- 140
- England
- PantalonUK
GT5's Game-Breaking Online Flaw (OP Updated: 11 Feb)
- Thread starter MGR
- 1,131 comments
- 96,741 views
- 2,385
- Retired to a padded cell
- GTP_Padded_Cell
- Microsoft sucks..
Probably a daft question but are PD actually aware as yet of the seriousness of our problem??
I don't think this is a daft question at all - BEST question more like IMO.
How come Polyphony were so quick to recognise the shuffle racing "bug" after the last update but not this "lap time bug"????
Nice one Polyphony, not only do many guys (including me) have to change our setups on all our online race cars (100+ cars) after you change physics or implement new features (that usually don't work 100% right straight away) in updates - now you do this to us??
👎
The changes made within the last 2 updates have caused carnage with our penalty systems in our race series too. We've had to retest our cars TWICE now, then change all the relevant info on the 'net TWICE too.
Managed to get that all done - start our new season only to be greeted with this new issue, meaning it's just not posible to race, too many guys being afected.
👎 x2
Extremely poor (make that NON existent) acceptance testing and customer service.
👎 x3
And all you can do is give us pretty race suits, helmets and "new" cars (which all sem to be Skylines, mostly - despite the E30 M3 being the most wanted car mentioned in feedback).
👎 x4
Online racing (????) - pfffft, yeah right..... 1 year after release and it's worse now than ever before, plagued and littered with problems for an entire year.
Polyphony - you suck.
A dream come true (GT online with friends), utterly ruined...
👎 x100
I've spent so much time organising race series in the past year, regularly spending 20-40hrs a week organising, hosting and helping others get decent online racing - biggest problem isn't us, it's the online connectivity and reliability.
I got cable internet, my PC runs off landline (phone) internet to not affect my PS (on the cable).
All those who are going moan at me for posting this - you try sacrificing all those hours every week for a year for others, and see how many times connection issues ruin or spoil things. Or how changes hidden in updates play havoc with races series rules or handicaps, meaning you have to spend hours and hours updating setups on your cars, re-testing all cars in series and amending rules / regulations on the net - then you'll undestand.
If you haven't hosted / organised for that amount of time - you can't understand what it's like.
It's just so, so, demoralising and frustrating....
- 1,695
- Norway
- ErikAndre71
My upload speed is 60mbit(usually I get 7,1MB/s upload speed), but still it's bad news. It lagged when I hosted (on fixed room ownership) with 13 players in the room. Race quality High (default) and mic quality on low
- 15,972
- Devon UK
- johnboy-1975-
Highlandor well im glad im not the only one
I too spent hours (i mean hours) testing cars for our V8 series trying to find up to 8 competitve well matched cars just as we think we have done all we can and after launching the thread (more time!!) Boom this hits us square on the chin possibly delivering the knock out blow. The cars feel totally foreign from the setups i had painstakingly spent hours tweeking the lap times show this some i can no longer even corner with anymore wtf!!
I for one will certainly not dedicate anymore time to another GT5 series if it was worth the paypal fees (another story
) it would go on Ebay tomorrow!!
Lucky your based in Japan PD i would be at the front of the protest march otherwise!
I too spent hours (i mean hours) testing cars for our V8 series trying to find up to 8 competitve well matched cars just as we think we have done all we can and after launching the thread (more time!!) Boom this hits us square on the chin possibly delivering the knock out blow. The cars feel totally foreign from the setups i had painstakingly spent hours tweeking the lap times show this some i can no longer even corner with anymore wtf!!I for one will certainly not dedicate anymore time to another GT5 series if it was worth the paypal fees (another story
) it would go on Ebay tomorrow!!Lucky your based in Japan PD i would be at the front of the protest march otherwise!
- 12,018
- Indianapolis
- BrandonW77
Hi, long time reader, first time poster. I've been following the thread since the beginning, think I've read the majority of it.
I've experienced all the issues: fast bug, slow bug, frame rate lag, loss of grip when other cars are near. I have been running a weekly series since the PSN outage ended and we very rarely have ever had any issues. For our new Alfa series I created a "private lounge ID" that we use to race in, like many other series have done recently. Since we've been racing in the lounge of the ID that is always offline, we have had all of these problems arise. Before this season we ran in my private lounge and I was always online, never had problems even with 13-14 cars.
I wouldn't think this would be a problem, but is there any possibility a lot of these things happening are due to racing in the lounge of an ID that's not online? Perhaps it's trying to send packets of info back to the host ID but keeps getting a dead end because that ID is offline?
I'm no expert in these matters, hell I'm not even a novice. This is just a thought that occurred to me last night that I haven't seen discussed yet so I thought I'd mention it.
