[RESOLVED]DCS 1.5.7 stutters heavily unit related

[BIGNEWY]

Since ED is aware and looking into it, and you have all the test missions you need, I'll just add this one thing.

 

I don't think it's just SAM units. We have been running a Vietnam mission series in our squad. The missions are very light on SAM units but heavy on infantry. After the 157 patch they became unplayable in the same way others have described in this thread. I think it's just AI numbers in general.

 

Our squad reverted to 156, and the missions all run fine again.

 

open a new thread, attach the mission I am happy to take a look at it

Edited September 4, 2017 by BIGNEWY

[Wingthor]

As a Field Specialist in TAWs DCS Division my responsibility is to keep members happy, I therefore takes feedback from members after our weekly event. Since 1.5.7 release most of the feedbaks is concerning the DCS bugs. This bug makes game unplayable for most users. We en TAW value each member of our community highly and have no one to loose. I will also claim most of our members is dedicated DCS World players. If leaving cause of the bug they also stop playing DCS world and ED is therefor loosing valuable customers. I should not have to say this.

 

Reason why I speak up now is that I feel we are on the limit. At 13. of July this year FlightControl reported this issue. Its nearly two month ago. I dare to say if ED had prioritized this issue it would have been resolved. There might be good reason ED has prioritized different. We are the dedicated DCS players, we created the missions and provide dedicated servers also for the causal players. We add value to DCS world and thereby also ED. If ED loose us the game will not be the same, and it will take years to rebuild. This as a fact.

 

I take the liberty to share with you some of the feedback I have to deal with every week after 1.5.7 came out. This after myself and others have spent hours in trying to making event as great as we can, and we now feel ED is not taking their part. We cannot do anything other than report and hope for a response, and that is frustrating.

 

Realize this is not my words, and it has been filtered as good as I can so I don't violate anything, and feedback's is given anonymously.

 

Lagging at start

Very low FPS, had to quit

Server crashed...

Server Crash..

A few lag spikes and one server crash after about 90 minutes.

Server crushed

Frame rate drops, lag.

Lag spikes

Server Lags (blame ED)

One plane got buried into the landing strip and Wingthor was bouncing around quite a lot for about 10 seconds, but thats what i would expect from DCS by now

Bad lag spikes server wide resulting in teammates dying unnecessarily.

Some quite severe lag spikes, especially towards the beginning of the evening

Freezes, lags, but most of all the Blue Transport flew like a mosquito on cocaine. Too bad ED won't fix their shit with lags and freezes. We had started a big engagement that would have been very spectacular if not the freezes.

Massive lagging/screen freezing going on from time to time

Large amount of lag

Massive lag due to too many AI's. to the point I'm thinking of leaving TAW.

Extreme lag.

server lag. no fun at all.

Lag for days,Only problem was lag other than that it was great.

The lag was terrible. I shot some video of what I was seeing, but I'm sure it's nothing new. Let's hope ED does a ton of work on netcode.

Severe lag spikes in the beginning, eased up as we went along

Lag spikes, freezing game (blame ED)

Stuttering and framedrops

Lag at & near airbase.

Verry laggy

Lag. But thats caused ED. We may reduce moving units around home base.

Server problems unplayable Sugestion please make a simple mission for next sunday. For testing for the community. That's probarbly the problem that the server can't handle. Minimize scripts for testing. If the test passed you can slowly pur scripts in it. I think this is better than the server goes crazy.

I hope you now realize how serous this matter is! Also provide dxdiag file from server.

DxDiag.txt

[BIGNEWY]

You don't have to tell me how serious it is, do you see anyone else replying to the threads, gathering the data for reports, and checking users missions?

 

If my comment upset any of you I am sorry, but there is not a lot I can do about that. Feel free to PM me.

 

ED is looking into it, if it was an easy fix it would be fixed already, it is high priority and they are working on it.

 

That's all I will say about it now, we have to be patient.

 

Keep the comments about the bug and for sharing data that might help.

 

Thanks

[Wingthor]

Ok, thanks! And wish ED all of luck!

[Looney]

Dxdiag attached.

DxDiag.txt

[NineLine]

1.5.7 broken that. Random freezes and stuff. Windows 8.1 x64

 

The 860M isnt doing you a lot of good though... Its probably gonna be considered below spec.

[NineLine]

Dxdiag attached.

 

That 770 is on the lower end. It should be ok in 1.5.7, but realize when the switch to 2.5 happens, you will probably be haveing many more issues. Even as we move towards 2.5, 1.5.7 is going to be getting more demanding...

[EntropySG]

As I said in me previous posts, its from our end less about individual performance, but rather about a problem we're having in multiplayer. FPS was fine, we had net code issues like desync and warping.

 

Anyway, as requested, here is my dxdiag.

BIGNEWY, I hope you were not offended by my post Sir, we greatly appreciate your help, really and all the time you spend in an effort to help.

DxDiag.txt

[evilivan]

As I said in me previous posts, its from our end less about individual performance, but rather about a problem we're having in multiplayer. FPS was fine, we had net code issues like desync and warping.

 

Anyway, as requested, here is my dxdiag.

BIGNEWY, I hope you were not offended by my post Sir, we greatly appreciate your help, really and all the time you spend in an effort to help.

 

...what @EntropySG said. My dxdiag attached.

DxDiag.txt

[BIGNEWY]

As I said in me previous posts, its from our end less about individual performance, but rather about a problem we're having in multiplayer. FPS was fine, we had net code issues like desync and warping.

 

Anyway, as requested, here is my dxdiag.

BIGNEWY, I hope you were not offended by my post Sir, we greatly appreciate your help, really and all the time you spend in an effort to help.

