AMD Radeon R9 285 Linux GPU Scaling Performance

[przemoli]

Are you telling us that you see, analyse, adapt and act 160 times per seconds ?? If you are a biological human being this is non sense...

What there is to "see, analyse, adapt" is FPS?

Its binary decision "is there head or nor?", after that its reflex of shooting (and hearing "head shoooooot!" )

Not that I'm saying that 120 vs 160 will save the world. Shaving off 30-40ms from network latency would have bigger impact in most competitive games.

Just I'm against putting artificial limits "because nobody will see the difference beyond that" based on simplified view of human body/mind.

[gutigen]

Are you telling us that you see, analyse, adapt and act 160 times per seconds ?? If you are a biological human being this is non sense...

If you want a prove that a human can see the difference between 60 and 120fps or more, just check how professional players play - unlocked framerate everyone with minimum of 120+ at 120hz monitors. Oh and I, a casual player, can tell a difference between 60fps and 120fps without any issues, even if monitor is only at 60hz. You know why? Cause I can fill the difference in input latency through my 1000hz (1ms latency in comparison to 8ms at 125hz) gaming mouse.

[niner]

Oh and I, a casual player, can tell a difference between 60fps and 120fps without any issues, even if monitor is only at 60hz.

No you think that you can tell the difference. You may even believe that. But as long as it's not the result of a double blind test it is nothing more than that: believe.

For some real numbers. The current best reaction time of the following click test is 124ms:

http://www.humanbenchmark.com/tests/reactiontime/

That's the best time of more than 11 million clicks of people who tried. And it's the simplest possible scenario. Detect that the large red box turns green and click. Extremely simple compared to the average FPS game.
If your game runs at more than 8 (eight!) frames per second, your own reaction time starts becoming responsible for the total. At 60fps only about 1/10 of the total reaction time comes from rendering. That's negligible.

[Veto]

If your game runs at more than 8 (eight!) frames per second, your own reaction time starts becoming responsible.

You are mistaking discrete reaction time from hand-eye coordination latency. I challenge you to ride a bike or fly an airplane with 125ms steering latency and feel the result. There are plenty of scientific studies showing that a latency much lower than that is detectable.
You are right about the DB testing and inflated ego of gamers though :-)

[flim]

If you want a prove that a human can see the difference between 60 and 120fps or more, just check how professional players play - unlocked framerate everyone with minimum of 120+ at 120hz monitors. Oh and I, a casual player, can tell a difference between 60fps and 120fps without any issues, even if monitor is only at 60hz. You know why? Cause I can fill the difference in input latency through my 1000hz (1ms latency in comparison to 8ms at 125hz) gaming mouse.

The problem is, most applications/ games poll events every 10ms (100hz).

[ua=42]

One does notice the difference between different frame rates.

Movies can only get away with 24 fps because of bluring. And even with bluring they are limited on what they can do because of the low refresh rate. (For example, they can't quickly pan because it would just be a blur)

Computer games don't blur (at least not without very expensive computations), so to provide the illusion of movement, it needs more frames.
The faster the frame rate, the faster the action can be on the screen. You can have a quite playable turn based strategy game running at 20 fps, while a shooter would be unplayable at that framerate.

Different people have different fps thresholds for the illusion of movement to work. For me it has to be at least 50 fps, any lower and I can't stand it.

Rendering more frames than your display can display is useless. So if you have a 60htz monitor, the number you look for is 60 fps. If you have a 120htz monitor, the 120fps.

[russofris]

Are you telling us that you see, analyse, adapt and act 160 times per seconds ?? If you are a biological human being this is non sense...

No. When I do act, even if it's only once per hour, I would like the computer to respond within 1/160th of a second.

[gutigen]

Rendering more frames than your display can display is useless. So if you have a 60htz monitor, the number you look for is 60 fps. If you have a 120htz monitor, the 120fps.

Not true, at 120fps even if you monitor is at 60hz, you still will benefit from lower input lag coming from frame latency (in good scenario it's 16.7ms at 60fps and ~8ms at 120fps).

[CrystalGamma]

No you think that you can tell the difference. You may even believe that. But as long as it's not the result of a double blind test it is nothing more than that: believe.

For some real numbers. The current best reaction time of the following click test is 124ms:

http://www.humanbenchmark.com/tests/reactiontime/

That's the best time of more than 11 million clicks of people who tried. And it's the simplest possible scenario. Detect that the large red box turns green and click. Extremely simple compared to the average FPS game.
If your game runs at more than 8 (eight!) frames per second, your own reaction time starts becoming responsible for the total. At 60fps only about 1/10 of the total reaction time comes from rendering. That's negligible.

You do know there is a difference between reaction time and the latency of an anticipated move, right?

[stqn]

Not true, at 120fps even if you monitor is at 60hz, you still will benefit from lower input lag coming from frame latency (in good scenario it's 16.7ms at 60fps and ~8ms at 120fps).

Rendering the frames is still useless though?