I've experienced all the issues: fast bug, slow bug, frame rate lag, loss of grip when other cars are near. I have been running a weekly series since the PSN outage ended and we very rarely have ever had any issues. For our new Alfa series I created a "private lounge ID" that we use to race in, like many other series have done recently. Since we've been racing in the lounge of the ID that is always offline, we have had all of these problems arise. Before this season we ran in my private lounge and I was always online, never had problems even with 13-14 cars.
I wouldn't think this would be a problem, but is there any possibility a lot of these things happening are due to racing in the lounge of an ID that's not online? Perhaps it's trying to send packets of info back to the host ID but keeps getting a dead end because that ID is offline?
I'm no expert in these matters, hell I'm not even a novice. This is just a thought that occurred to me last night that I haven't seen discussed yet so I thought I'd mention it.
- 15,972
- Devon UK
- johnboy-1975-
Pole Postion now theres a proper game!!
Here is the data from trying 14 people in a fixed host room.
Hosted on a comcast business class cable modem, typical U/D to many speedtest.net servers across USA 12/65Megabit, latency 5ms-75ms. PS3 gigabit wired.
Track: Cape Ring Inside (tight track worst case)
Room Quality: High Voice: Standard
Aids off, except ABS 1
Grid Start with False Start check
Boost: Off
Penalty: Weak
Visible Damage: Off
Mechanical Damage: Light
Slipstream Strength: Strong
Grip Reduction: Real
PP: 575 (rx-7 tc, elise rm were primary cars).
The game is trying to use a ton of upload bandwidth, based on the usage there in a 14 player non fixed host room it would require every player to have at a minimum 2Mbit upload, and with some headroom 3-4Mbit for 14 players. I wish I could monitor PS3 CPU usage somehow because in this case I honestly think my host PS3 was maxed out and accounted for the lag during the race. I would really like to try with a second PS3 hosting but not racing, but I don't currently have access to another PS3. I think beyond that the only possibly thing to try to work around this would be to put 16 PS3's on a local network, which is pretty unrealistic.
Please Polyphony Digital put your best guys on the netcode for the next patch.
Ahh Tomato.
So you're seeing this in fixed host rooms as well? I'm beginning to think it's less and less a network issue and more to do with boost sneaking into rooms where it's supposed to be disabled and throwing a monkey wrench into the whole thing.
And 2mbit per peer in non-fixed room isn't accurate. You're using just over 2mbit TOTAL for all 14 clients as the host. Each peer would use a percentage of the 2+mbits if it where a lounge or non-fixed lobby. I've never seen more than 40KB/sec in rooms like this.
- 10,456
- United States
- WrecklessAbandon
Nope, happens in every type of room from what I hear.
- 12,018
- Indianapolis
- BrandonW77
^ Ok, makes sense. Just a thought
- 347
- MGR
Here is the data from trying 14 people in a fixed host room.
Hosted on a comcast business class cable modem, typical U/D to many speedtest.net servers across USA 12/65Megabit, latency 5ms-75ms. PS3 gigabit wired.
Track: Cape Ring Inside (tight track worst case)
Room Quality: High Voice: Standard
Aids off, except ABS 1
Grid Start with False Start check
Boost: Off
Penalty: Weak
Visible Damage: Off
Mechanical Damage: Light
Slipstream Strength: Strong
Grip Reduction: Real
PP: 575 (rx-7 tc, elise rm were primary cars).
The game is trying to use a ton of upload bandwidth, based on the usage there in a 14 player non fixed host room it would require every player to have at a minimum 2Mbit upload, and with some headroom 3-4Mbit for 14 players. I wish I could monitor PS3 CPU usage somehow because in this case I honestly think my host PS3 was maxed out and accounted for the lag during the race. I would really like to try with a second PS3 hosting but not racing, but I don't currently have access to another PS3. I think beyond that the only possibly thing to try to work around this would be to put 16 PS3's on a local network, which is pretty unrealistic.
Please Polyphony Digital put your best guys on the netcode for the next patch.
That's very good data. Cheers 👍
Although I don't understand how you came to the conclusion that every player in a non fixed hosted room would need 3-4Mbit upload as well? The whole point of running a non-fixed room is to reduce the amount of lag by sharing the network bandwidth requirements.
Using your fixed host data for example (I'll use the peak values in this case) would suggest that every PS3 is providing at most around 51Kbits/sec (ie 666kbits / 13 players - as one player is the host). In a fixed host connection the host must collect all data from every PS3 and then re-distribute the locational data of all cars to every PS3. The total upload in you example is 3451kbits which would suggest the host is providing every client PS3 with around 265kbits of data. This also suggests the host PS3 is consolidating / compressing the incoming data from 666kbits to under half the size - probably an attempt to reduce network lag.
