.

Imo, GPUboost and heavily capped frame rate (ie 60 fps) gives higher minimum frame rate, since there is more thermal headroom for the boost frequency, should you need it.

Also, the frame rate is still capped at 143 fps, even with no frame rate cap. And the Core will still throttle itself, if it only takes say 50% GPU usage to provide you with 143 fps at any given point in the game. So capping it slightly lower isn't going to give significantly more throttling.

I suppose you're right. I hope Nvidia fixes the input lag at max frame rate in the future so we don't have to worry about 3rd party apps to limit frame rate.

G-sync at max refresh is never worse than v-sync's best case scenario. Nvidia and third party tools don't have control over the code necessary to limit framerate in a low latency manner. The game developer has to implement that. I suppose you could hack the IO libraries and insert the delay there, in the hopes the game engine will stall the render pipeline while waiting for those IO calls to return, but that won't work with every game.