FCAT benchmarking an article introduction [Guru3D.com]

Incredible article hilbert.

I always new about the one frame delay in SLI (being an SLI user) and have always been a tiny bit turned off by it.

BUt still... great to see this research being done. I laughed at your "website begins with an "A" joke too.

 

Higher input latencies? Cap FPS at refresh rate, force Triple Buffering on, and enjoy almost as low input lag as no VSync.

Capping FPS to half-refresh? With single cards, Triple Buffering solves that issue and lets FPS run between 30 and 60. With multi-GPU setups, a technique in AFR inherently similar to Triple Buffering is used, so VSync on multi-GPU does NOT cause dips to half-refresh FPS.

I don't know about AMD drivers and VSync and Flip queue, all I know is that I can force VSync + TB on in RadeonPro, and set Flip Queue size (and yes it does have an effect, I felt it).

The experience would be identical with VSync as you'll see it is butter smooth regardless of the card driving the VSynced game / benchmark. No dropped frames / runt frames would be observed.

It's only when the GPUs are at 99% and VSync on dipping below refresh rate FPS that microstutter might (not confirmed) rear its ugly head again.

I always used to play without VSync, but the advantages of VSync far outweigh the disadvantages regardless of whether you have an AMD or Nvidia graphics card.

 

Good stuff.

p.s. doing some "Tom" raiding lately?

 

So when should VSync ever be used?
Personally I never enable it but again under what condition does it help if at all?

 

Well I guess that would depend on your monitor. I always have terrible screen tear without vsync I cant stand it. This tv seems to be better in that regard then my last one but still http://www.youtube.com/watch?v=K9z3MQeogcY that shows you what screen tearing looks like.

 

I have to use Vsync on the original Dead Space otherwise the flashing lights look like ****, the mouse is faster with the 30 fps cap moved to 60 fps with vsync.



Btw, in Soviet Russia games frame interval latency benchmark you!

 

Too bad there is no Metro2033 tested here. That game has some issues on NVidia cards. There are some huge (visible) latency spikes in the benchmark.

 

I agree, though I typically cap my FPS a couple of frames below refresh rate to prevent any input lag caused by additional pre-rendered frames.

 

This is so old so i hope Hilbert sees this:

Hi,
I just found your An introduction to FCAT benchmarking - A FCAT Conclusion article hoping it will be step by step guide how to do everything [Including links to all the tools since something you spoke about like adding colored line to every frame, i have no idea what tool can do that], but alas it wasn't a guide but introduction.

What was specially interesting is your usage of "capture PC" with capture card to analyze PC games, which should open up doing the same for console games.

Right now there are no real tools to analyze console FPS, Digital Foundry uses in house tools that they build and otehr YT that analyze console video use FCAT [I think, since there is no otehr tools to analyze captured video and create a graph].

Its been 5 years since then, can you consider creating a full guide step by step and all that, how we can do it using a second Capture PC, the usage of capture card is important since it will also allow doing the same for console games.

Regards