So basically
Host U/D : 3451 / 666
Client U/D : 51 / 265
In a non-fixed room every PS3 is communicating with every other PS3. Assuming the same 51kbits of data is being sent then every PS3 is uploading 666kbits and also downloading 666kbits. The network load is distributed evenly across all consoles.
The downside is that because every PS3 must be communicating properly with every other PS3 in the room then connection issues can be more common. Good routers utilising UPnP should be able open the appropriate ports to ensure proper connection between consoles. Correctly set up port forwarding (or at the very least putting the PS3 IP in the routers DMZ) should provide an even more robust connection.
That all said, I still believe this issue is not related to network speeds.
- 591
- London, UK
- TAM-Spencer
As you just said this issue isnt 100% network speed related, we didnt used to have these problems, they have cropped up from one of the patches pre Xmas, im keeping my fingers crossed that it will be fixed in the next patch as it seems to me that they have changed something for the worse, wether its thier already terrible net code or somthing else is impossible to tell, but somthing has changed as i used to run full rooms with no worries all the time.
- 591
- London, UK
- TAM-Spencer
Btw Highlandor, I also run a league with my friends and you summed up exactly how i feel m8
chuyler1
Premium
- 4,548
- New Hampshire
- chuyler1
I ran a fixed host last night. 7 drivers planning to run a 1.5hr enduro on Nurburgring.
9:30-10:15 EST - No issuses during practice free run, race started at 10:15
10:18 - 3 minutes into the first lap of the race, I got booted from PSN network and had to restart the game. Everyone was dropped from my lobby, a few had to restart their game.
10:30 - race starts back up again with a new fixed host lobby
10:50 - after two laps, one driver complains of some lag. He races with chase cam and said the combo of the new motions + the lag was making him nauseous. A few minutes later it improved enough for him to make up some time.
11:25 - around lap 5, a different player gets booted from the race. I was racing right behind him when he disappeared. There was absolutely no lag between he or I before it happened.
11:30 - we all exited the race when the booted player returned. We gave up on the endurance and just ran a few 10 minute races on some other tracks. No lag to speak of for the next 40 minutes.
What a waste of time. I guess online racing is only good for sprints. If the network cannot gracefully handle a few hiccups then it will be impossible to run extended races.
9:30-10:15 EST - No issuses during practice free run, race started at 10:15
10:18 - 3 minutes into the first lap of the race, I got booted from PSN network and had to restart the game. Everyone was dropped from my lobby, a few had to restart their game.
10:30 - race starts back up again with a new fixed host lobby
10:50 - after two laps, one driver complains of some lag. He races with chase cam and said the combo of the new motions + the lag was making him nauseous. A few minutes later it improved enough for him to make up some time.
11:25 - around lap 5, a different player gets booted from the race. I was racing right behind him when he disappeared. There was absolutely no lag between he or I before it happened.
11:30 - we all exited the race when the booted player returned. We gave up on the endurance and just ran a few 10 minute races on some other tracks. No lag to speak of for the next 40 minutes.
What a waste of time. I guess online racing is only good for sprints. If the network cannot gracefully handle a few hiccups then it will be impossible to run extended races.
- 10,456
- United States
- WrecklessAbandon
I'm confused why you would quit 1 hour into a 1.5 hour race just because 1 player DC'd. ???
Fixed host = Potential room destruction, always has.
Chase Cam = People have been complaining that it makes them nauseous, not related to network issues or lag at all though.
No offense, but the "problems" you had in that room seem like jack squat to a worst case scenario.
Fixed host = Potential room destruction, always has.
Chase Cam = People have been complaining that it makes them nauseous, not related to network issues or lag at all though.
No offense, but the "problems" you had in that room seem like jack squat to a worst case scenario.
I agree with the theory that P2P should use less bandwidth, but the data I gathered earlier below, it looks as if in the second case the total traffic is being limited. In any case it looks like you'd still probably want atleast a 1Mbit upload for 16 P2P, if the network is the issue.
I wish we had some recorded bandwidth from 1.x patches, or even 2.0, but I couldn't find any to compare.
I also agree that I don't think the root cause is the network limitations, this was just the only metric I could gather (would really like ps3 processor utilization). Whatever change happened I believe has put more data into the network ticks, and this is causing increased network usage, and also increasing the CPU load on the PS3. So if you have fast connections the network is fine, but you still start having issues from CPU load, and if slower network conditions are involved you get a mixture of both.
