Simulator technology

Here are our CFD links and discussions about aerodynamics, suspension, driver safety and tyres. Please stick to F1 on this forum.
beelsebob
beelsebob
85
Joined: 23 Mar 2011, 15:49
Location: Cupertino, California

Re: F1 Simulator - Robot Arm

Post

HampusA wrote:It´s not placebo or any other effect. It clearly delays what i´m doing adn what is shown on screen.
Remember, i can turn it on and off in a matted of seconds so there is no placebo effect going on.
The placebo effect has nothing to do with whether you can turn it on or off rapidly. Again, vsync will not cause this, so either you have some other issue, or you're seeing placebo in action.

User avatar
HampusA
0
Joined: 16 Feb 2011, 14:49

Re: F1 Simulator - Robot Arm

Post

Again there is no placebo effect going on, nor is it my settings since 20-30 people at iracing forum can vouch for this. They all have the same issue.
It might be iracing that is the cause but there is no denying that there is infact a delay from what you do to what the wheel in game do.

This is only not in iRacing but in Portal aswell.
The truth will come out...

beelsebob
beelsebob
85
Joined: 23 Mar 2011, 15:49
Location: Cupertino, California

Re: F1 Simulator - Robot Arm

Post

HampusA wrote:Again there is no placebo effect going on, nor is it my settings since 20-30 people at iracing forum can vouch for this. They all have the same issue.
It might be iracing that is the cause but there is no denying that there is infact a delay from what you do to what the wheel in game do.

This is only not in iRacing but in Portal aswell.
Then again – you have an independent issue going on – I have explained what vsync does technically; you can verify for yourself that this causes no extra latency of input.

Iit's common for the gamer community to claim that it introduces latency because they see their frame rate drop from 150* to 60. This is bollocks for two reasons:
1) Their frame rate hasn't dropped at all, all that's dropped is the number of redundant frames that the graphics card has generated.
2) It is typical for the engine's simulation and graphics loops to run independently of each other, meaning that simulation actually runs faster, because there isn't a load of redundant graphics work going on.

There is one scenario where you might see minutely increased latency. This is where the engine predicts that it can produce a frame in 0.001 seconds*, but ends up guessing incorrectly, and causing the framebuffer to still be filled with the last frame when the monitor swaps. This is obvious when it happens, because your frame rate drops significantly (read instantly drops to 30fps from 60, and causes a jitter in the output). The chance of this happening is more than offset by the fact that the simulation loop gets more time, and frames are produced for a predicted screen refresh time rather than for now.


* Numbers made up on the spot

User avatar
HampusA
0
Joined: 16 Feb 2011, 14:49

Re: F1 Simulator - Robot Arm

Post

Either way, there is a noticable delay when turning your wheel vs turning the wheel in game with Vsync on...
It´s not my settings nor is it a placebo effect. There is a distinct delay so when for example your driving a F1 car you will not be as competitive as with Vsync without it.

This thing only occurs with Vsync on.
The truth will come out...

User avatar
HampusA
0
Joined: 16 Feb 2011, 14:49

Re: F1 Simulator - Robot Arm

Post

Quote frim IR forum,
I just downloaded the missing dll but I just install the runtime components since other programs missing stuff too...

The cross behind the cursor is just plain nailed...

This whole discussion is just much easier then it is discussed. Take Fraps and make a Benchmark with frametime logging. I only post the first 10 frames rendered for every situation.

vSync off

Frame, Time (ms)
1, 0.000
2, 7.063
3, 14.724
4, 22.789
5, 31.154
6, 39.119
7, 47.258
8, 56.659
9, 64.677
10, 73.379


vSync on

Frame, Time (ms)
1, 0.000
2, 14.699
3, 31.985
4, 47.908
5, 64.943
6, 81.112
7, 98.282
8, 114.528
9, 132.041
10, 148.120


With vSync is half as fast as without. In my case that are > 4 meters for 10 frames, behind the position where the care really is... This is on my machine and can be different on other machines but I don't think so. 40 cm per frame is not less to compensate and makes it clear why the car is easier to handle without vsync.
The truth will come out...

Giblet
Giblet
5
Joined: 19 Mar 2007, 01:47
Location: Canada

Re: F1 Simulator - Robot Arm

Post

There is a full on issue with iRacing and vsync. I have yet to experience it myself, but it is not within the normal scope of things. What is being said is true, vsync causes an input lag that directly affects people's lap times.

Vsync is not something that would cause an affect in other games, but iRacing does have this specific bug to it.
Before I do anything I ask myself “Would an idiot do that?” And if the answer is yes, I do not do that thing. - Dwight Schrute

User avatar
hollus
Moderator
Joined: 29 Mar 2009, 01:21
Location: Copenhagen, Denmark

Re: F1 Simulator - Robot Arm

Post

Removing frame 1 where the monitor has no say, for the next 8 frame gaps you have

(148.120-14.669)/8*60=1000.8825ms.

That is what Vsync is supposed to do.
Rivals, not enemies.

beelsebob
beelsebob
85
Joined: 23 Mar 2011, 15:49
Location: Cupertino, California

Re: F1 Simulator - Robot Arm

Post

HampusA wrote:Quote frim IR forum,
I just downloaded the missing dll but I just install the runtime components since other programs missing stuff too...

