Misc ACC Results Companion 1.24.0

[omhent]

Hi Doug,
Just wanted to thank you for this great app and the support you have provided. Amazing! I sent you a coffee for your efforts today.

Apologies if my question has already been addressed, but I searched and looked back a few months and couldn't find anything.

I'm noticing that if I score a podium, the #P1/2/3 does not increase in 'result statistics', even though the # of laps and #races does correctly increase. It gets recorded correctly in 'session' as a position 1,2, or 3 in a race which results in there being many more podiums in the sessions than results stats.
I thought it might have something to do with # of racers for an official race to count, but I have seen this with a race of 5 and a race with 20 and the same thing happens. I only race MP so can't comment on SP
I have tried rebuilding and it doesn't add up.
Am I doing something wrong?

 

[Doug Duthie]

Hi Doug,
Just wanted to thank you for this great app and the support you have provided. Amazing! I sent you a coffee for your efforts today.

Apologies if my question has already been addressed, but I searched and looked back a few months and couldn't find anything.

I'm noticing that if I score a podium, the #P1/2/3 does not increase in 'result statistics', even though the # of laps and #races does correctly increase. It gets recorded correctly in 'session' as a position 1,2, or 3 in a race which results in there being many more podiums in the sessions than results stats.
I thought it might have something to do with # of racers for an official race to count, but I have seen this with a race of 5 and a race with 20 and the same thing happens. I only race MP so can't comment on SP
I have tried rebuilding and it doesn't add up.
Am I doing something wrong?

This sounds like a bug. The query building the result stats has become ridiculously complicated, so it's possible the aggregation is incorrect. On the basis that I rarely finish in the top 3, it might be quicker if you provide your diagnostics, and I'll trace through to see what is going on - instructions can be found here

EDIT: Also, can you list a few examples of where things are wrong to make it quicker for me to find

 

[omhent]

Thanks, I sent the file.
user: QyXQsTXLwNRH0YTZnagsBlhVHC4QKVlcIH80wNvnOos

The most recent example was Feb 4 2nd place zolder. Laps and race added to #s, but the #P2 in stats didn't increase.
I can't be certain, but it seems to me that the more podiums already in the stats, the less likely it is to be added. If its the 1st podium at a track, it seems to get added more readily. Such as when I change the track year in 'fix session' and a new row is added into the results, it will show the podium that wasn't showing before I changed the track year (hope that makes sense).

Feb 3 2nd place at nurburgring is another example that didn't get added to podium stats but did get added to laps and races.
I think those 2 examples are correct but I didn't keep track unfortunately of the exact sessions, but should have.

Thanks for the help.

 

[omg.imba]

Yes - can you upload the diagnostics - instructions here

Doing so right now. UserID: 1fr2BAoM7Q6wIQ-isARR7YuO1_5lDMgMFE2DuIZgtao
(trying at least. The dialog seems to be stuck on a full upload bar. Will try again and update the userid)

Can't remember when the first practice session was missing. But the most recent one was a ~90min session on Kyalami yesterday evening (2021-02-05 ca 22:30)

In addition, in the second to last race (2021-02-06 20:18, Imola, 25min) Laps 2-5 were not recorded for me. All other cars seem to be present, although I didnt check that in detail.

 

[Doug Duthie]

@omhent , @omg.imba - I've received the diagnostics. I'll have a look later

 

[Doug Duthie]

Thanks, I sent the file.
user: QyXQsTXLwNRH0YTZnagsBlhVHC4QKVlcIH80wNvnOos

The most recent example was Feb 4 2nd place zolder. Laps and race added to #s, but the #P2 in stats didn't increase.
I can't be certain, but it seems to me that the more podiums already in the stats, the less likely it is to be added. If its the 1st podium at a track, it seems to get added more readily. Such as when I change the track year in 'fix session' and a new row is added into the results, it will show the podium that wasn't showing before I changed the track year (hope that makes sense).

Feb 3 2nd place at nurburgring is another example that didn't get added to podium stats but did get added to laps and races.
I think those 2 examples are correct but I didn't keep track unfortunately of the exact sessions, but should have.

Thanks for the help.

It was a rather stupid bug. If you had more than one P1/P2/P3 for a given car/track combination, it just set the count to 1 rather than to the number of podium positions

Before

After

EDIT: Also, there was a fllter to discount sessions where no position was reported. I've included this now as there are some circumstances where it might not be able to have calculated the final position in the past

 

[MB190E]

I can't reproduce this one at the moment, but I'm going to keep an eye on it. It sounds as though events are being missed - whether this is the control, the app or Windows, I can't tell at the moment.

