UKGPL Season 22 - Monaco 65 and Silverstone 67 - Ferrari Challenge - Jan 03

[clouds]

Hi.
Like I said yesterday nite, at first I'm sorry to have rearended Bernie just at the last corner before the chequered flag because I misjudged the braking distance so my apologies to Bernie are due, and at second after the heap just a few meters after the green flag in wich I've involved I...unfortunately forgot to stop & go in the pitlane so take your decisions.

For the second race at Silverstone I've been suffering of a throttle command inversion (and this is not the only one case I've had). While I was slowing down approaching the Stowe turn chased by Hristo who was closing on me very fast, when I've raised the throttle and  pressed on the brake pedal, the trottle had a strange behaviour and seemed like to be pressed at maximum overreving the engine and making me hitting the wall just in front of me.
Fortunately my Ferrari seemed not having any damage but it has been difficult to get the car on track again because this kind of uncalibration in pedals has lasted after 4 or 5 seconds making me almost spinning out returning on track while the throttle pedal was having an On/Off behaviour.

And last but not least...why I was not been able to save my replay ?
When I try to run a replay in 60fps why GPL Crashes To Desktop ?

Thanks

[Cookie]

Sergio,
please read the second post here

Fulvio and Nicky, thanks again for your help!
I never had this overload before with my server, and the races of last week did not show any signs of this.
But it worked fine in Silvy after the change!

[fpolicardi]

Sergio,
please read the second post here

Fulvio and Nicky, thanks again for your help!
I never had this overload before with my server, and the races of last week did not show any signs of this.
But it worked fine in Silvy after the change!

Link doesn't work, see last post for the right link.
Your welcome!
Ciao

[clouds]

Rgr that Axel.

Thanks!

[Hristo Itchov]

And no, I didn't notice any difference in handling with the 60 fps patch.

I do not agree with you: the game's physic does not change, but, imho, the game is more quickly in response to input.

That might be the case for some PC and wheel/pedal configurations, but I really didn't see any difference on my end. If anything, it even makes it easier to drive, because everything goes smoother.

@Axel - I know the Monaco times issue is not related to the 60 FPS patch, it was due to the bandwidth settings causing trouble. I'm not sure how you're going to correct the lap times, but I know we've done that before in UKGPL so somebody else knows how for sure. I know for certain that Kleinfeld, Sky and Bernie were actually lapped (nothing personal), but I don't know about the others down the order.

[Cookie]

If anything, it even makes it easier to drive, because everything goes smoother.

@Axel - I know the Monaco times issue is not related to the 60 FPS patch, it was due to the bandwidth settings causing trouble. I'm not sure how you're going to correct the lap times, but I know we've done that before in UKGPL so somebody else knows how for sure. I know for certain that Kleinfeld, Sky and Bernie were actually lapped (nothing personal), but I don't know about the others down the order.

Thanks for this compliment!

As we all learn, and as I have never before seen those bandwidth settings on a 60fps server before, we hope that this problem is solved now!

About the Monaco results, we will find a solution, but it will need some time as I am a newbie in LM2...

[NickyIckx]

Maybe we can test this with BREASTS on a full grid with F2s and 65s tonight?

good idea , I would then suggest 65s at Monaco  and F2s at Silverstone .

I will set up my server for both , using recommended settings outa mentioned readme.
( 5/84 5/384 ) .afaik my internet provides enough down/upload to handle 384 .

But , Axel , if you want to test your server , of course I will be fine with that .



  thx Fulvio for hint ! mostly I do join via IP . the moment I have to join my own server , I had to .
so (If you join by Igor you don't need to set anything on client as usual, only if you join direct IP you need to set core.ini parameters. ) was very important information to me

[Cookie]

Great Nicky!

If Clive agrees:
Lets try your server

I have this 6000 Internet by 1&1

I have set my server to 5/84 and 5/196 now for FC practice Kyalami  

[Phil Thornton]

Firstly a big thanks to Axel for organizing this! 

