AC Multiclass Abarth500/MazdaMX5@Fuji, Wed 22nd March 2023

[stigs2cousin]

That's why I introduced a safety margin:

when your "rich man's PC " runs 45 AI we peasants should be good with 30 racers online

 

[HF2000]

"rich man's PC "

My computer and GPU is 5 years old! Enzo built it for me.
No idea which processor to be honest. Does I7 ring a bell?
GPU something with 2080 I think.

 

[stigs2cousin]

Sounds legit.

So long story short,
Will we "live risky" and rely on luck with big grids and untested mods or will we set a maximum grid size for reliability?

You can probably guess which would be my prefered solution

 

[Allan Ramsbottom]

Could it be server related, i have a I9 with a 3070ti,got real bad stutter and split freezing last nite.Ping was 25

 

[Shed 17]

Switching cars produced my best result for a long time. I'd expected the Abarth to be more of a struggle but John's setup seemed to work well. If more people give it a go then we could have a 50:50 split between cars

Well done on your debt drive in the Abarth Colin . I drove it all of last year with very varying degrees of success (and failure). It seems like John's set up worked well last night, maybe I'll have a play if I get time to see if I can be tempted back, but I am enjoying the stability and late braking capabilities of the MX5 at the moment.

After getting tagged in the T1 chaos and taking to the grass, I settled down to some steady consistent laps behind Giando and managed the undercut after the pitstop. We then had a good battle behind Ricardo and Doug and were very evenly matched.

Well done podiums and thank Han for organizing, we all just need to be a bit more careful on the first lap, we don't want to get a reputation of being as bad as public servers. Maybe I'll go off spider catching and start from the pits next time.

 

[HF2000]

Well done on your debt drive in the Abarth Colin . I drove it all of last year with very varying degrees of success (and failure). It seems like John's set up worked well last night, maybe I'll have a play if I get time to see if I can be tempted back, but I am enjoying the stability and late braking capabilities of the MX5 at the moment.

After getting tagged in the T1 chaos and taking to the grass, I settled down to some steady consistent laps behind Giando and managed the undercut after the pitstop. We then had a good battle behind Ricardo and Doug and were very evenly matched.

Well done podiums and thank Han for organizing, we all just need to be a bit more careful on the first lap, we don't want to get a reputation of being as bad as public servers. Maybe I'll go off spider catching and start from the pits next time.

I don't mind to have a rolling start if the majority wants that. After the lights go out everyone keeps in position behind the leader who starts slowly and chooses the moment to start racing. Just like in the F1 when safety car is ending.

 

[JamKart22]

Could it be server related, i have a I9 with a 3070ti,got real bad stutter and split freezing last nite.Ping was 25

Han, was the server configured with SOL last night as a carry over setting of last weeks night race? I think SOL can be a resource drain so if we don't have a night race then I assume its not needed.

Some of last nights crash fest was due to pinging so a rolling start won't have made any difference. Saying that, the pinging seemed to start after the non-ping crash between the Abarths.

 

[KDiesel]

Could it be server related, i have a I9 with a 3070ti,got real bad stutter and split freezing last nite.Ping was 25

I think it is. When server provider changed I have issues since then, not big ones but still...

 

[HF2000]

Han, was the server configured with SOL last night as a carry over setting of last weeks night race? I think SOL can be a resource drain so if we don't have a night race then I assume its not needed.

Some of last nights crash fest was due to pinging so a rolling start won't have made any difference. Saying that, the pinging seemed to start after the non-ping crash between the Abarths.

So no but the participants who had problems, might have had WeatherFX with SOL activated. I had it disabled because I don't like the redlight reflections of the brakelights on the tarmac.

 

[Kek700]

it is what it is, you will not make 25 drivers neatly sit in line like this old git, I very rarely cause a problem in the first corner. an opportunity has to be bleeding obvious before I go for it, even then I think twice.
My point...
Good luck with making everyone do this, I gave up, I came to realise that the field is just 25 drivers with different psychologies all working in tandem against one another, which is not a good receipt.
Just accept it, and with time and a stable field of every week drivers it will eventually get a self-made cure.
I think our good manners; this has a big impact on new drivers, we get very small problems.

