Blur Busters Forums

It seems like there is currently no further explanation in this but they claim it is essentially decresing input lag relative to the framerate by untying the input and the fps. If I misunderstood it please do correct me but that would be a huge leap forward in terms of software. Imagine higher polling rates combined with this technology. I'm hyped.

The first clue is that they said it has to do with software that exists between the gpu and the monitor. I'm not sure if they meant software specifically on the gpu or not. But those were the words they used.

The second clue is that they showed data that they said was from a 60fps situation. The lag goes from ~55ms to ~45ms.

This may be a vsync related optimization. At 60fps and with the vsync mode not specified there are certainly many possibilities for input lag reduction. I would love to see it be something cool and new.

Perhaps they dynamically shift the rendering constraints to start work at a later time in constrained vsync rendering. Perhaps it works like fast sync requiring much higher fps than refresh rate for maximum benefit. Perhaps it involves more with extra driver optimizations working in conjunction with refresh timing related techniques.

open wrote:Perhaps it works like fast sync requiring much higher fps than refresh rate for maximum benefit.

Nah, they already have that ("AMD Enhanced Sync"). It has to be something else. But everything they said so far is too confusing.

Don't get your hopes up. There is a reason why they used 60fps scenario. My bet is it won't do anything at high fps/high hz scenarios.

AMD doesn't currently have a user-side max pre-rendered frames setting exposed/available, right? Perhaps this trick (at least partially) relies on a reduced pre-rendered CPU and/or GPU frames queue?

That, or some sort of "early" detection or prediction of user input, as they said it's going to be decoupled of framerate, apparently. I've seen similar techniques used in emulators. E.g. RetroArch's "Early" "Poll Type Behaviour" option, where user input is supposedly read before the frame is processed.

I'm sure we'll find out soon enough.

GPU bottleneck can cause input lag in some games, this seems to be a fix/optimization for it

Did they made this test with vsync off? I remember that Nvidia promoted the GTX 750 with a big input lag reduction. But there was no difference compared to ohther graphic cards.

jorimt wrote:AMD doesn't currently have a user-side max pre-rendered frames setting exposed/available, right?

I think radeon pro software has some sort of option like that. Im not too sure

Simon95 wrote:Did they made this test with vsync off? I remember that Nvidia promoted the GTX 750 with a big input lag reduction. But there was no difference compared to ohther graphic cards.

This is a software/hardware reduction. This most likely is not bs. Espescially since the developer explained it as untying the framerate from the input polling. What concerns me is them saying it is mechanical input to monitor pixel response time. Could also be software driven overdrive or something. I hope it's untied polling.

For what it's worth:
https://www.pcgamesn.com/nvidia/nvidia- ... lternative

Tony Tamasi, VP of technical marketing, says [Nvidia] has an alternative to [Radeon Anti-lag] too… and has done for over a decade [...] That’s the setting ‘maximum pre-rendered frames’ in the GeForce control panel, which controls the size of the Flip Queue. The higher the number here, the higher the lag between your input and the displayed frame, while lower queue lengths result in less lag but increased strain on your CPU.