The first race at Monaco was a bit of a disaster for me.  I arrived late and so had to start at the back of the grid.  When the flag dropped everyone disappeared and then it started raining Ferrari's.  Cars were appearing from everywhere.  When I did get going the field was pretty strung out so I couldn't tell if there was more or less warping.  When I did catch another car (and when I got lapped) the connection seemed pretty good and there was no discernable warp.  As far as car control was concerned, the steering was more precise and the braking more responsive.

The second race at Silverstone was much better.  The very tight grid resulted in me banging wheels with someone and I spun out shortly after the start.  However there wasn't any warp as far as I can remember.  If it hadn't been for the tight grid the start should have been pretty clean.  During the race the server was rock solid.  No warping and the cars were always visible (until I got left behind of course).  Again the car control was much improved.

Overall a good test for me but I suggest we need to do a few more tests before we consider running the 60fps patch for a league race.  Personally I think it is the way to go but we have to be sure. 

As far as I can ascertain the 60fps patch should (in theory) give us less warp contacts and (my experience so far confirms this) better car control.  Warp is the single most problematic issue in on-line racing as far as I'm concerned so anything that reduces it has to be worth investigating.  If the tests don't back up the theory then we'd have to decide if the extra complications with using the 60fps patch was worth the effort.

Good advice from Fulvio regarding the core.ini settings.  Particularly the bit about only having to change the client and server send rates from 3 to 5 for the client when connecting by IP.

However this is the bit that I don't quite understand about the 60fps patch.  The 384 is the bandwidth that sets the number of cars that you can see.  The 3 or 5 dictates how many ticks between the server or client sends,  i.e. the rate at which the server and client exchange data about the position of the cars.  To reduce warp we need better position data so surely we need less ticks between sends. 

According to Bart Westra "The time unit in GPL. 1 Tick = 1/36 seconds or ~27.778ms. When there are enough system resources, GPL displays a new frame every tick to give you 36 fps (frames per second)."

So using 5 with the 60fps patch means we have pretty much the same latency with the 60fps as we have with the standard 36fps.  So how is that going to help with the warp?

[Hristo Itchov]

According to Bart Westra "The time unit in GPL. 1 Tick = 1/36 seconds or ~27.778ms. When there are enough system resources, GPL displays a new frame every tick to give you 36 fps (frames per second)."

So using 5 with the 60fps patch means we have pretty much the same latency with the 60fps as we have with the standard 36fps.  So how is that going to help with the warp?

Good point, it should produce the same kind of warp contacts as with 36 fps. Perhaps it's only less pronounced warp contacts when we run on setting 3 and 60 fps, but as we saw in these races, that requires less than 12-13 people on the server, at least on Axel's.

[BadBlood]

I believe that part of the 60FPS patch is to make GPL reset the tick to 1/60th of a second.

We currently use send_every=3 so that give 12 updates a second. Using 5 with the 60fps would imply 5/60 of a second per update = 12. There is nothing to stop us trying other parameters but we are limited by bandwidth whatever we do so it is a case of finding a sweet spot.

I think when Cookie had it set to 3 he was trying to update 20 times a second - 4 would be 15 times a second. It may help with warp but lose with lag where frames are dropped. Case of suck it and see. I had noticeable 'wobble' at various points but connectivity has never been a big issue for me. The major effect of 60FPS is client side.

[clouds]

Honestly speaking I've not noticed so much difference between 36 and 60fps in a stand alone stint, at least not so much like I was reading since the 1st exit of the 60fps patch. At monaco I had much warp and I've seen cars flashing in front of me while it has been not at Silverstone.

[fpolicardi]

Both frequent ratio (3,4 or 5) and number of cars visible(132 for 8, 384 for all) have big impact on Upload bandwith of the server, while they have less impact on Download bandwith and more impact on CPU/GFX performance of clients.

I'll install 60fps patch on T7 server so we could test next race on T7 server that has a commercial 10/10 Mbs connection and see how things go with 3/XXX or 4/XXX settings. As we saw last race connection problems arised soon on the server in practice as clients went over 12/13 so we can easy adjust the settings in a little time.