Can you go through all the steps in order to reproduce this (from starting ARC). Once it starts working, is it good, or does it always revert to this unresponsive state? Is the rest of the app unresponsive while this happens. I've still got a copy of your ARC environment from December, so I'll see if I can reproduce

Also, while ACC and ARC are running, can you give me a screen shot of your memory usages (from Task Manager)
View attachment 441700
Also, can you show me the summary page of System Information
View attachment 441701

Hi Doug,

I’m following up on this but have nothing to show. As it happens, since I reported it, I have not been able to produce the problem. Not that that is a bad thing, of course.

If it starts again I’ll do as you suggested.

 

[Doug Duthie]

Doing so right now. UserID: 1fr2BAoM7Q6wIQ-isARR7YuO1_5lDMgMFE2DuIZgtao
(trying at least. The dialog seems to be stuck on a full upload bar. Will try again and update the userid)

Can't remember when the first practice session was missing. But the most recent one was a ~90min session on Kyalami yesterday evening (2021-02-05 ca 22:30)

In addition, in the second to last race (2021-02-06 20:18, Imola, 25min) Laps 2-5 were not recorded for me. All other cars seem to be present, although I didnt check that in detail.

Sorry - the diagnostics dialog isn't the best. When the progress bar goes all the way to the right, it has completed the upload - it should be a bit more obvious it has done something.

I can see the root of the problem. I look for two events. One if checking if a car appears to have crossed the line by position. There is also a broadcast event that indicates that a car has just crossed the line, with the time that it did so. I need to look for both as they hold different information (one holds the lap/sector times and the other logs the time of the end of lap - I need the latter to determine car positions).

The positional check if pretty much synchronous and real time, whereas the line cross event is asynchronous. I need to wait for this before storing laps. On my system, the lap cross event occurs before the position check 95% of the time. The other 5%, it occurs within the next second. On your system, it is taking up to 105 seconds. This is suspiciously close to the lap time so I'm not sure at this point whether it is just slow, or events are being missed.

I'm going to do some more investigation to see if there is anything obvious, but I think i may just need to add more logging to help resolve the issue

 

[Doug Duthie]

Sorry - the diagnostics dialog isn't the best. When the progress bar goes all the way to the right, it has completed the upload - it should be a bit more obvious it has done something.

I can see the root of the problem. I look for two events. One if checking if a car appears to have crossed the line by position. There is also a broadcast event that indicates that a car has just crossed the line, with the time that it did so. I need to look for both as they hold different information (one holds the lap/sector times and the other logs the time of the end of lap - I need the latter to determine car positions).

The positional check if pretty much synchronous and real time, whereas the line cross event is asynchronous. I need to wait for this before storing laps. On my system, the lap cross event occurs before the position check 95% of the time. The other 5%, it occurs within the next second. On your system, it is taking up to 105 seconds. This is suspiciously close to the lap time so I'm not sure at this point whether it is just slow, or events are being missed.

I'm going to do some more investigation to see if there is anything obvious, but I think i may just need to add more logging to help resolve the issue

@omg.imba - It looks like lap completed broadcasting events are sometimes not being sent from ACC. I'm going to see what extra logging I can put in to see what is going on

 

[omhent]

It was a rather stupid bug. If you had more than one P1/P2/P3 for a given car/track combination, it just set the count to 1 rather than to the number of podium positions



EDIT: Also, there was a fllter to discount sessions where no position was reported. I've included this now as there are some circumstances where it might not be able to have calculated the final position in the past

Thanks Doug, much appreciated!

 

[omg.imba]

@omg.imba - It looks like lap completed broadcasting events are sometimes not being sent from ACC. I'm going to see what extra logging I can put in to see what is going on

Nice to see tha you're getting more info from my diagnostics. Yes, the dialog confused me a bit, but it is working

Will upload a new set, this time with data from a 4h kyalami endurance race with driver swap. (The race finished after ~2.5h due to a server disconnect). Here not only some laps are missing, but the big majority. I was driving from (probably) lap 40 to 70. From this stint, 4 laps are in the results. The rest of the time I was spectating my co-driver, and ARC didnt collect any lap from before that and 5 of the remaining 8 laps.
Similar for other cars, there are a lot of laps missing. For example, lap 53 was only recorded for two out of 46 cars.

I assume there are additional oddities at play with the driver swap, but I guess it's connected to the same issue as before.