Then you throw in Ping, Formula Agile that spins at a drop of a hat at the first corner, someone seeing an opportunity, elevated heart rates, cold tyres, tunnel vision, single monitors, lag, and inexperience.
I think you are expecting too much, annoying yes, but inevitable.

 

[Mikke73]

Could it be server related, i have a I9 with a 3070ti,got real bad stutter and split freezing last nite.Ping was 25

I have a i5-9400F with GTX 1650, but I had no stutters or freezing at all yesterday, ping was around 30 during the race.
But I did see much more lag spikes than usual, also with the drivers who had a good ping. I think it could actually be server related.

 

[stigs2cousin]

Ernie,

I see your point but I´m not prepared to give up yet.
The discussion shows that we have good awareness of the problem.
If only we could beat some patience and disciplin into these boneheads

Of course I´m kidding

Just keep on trying, we´ll get it right at least once this year

Oh, and I am mellowing out. You might have noticed that lately I´m skipping qualifiing to stay out of the T1cluster****(TM)

 

[RasmusP]

Does I7 ring a bell?

Could it be server related, i have a I9 with a 3070ti,got real bad stutter and split freezing last nite.Ping was 25

i7 reaches back to 2009 iirc. It's like "VW Golf". Definitely a solid car no matter which one, but still a clear difference between the Golf 2 and Golf 7.

Which i9 is it, Allan? The first i9 was a 9900k iirc?

About everything else:
- Server related: Could be, but unlikely since the traffic from AC is basically nothing and the CPU load of an AC server is also basically nothing. But if someone else was tinkering on the Server via Remote View, you could have some spikes for sure.

- About the stuttering: Was YOUR game stuttering or only the other cars but you and the track were smooth? That's the important bit since server lags will NEVER make your own game stutter. It just teleports the other cars but your own car and the track will always stay perfectly fine.

So if your game stuttered/freezed, then it was on your end.

Can someone load his replay and check? Server lags show on replays (jumping opponent cars), stutters usually don't, as long as your physics calculation was still okay.

To check your CPU performance:
Activate the render stats apps. The one from AC and the CSP one:

You can clearly see which Frame time is higher, CPU or GPU. The higher one determines your FPS!
And then you can see which part of AC is taking the most time. In this screenshot: Map Display, SubStanding and CrewChief are using a lot of CPU time!

In the AC app, you can see the Main_T being at 93%. If it goes too close to 100%, you'll get physics calculation errors!
Which will show in the PHY_LATE value.
This is from a replay, so no physics calculations running!

If your PHY_LATE value goes up by more than 50 per lap, you're in trouble.

 

[RasmusP]

Second post with an idea how to check if a combination would be okay for everyone:

We need a test-scenario like a replay with WeatherFX and LightingFX disabled, which is causing problems for some PCs.
Then Han opens these 2 apps and writes down his Main_T and the CPU Frame time.

He then knows that the values that are too much for some drivers.

When he's testing a new combination, he can do an AI race and check Main_T and CPU Frame time.
If it's higher than what he wrote down, he needs to lower the amount of cars.
If it's below the values, then he can go for it.

 

[WRT 53]

Grâce à une araignée et à ma femme, j'ai eu une course très agréable. En partant des stands avec une pénalité de 20s (j'ai oublié d'appuyer sur le bouton drive (encore as digne:!)), j'ai réussi à remonter de la 29ème à la 9ème. Dans cette course, je me sentais comme Max dimanche dernier avec tous ces dépassements sur la plupart des Mazda. C'est super d'avoir cette vitesse supplémentaire dans la ligne droite pour pouvoir passer juste avant T1.
Mais j'ai été surpris que l'Abarth soit si forte ici. Je n'ai perdu que du temps dans le long et rapide conducteur droit où le sous-virage était un gros désavantage. Le plus surprenant était que je pouvais aussi gagner sur Mazda dans la chicane.