But as fun as that speculation is, I am really trying to just figure out a workaround, and sadly I do not think there is one other than reducing to 10-12 player rooms. I would love to try/hear how a 16 player P2P all on a local network goes, but even if that works great its not much help. It is really all in Polyphony's hands now.
Code:
7 players non-fixed host (no-lag)
432kbit/s upload - 547kbit/s download bandwidth during race. (peak)
16 players non-fixed host (lag):
597kbit/s upload - 457kbit/s download bandwidth during race. (peak)chuyler1
Premium
- 4,548
- New Hampshire
- chuyler1
Worst case scenario? No...just frustration that every night something goes wrong...but if you were the one driver with the lag issue, I'd say it is worst case scenario. In any case, "fixed host" does not prevent lag from occuring, it just means that the race will end if the host gets dumped.
Normally I wouldn't DC, but when its the one driver who has a chance at beating me...who is in the lead by 5 seconds...and the rest of the drivers are 30+ seconds behind me. What's the point of continuing? We are a tight group...and enjoy racing with one another. If someone DC's that is part of our core group, we'll usually just run a new race so all can participate.
- 347
- MGR
I have a theory which may explain the link between the missing replay data and strange physics feeling. And it relates to a post by SuperSic earlier in this thread. Initially I disagreed, but now I think I understand the point SuperSic was trying to make.
It's like those affected by the bug are suffering 'micro-pauses' within their game world. I'll try to explain.
Missing replay data
Earlier in this thread I posted some replay timing data from a bugged replay;
Free-run lap : 2:11.8 (game clock)
2nd flying race lap : 2:17.6 (game clock) - 2:11.6 (replay data)
Around 6 seconds lost.
There seems to be a strong correlation between the amount of lap time lost and the amount of missing replay data based on this example (and other bugged replays I myself and other posters have tested). The time difference in this example is around 4.5%
Working on the assumption that GT5 is running at 60fps (I know it's not solid 60fps but the actual framerate is not really relevant) then the 'stuttering' could be from the game pausing frames 2 or 3 times per second. Tiny pauses of this nature will be detectable to the naked eye - and will be worse if the time difference is something in the realm of 15 seconds like some have been reporting.
The game timer clock continues to run during these 'pauses' but the paused frames are not recorded by the replay data. This would account for the time difference between the game clock in a bugged replay and stopwatch time.
So while your bugged car is pausing 2 or 3 times per second (each pause less than two hundreds of a second) the unaffected car next to you has moved on up the road just a little.
Physics change
I described the change as a loss of grip, power and braking. But now I think I see what SuperSic was getting at regarding our perceptions of speed and grip.
I noted before that the time lost for this example lap was 6 seconds (approximately 4.5%)
Now consider your car is travelling at 100km/h - but the game is 'micro-pausing' 2 or 3 times per second. Because of these added paused frames the driver perceives or 'feels' the car to be travelling slower than it actually is in the physics engine. Because the driver perceives the vehicle to be moving 4.5% slower (say 95.5km/h) than it actually is then they will brake at an appropriate point - based on their perceived (slower) vehicle speed.
However, the car is not travelling at 95.5km/h in the game engine. It's actually travelling at 100km/h and the braking point selected by the driver is too late! Not only that, but now that they've hit the brakes (and the game is still micro-pausing) the car is now taking 4.5% longer (in perceived time by the driver) to pull up for the corner.
Cornering has the same 'slow' feeling. Due to the micro-pauses the car is perceived by the driver to be travelling slower than it should! And relative to the rest of the unaffected cars racing - it IS travelling slower!
This theory would explain:
- The slow lap times
- The framerate stuttering
- The lost replay data
- The perceived loss of grip and power.
It's like those affected by the bug are suffering 'micro-pauses' within their game world. I'll try to explain.
Missing replay data
Earlier in this thread I posted some replay timing data from a bugged replay;
Free-run lap : 2:11.8 (game clock)
2nd flying race lap : 2:17.6 (game clock) - 2:11.6 (replay data)
Around 6 seconds lost.
There seems to be a strong correlation between the amount of lap time lost and the amount of missing replay data based on this example (and other bugged replays I myself and other posters have tested). The time difference in this example is around 4.5%
Working on the assumption that GT5 is running at 60fps (I know it's not solid 60fps but the actual framerate is not really relevant) then the 'stuttering' could be from the game pausing frames 2 or 3 times per second. Tiny pauses of this nature will be detectable to the naked eye - and will be worse if the time difference is something in the realm of 15 seconds like some have been reporting.