You were saying, ACC lost / forgot to send out broadcasting events. You have any idea what could cause that? high CPU load? network issue? If others are not seeing this, I guess the issue is on my end.

userid: 1fr2BAoM7Q6wIQ-isARR7YuO1_5lDMgMFE2DuIZgtao

 

[Doug Duthie]

Nice to see tha you're getting more info from my diagnostics. Yes, the dialog confused me a bit, but it is working

Will upload a new set, this time with data from a 4h kyalami endurance race with driver swap. (The race finished after ~2.5h due to a server disconnect). Here not only some laps are missing, but the big majority. I was driving from (probably) lap 40 to 70. From this stint, 4 laps are in the results. The rest of the time I was spectating my co-driver, and ARC didnt collect any lap from before that and 5 of the remaining 8 laps.
Similar for other cars, there are a lot of laps missing. For example, lap 53 was only recorded for two out of 46 cars.

I assume there are additional oddities at play with the driver swap, but I guess it's connected to the same issue as before.

You were saying, ACC lost / forgot to send out broadcasting events. You have any idea what could cause that? high CPU load? network issue? If others are not seeing this, I guess the issue is on my end.

userid: 1fr2BAoM7Q6wIQ-isARR7YuO1_5lDMgMFE2DuIZgtao

Thanks for the extra diagnostics. I'll look at these later to see if there is a pattern.

I'll have to check again, but it looked as though lap completed events were only dropped for you but appeared present for all the other drivers. It looks as though ACC isn't sending those messages, but I need to do low level logging of the UDP messages to be sure.

I've not heard of anyone else having this issue. It would be worth checking your cpu usage before ACC and ARC are running and with them both running (when ACC is just in the UI, you should be seeing <= 10% cpu. I might add just extra optional diagnostic logging to monitor cpu usage in case you are maxxing out.

It is good you have tried driver swaps - I've no idea if this will work well or not at all in ARC. It will give me a chance to see if I can improve this area

 

[omg.imba]

CPU usage of ARC is in the 8-10% range whenever I check during a race (i5-6600K).

 

[ajclarke]

I can't tell if there's bugs, or if its me mis-using the software, or a limitation of what ACC itself outputs but I have two quesetions:

• My results companion seems to think every track I race on is 2018, and doesn't detect the "year" of the track. I think this might make a slight difference with quickest lap times, since v1.6 (so if I've understood correctly any 2020 European track) lap-times have gotten slower
• Sometimes my fastest lap in game which ACC thinks is valid, is considered invalid by results companion.

Awesome tool though, thank you very much for all your work on it. The fuel calculator alone is amazing, so much easier than typiing laptimes and race times in a calculator!

 

[Doug Duthie]

I can't tell if there's bugs, or if its me mis-using the software, or a limitation of what ACC itself outputs but I have two quesetions:

• My results companion seems to think every track I race on is 2018, and doesn't detect the "year" of the track. I think this might make a slight difference with quickest lap times, since v1.6 (so if I've understood correctly any 2020 European track) lap-times have gotten slower
• Sometimes my fastest lap in game which ACC thinks is valid, is considered invalid by results companion.

Awesome tool though, thank you very much for all your work on it. The fuel calculator alone is amazing, so much easier than typiing laptimes and race times in a calculator!

1. The first is a restriction of ACC's interfaces - it basically doesn't report which year of track you are running on. The workaround to manually change the track in ARC can be found here
2. This is an interesting one. I don't know exactly why, but I have some theories. ARC flags a lap as invalid whenever ACC reports it as such on it's interfaces. I think the ACC UI applies different tolerances to track limits depending on the session type (eg qualifying is much stricter than races). However, I think the ACC interface always applies the same tolerance - if you don't keep wheels within the track limits/kerbs it will flag as invalid. I'm comfortable with this behaviour as it means there isn't an exploit to get faster times on the leadboard by pushing limits in races

 

[ajclarke]

Thanks for the information. That makes sense.

 

[MB190E]

I have events at the new BTCC tracks. If I leave ARC running to not lose out on session data, will I be later able to correct the “unknown track to reflect what it really is once ARC is updated?

 

[Doug Duthie]

I have events at the new BTCC tracks. If I leave ARC running to not lose out on session data, will I be later able to correct the “unknown track to reflect what it really is once ARC is updated?

Yes... although the new tracks should show as "donington", "oulton_park" and "snetterton" until I release an update to properly recognise them.

Are yours being reported as "Unknown track"?

 

[MB190E]

I tried it today and I see “oulton_park.” I didn’t expect It to work.

 

[unpierrot]

Strange thing because I didn't remember seeing it before but label for track is "Row Limit" in 1.10.0 (see picture)