Bonne course de la part de tous (sauf tour 1, T1) et merci de vous être joints à vous tous les 28 !
Cette taille de grille m'oblige à rechercher des pistes avec plus de 30 pitspots. Cela limite les choix qui sont bons et mauvais.

Félicitations au podium.
J'espère vous voir tous la semaine prochaine. J'avais en tête l'Arctic Circle Raceway en Norvège, mais il s'est avéré que c'était une déchirure de rFactor et je ne sais pas si c'est légal. Il n'est pas disponible en téléchargement sur RD donc je dois chercher un autre morceau à moins que l'un de vous puisse me donner la preuve qu'il s'agit d'un rip légal.

Salut !

Pourquoi ne pas faire des circuit originaux d'Asseto ?

 

[Mikke73]

i7 reaches back to 2009 iirc. It's like "VW Golf". Definitely a solid car no matter which one, but still a clear difference between the Golf 2 and Golf 7.

Which i9 is it, Allan? The first i9 was a 9900k iirc?

About everything else:
- Server related: Could be, but unlikely since the traffic from AC is basically nothing and the CPU load of an AC server is also basically nothing. But if someone else was tinkering on the Server via Remote View, you could have some spikes for sure.

- About the stuttering: Was YOUR game stuttering or only the other cars but you and the track were smooth? That's the important bit since server lags will NEVER make your own game stutter. It just teleports the other cars but your own car and the track will always stay perfectly fine.

So if your game stuttered/freezed, then it was on your end.

Can someone load his replay and check? Server lags show on replays (jumping opponent cars), stutters usually don't, as long as your physics calculation was still okay.

To check your CPU performance:
Activate the render stats apps. The one from AC and the CSP one:

You can clearly see which Frame time is higher, CPU or GPU. The higher one determines your FPS!
And then you can see which part of AC is taking the most time. In this screenshot: Map Display, SubStanding and CrewChief are using a lot of CPU time!

In the AC app, you can see the Main_T being at 93%. If it goes too close to 100%, you'll get physics calculation errors!
Which will show in the PHY_LATE value.
This is from a replay, so no physics calculations running!

If your PHY_LATE value goes up by more than 50 per lap, you're in trouble.

View attachment 649588

I've looked at my replay yesterday, and in the T1 chaos there was at least one Mazda that kept jumping all over the place. I've never seen anything like that, it was practically teleporting all around.
Also in the lag spikes everything was smooth for me, but suddenly a car that had been 10 meters in front of me jumped 9 meters back. That woke me up...

I also happened to look at a practice session replay earlier, and as I was watching @JamKart22 's lap, his car suddenly had a liftoff - it just went straight from a corner and into the air. Then it jumped back to track.

 

[stigs2cousin]

I had both, stutter/microfreezes which I cured with the FPS cap and brake dancing cars (get it) which suddenly jumped around or even backwards on the Abarth rehabilitation straight .

 

[RS242]

I may have been very lucky or it is indeed on the client side and my rig held good.
I observed no stutters and very little lag from drivers around me - things overall were quite smooth.
It was only a problem on T1 because every car is clustered together, and a small hiccup of +/- 1 meter can indeed cause sudden breaking and collisions.
Oh and I also saw a flying Mazda on T1, launched by the pit exit traffic cones. However the video of Fulvio only shows it spinning in that area - could this be caused by those physics errors Rasmus mentioned?

 

[HF2000]

Salut !

Pourquoi ne pas faire des circuit originaux d'Asseto ?

I'm sorry but I don't understand your question.

 

[Alfred Wayne]

Dont have my replay to look through, but I feel like the first time I noticed this was back on the banking at Daytona, or maybe a race or two before that (maybe a back straight at a Fat Lafie track?) But I only notice it on tracks where there is a lot of long straight drafting, and the issue only seems to pop up after drafting for 5s or so, not immediately (just when you've got your run and want to move to pass ) I haven't noticed any other issues outside of long straights.

And the drivers that it happens to don't have bad ping, aren't jumping around otherwise, nor am I having any game stutters or drops