Ciao

[Cookie]

Thanks Fulvio for this offer!

I am not happy with the Monaco race inconvenience we all had,
but I am happy to have made this experience, because we learned a lot about this connection related stuff.

To me it seems that there are 2 different things intermingeled:

1. The 60 times per second updated collisionbox of the cars  <-> 2. the X times per second client - server updates

These 2 factors are independent and do not interfere!

@1 the 60fps collisionbox with its smaller peaks is less than the 36fps collisionbox, so you will have less warp and you can have closer fights!  


@2 The client-server update rate has an effect only to the server!
      In this "Monaco" case the traffic just was too high for my servers connection as it was nearly double upload Kb/s and download Kb/s!
      

[Cookie]

Now I need your help to get the right Monaco results!

The problem is caused by a bug in GPL or the Monaco track!

RACE FASTEST LAPS

Pos Driver                                 Time Lap
 1  Alex Barresi                        12.606s   1
 2  Albrecht Kleinf65                   13.048s   1
 3   Pierre D                           17.211s   1
 4  Bernie65 Darwin                     20.833s   1
 5  Nigel Smith65                       20.906s   1
 6  Hristo Itchov65                   1m27.809s  16
 7  Bob Whitwell                      1m30.140s  14
 8  Mark75 Barresi                    1m30.918s  15
 9  Dave Rainier65                    1m31.451s  15
10  Sergio Lonzar 65                  1m31.696s  16
11  Skymole 2173 -m65                 1m31.753s  17
12  65 evilclive                      1m31.765s   3
13  Axel Cookie                       1m32.132s   2
14  Nicky Ickx65                      1m33.169s   6
15  Jason Blito                       1m33.222s  10
16  Phil Thornton65                   1m33.384s  14
17  Axel Cordt                        1m41.660s   4

RACE RESULTS (After 18 laps)

Pos No Driver                          Team     Nat Laps   Race Time       Diff
 1   6 Hristo Itchov65                 Ferrari  JPN   18  26m55.339s           
 2  10 Albrecht Kleinf65               Ferrari  GER   18  27m26.296s    30.957s
 3  14 Mark75 Barresi                  Ferrari  ITA   18  27m39.980s    44.641s
 4   9 Sergio Lonzar 65                Ferrari  ITA   18  28m22.605s  1m27.266s
 5   1 Skymole 2173 -m65               Ferrari  JPN   18  28m25.549s  1m30.210s
 6  16 Bernie65 Darwin                 Ferrari  GBR   18  28m33.436s  1m38.097s
 7  12 Dave Rainier65                  Ferrari  MON   17  27m06.571s   1 lap(s)
 8   8 Jason Blito                     Ferrari  GBR   17  27m43.281s   1 lap(s)
 9  17 Phil Thornton65                 Ferrari  GBR   17  28m02.005s   1 lap(s)
10  22  Pierre D                       Ferrari  CAN   17  28m10.252s   1 lap(s)
11   2 Bob Whitwell                    Ferrari  GBR   16  26m59.436s   2 lap(s)
12  11 Nigel Smith65                   Ferrari  GBR    9  13m44.613s   9 lap(s)
13   3 Axel Cookie                     Ferrari  GER    6   9m52.814s  12 lap(s)
14  21 Nicky Ickx65                    Ferrari  RSA    6   9m55.463s  12 lap(s)
15   4 Axel Cordt                      Ferrari  GER    4   7m59.003s  14 lap(s)
16   5 65 evilclive                    Ferrari  GBR    3   4m52.254s  15 lap(s)
17   7 Alex Barresi                    Ferrari  ITA    1     13.621s  17 lap(s)

- Hristo won it! no doubt
- Albrecht claims to have been only overtaken once, by Hristo!?
- Mark is your 3rd position correct?
- Sergio is your 4th position correct?
- Sky is your 5th position correct?
- Bernie is your 6th position correct?