The game timer clock continues to run during these 'pauses' but the paused frames are not recorded by the replay data. This would account for the time difference between the game clock in a bugged replay and stopwatch time.
So while your bugged car is pausing 2 or 3 times per second (each pause less than two hundreds of a second) the unaffected car next to you has moved on up the road just a little.
Physics change
I described the change as a loss of grip, power and braking. But now I think I see what SuperSic was getting at regarding our perceptions of speed and grip.
I noted before that the time lost for this example lap was 6 seconds (approximately 4.5%)
Now consider your car is travelling at 100km/h - but the game is 'micro-pausing' 2 or 3 times per second. Because of these added paused frames the driver perceives or 'feels' the car to be travelling slower than it actually is in the physics engine. Because the driver perceives the vehicle to be moving 4.5% slower (say 95.5km/h) than it actually is then they will brake at an appropriate point - based on their perceived (slower) vehicle speed.
However, the car is not travelling at 95.5km/h in the game engine. It's actually travelling at 100km/h and the braking point selected by the driver is too late! Not only that, but now that they've hit the brakes (and the game is still micro-pausing) the car is now taking 4.5% longer (in perceived time by the driver) to pull up for the corner.
Cornering has the same 'slow' feeling. Due to the micro-pauses the car is perceived by the driver to be travelling slower than it should! And relative to the rest of the unaffected cars racing - it IS travelling slower!
This theory would explain:
- The slow lap times
- The framerate stuttering
- The lost replay data
- The perceived loss of grip and power.
- 1,106
- NC/NoVA
- dudeondacouch
Probably a daft question but are PD actually aware as yet of the seriousness of our problem??
Are they aware of anything would be a better question.
dsgerbc
Premium
- 2,577
- dsgerbc
What I don't get, is why this thread wasn't advertised in the 'news' section. This is way more important than tire degradation test. The only way for PD to take notice if this becomes a threat to future GT5 sales and this _should_ be publicized on the front page of this web-site.
chuyler1
Premium
- 4,548
- New Hampshire
- chuyler1
+1
It was my thread that was featured. I'm flattered that the two hours I put into testing tires got a little recognition...but since then, I haven't been able to get a long enough race in with other drivers to see my results in action. This network lag/ frame rate / lap time issue is much more important than tire wear.
- 218
- nonmovingtarget
Signed.
I had a couple of >14 player races yesterday; ran into the same bug. Stopwatch data differed from GT data by around 4%, with all of the above symptoms.
Thinking about it, similar does happen even free run - when you are in a qualifying lap, and another player enters or leaves the track - there is a micro stutter, which will be felt immediately and also is reflected in the lap time.
I am crossing my fingers that PD is already aware of this issue and will address it in one of the next updates.
- 15,972
- Devon UK
- johnboy-1975-
- 1,019
- johnkiller2
The makes perfect sense. Great thinking connecting all symptoms into a logical conclusion. All fits together.
Could the root cause be communication. The game is sampling the race at a certain rate. What should the game do if at one sample one or several cars data is missing (coming too late). It has to do an assumpion and the best assumption would be that the car(s) would be continuing at the same speed/direction, so all acceleration/breaking/steering data is lost.
Question. When a race is being computed, is the race itself being ran at each individual PS3 and the communication is synchronizing all cars, or is it being computed at the server end and the cars position being uploaded/downloaded to the server from each PS3as the race continues. If the latter, maybe the problem and fix lies to increase server capacity.
Last edited:
cicua
Premium
- 944
- Copenhagen
- GTP_cicua / cicua
I and several others have reached the same conclusion during this thread 👍
This idea requires the assumption that physics calculations are tied to framerate. But once you get it, it all makes sense.
But is it caused by overloaded CPU's, or is it simply a coding bug that causes these skips?
If CPU, it makes sense that fewer cars and no damage/weather could lessen the problem.
If other bug, what triggers it, and can it be avoided? This is where network analysis might provide answers.
- 231
- Rio de Janeiro
- RowdLL
GT Planet has become (or maybe always been) too positive towards the game, PD and Kaz... maybe they fear posting critics will hurt their relationship. Unfortunetaly it is starting to become less relevant as a news source.
Right now, the forum is a much more reliable source than the news page. But ultimately, if even the good and solid arguments raised in this forum are not adressed by GTPlanet or PD, even the forum can loose its relevance and die.
GTPlanet MUST make sure this issue reaches PD in a PUBLIC way so we keep on trusting it.
- 15,972
- Devon UK
- johnboy-1975-
x2:tup:
cicua
Premium
- 944
- Copenhagen
- GTP_cicua / cicua
+1 sadly
Similar threads