The cross behind the cursor is just plain nailed...

This whole discussion is just much easier then it is discussed. Take Fraps and make a Benchmark with frametime logging. I only post the first 10 frames rendered for every situation.

vSync off

Frame, Time (ms)
1, 0.000
2, 7.063
3, 14.724
4, 22.789
5, 31.154
6, 39.119
7, 47.258
8, 56.659
9, 64.677
10, 73.379


vSync on

Frame, Time (ms)
1, 0.000
2, 14.699
3, 31.985
4, 47.908
5, 64.943
6, 81.112
7, 98.282
8, 114.528
9, 132.041
10, 148.120


With vSync is half as fast as without. In my case that are > 4 meters for 10 frames, behind the position where the care really is... This is on my machine and can be different on other machines but I don't think so. 40 cm per frame is not less to compensate and makes it clear why the car is easier to handle without vsync.
As I said above, vsync *will* cut the frame rate reported by the engine, it will not cut the number of frames you see – all it does is stops producing a bunch of frames that are never copied out of the framebuffer and onto the screen.

With vsync on, you will render at most at 60fps, which is exactly what's going on here.

Caito
Caito
13
Joined: 16 Jun 2009, 05:30
Location: Switzerland

Re: F1 Simulator - Robot Arm

Post

First you have to say that it doesn't matter your fps, you will never watch more frames than your refresh rate on your monitor.

So if your monitor has a refresh rate of 60hz, you'll never see more than 60 frames a second.

What vSync does is eliminate image tearing.


Lets put it this way: we have a box(buffer) where the graphic card puts the frames. The monitor checks the box and reflects that frame.

The monitor will check the box 60 times a second. If your game fps are at 80, the video card is filling in the box 80 times a second. The box has space for only 1 frame.

So, what is tearing? The monitor gets a frame and display it, then the graphic card updates a full frame, plus a third of other. Then the graphic card fills the missing 2/3 plus 2/3 of a new frame, again a new cropped image. Finally the graphic card puts the missing third, and a complete 3/3 of a new image. So you get an un-cropped frame every 3.
(EDIT: this is because fps is 33% higher than refresh rate, so for every 1 frame updated to the monitor, the game has created 4/3, or 1 and a 1/3)


vSync was invented to solve that problem. vSync basically is the boss who checks the box. It tells the graphic card that it can't fill the box until right after the monitor retired an image. This is good as long as your fps is higher than your refresh rate. If they're lower, let's say 50hz, you'll get an effective fps of 30hz.(I'll not expand on this unless you really need it, triple buffering came to solve this problem)



So if your FPS are ALWAYS above your monitor refresh rate, you'd rather have vSync on. Triple buffer may actually slow you down.

Basically vsync wouldn't cause the problem.

That's like saying a deaf "I'll give you 1000usd if you say hi to me", and then conclude that the guy doesn't want 1000usd.
Come back 747, we miss you!!

User avatar
HampusA
0
Joined: 16 Feb 2011, 14:49

Re: F1 Simulator - Robot Arm

Post

Vsync comes at the expense of delay, not very good for catching tenths.
The truth will come out...

User avatar
HampusA
0
Joined: 16 Feb 2011, 14:49

Re: F1 Simulator - Robot Arm

Post

I was wrong about the part where i said it was a Mathematical model i think.

I just read a comment in the forum,

"The iRacing NTM is a theoretical based model... the first of its kind in the world."

I´m assuming it´s the first of it´s kind in sim racing right?
The truth will come out...

beelsebob
beelsebob
85
Joined: 23 Mar 2011, 15:49
Location: Cupertino, California

Simulators

Post

To take this stuff out of the Ferrari F150° thread, lets have a thread about the various teams' simulators.

Massa stated three interesting things about the Ferrari one:
• It runs on only 60MB of RAM – said apparently in a tone saying that it was a severe limitation.
• It can't simulate the Pirelli tyres
• It doesn't have all the tracks available.

User avatar
Javert
5
Joined: 10 Feb 2011, 14:14

Re: Simulators

Post

Massa's interview is taken from italian "Autosprint" newspaper. The last two points beelsebob said are confirmed

For the first one, I asked the journalist who made the interview to confirm

User avatar
JohnsonsEvilTwin
0
Joined: 29 Jan 2010, 11:51
Location: SU 419113

Re: Simulators

Post

What are the limits imposed on Data, speeds and information transfer?

Mercedes have a new sim nearly fully operational, but what good would a new one be if its abilities are limited to that of the old one by the rules?
More could have been done.
David Purley

wesley123
wesley123
204
Joined: 23 Feb 2008, 17:55

Re: Simulators

Post

is there info for what they run for example? It sure cant be rF Pro, I doubt that even starts with 60Mb of ram.

I cant even understand how this is done, 60Mb was the amount we ran 15 years ago, I even doubt why they would even have to let a manufacturer make a 60Mb ram kit for way too much when a 2Gb kit is way cheaper.

I seriously cannot understand it, in the simulator you have loads and loads of information, and you just want that information always available, so unless they have some sort of special CPU with sh*tloads of cache, I cannot understand how this is done
"Bite my shiny metal ass" - Bender