 

Rarely offended, just sometimes old and grumpy :)

 

thanks for the info, it really helps when ED are looking into issues.

[Pikey]

Just for clarity, is the "Warping issue" (low client update speed) the same as the "stutters"? I've been trying to detangle folks' terminology on the above and often single players chime in with their take on it.

 

What I have with just three connected clients in MP to a single dedicated server is a low update speed, like one update per second of objects, such as client, but also the AI aircraft. The server frames show 30 but the CPU is maxed and the app eventually becomes unresponsive.

 

The mission is what was previously a "mid tier" complexity. One SA-3, a few ZSU's, up to 6 AI planes and 8 moving vehicles, rest static. It's nothing like the things I used to put together going back to 1.12 which had unit counts way over 100.

 

I can't work on testing because my friends simply won't play the mission through bad experience. I can make other missions without these issues, and have to admit the exclusion to date is SAMs.

 

The symptom client side is great frames, but server is dying worse than early Blueflags at the end of a four hour play session. The updates are down to 1 second intervals, planes pointing in the sky, darting backwards and forwards, radar useless, clearly the server having the issue, and like the others here never used to be a problem, I've hosted since 2015.

 

Is this the right thread or is there another critical issue (that I'm not experiencing as 80-120FPS in NTTR is beautiful?

[SnorreSelmer]

dxdiag attached.

 

I flew the same mission with Entropy and Looney

DxDiag.txt

Edited September 4, 2017 by SnorreSelmer

[NineLine]

As I said in me previous posts, its from our end less about individual performance, but rather about a problem we're having in multiplayer. FPS was fine, we had net code issues like desync and warping.

 

Anyway, as requested, here is my dxdiag.

BIGNEWY, I hope you were not offended by my post Sir, we greatly appreciate your help, really and all the time you spend in an effort to help.

 

No doubt there is a bigger issue here, a report about units, especially WWII units was made some time ago already. Obviously there is an issue here, but most data from below spec computers isnt going to do us much good...

[Looney]

For clarification, I did not see many stutters. I did get the occasional frame-load freeze but nothing major that would prevent me from flying. I'll upgrade in due time before 2.5 hits.

 

I flew the Mi-8 in that mission, so low and slow. What struck me was the warping and random moving around of all the units when connected to the server, be it player, aircraft, ground or AI. I never experienced anything like that.

[BIGNEWY]

We have a bit of a mixture of bugs in this thread.

 

The main problem is related to SAM units which is giving a heavy fps hit currently, then we have the general stutters that some users see made worse by the fps issue, and also MP issues like the warping.

 

ED is aware of all of them, and we are testing internal builds as adjustments are made.

[Wingthor]

That is good news, Bignewy. If you and ED want mass testing let me know. Maybe TAW can contribute in such matter. We all want this issue solved as soon as possible :)

 

Again, most of luck from me!

[FlightControl]

We have a bit of a mixture of bugs in this thread.

 

The main problem is related to SAM units which is giving a heavy fps hit currently, then we have the general stutters that some users see made worse by the fps issue, and also MP issues like the warping.

 

ED is aware of all of them, and we are testing internal builds as adjustments are made.

 

Indeed. Let us clarify a bit further what the issue really is, and what the (possible) symptoms are, and the possible causes, because indeed there are different levels of understanding.

 

The problem is binary, true or false. It works or it does not.

 

The problems started to appear since the last patch upgrade from DCS 1.5.6 to DCS 1.5.7. Something has been added or changed, that causes a severe stutter in existing missions.

 

1. Let us take first the clarification in Single Player.

 

Take a mission that runs smooth in DCS 1.5.6 with lots of AI activity. That means, AI that is activated, and preferably has radar capability. With smooth we mean, no stutter (or very limited that does not hinder the simulation experience). When you run this mission in DCS 1.5.7, you'll experience stutter. Actually the mission will be unplayable. Not all missions have lots of AI, so this appears only at specific missions.

 

2. Let us move now the clarification up to Multi Player.

 

Multi Player missions that ran fine in DCS 1.5.6 will suddenly not perform at all in DCS 1.5.7. Again due to lot of AI Activity, and there must be (lots of) AI with radar capability.

You have two components, client and sever. The server runs the mission, the client connects to the server and synchronizes the mission.

The 'smoothness' experience however is something different here.

The client is doing no mission processing. The client is simply taking the synchronized inputs from the server, that is running the mission, and render the received information in the simulation. Now if the server becomes CPU overloaded (= does not perform anymore), you'll see lag, but no "stutter" on the client. The symptoms you see is "lag", but this is very subjective, and really not a measurement to judge this issue at all. Measurements need to be done on the server, in terms of CPU load and "FPS" (on the server).

 

FC

[EntropySG]

We have a bit of a mixture of bugs in this thread.

 

The main problem is related to SAM units which is giving a heavy fps hit currently, then we have the general stutters that some users see made worse by the fps issue, and also MP issues like the warping.

 

ED is aware of all of them, and we are testing internal builds as adjustments are made.

 

Thank you!!

[Wingthor]

Yes, Thank you FlighControl. Its important we stay focused and keep too topic.

 

"We are all in same boat, lets bring this ship safely the haven"

[lemoen]

Indeed. Let us clarify a bit further what the issue really is, and what the (possible) symptoms are, and the possible causes, because indeed there are different levels of understanding.

 

The problem is binary, true or false. It works or it does not.

 

The problems started to appear since the last patch upgrade from DCS 1.5.6 to DCS 1.5.7. Something has been added or changed, that causes a severe stutter in existing missions.

 

1. Let us take first the clarification in Single Player.

 

Take a mission that runs smooth in DCS 1.5.6 with lots of AI activity. That means, AI that is activated, and preferably has radar capability. With smooth we mean, no stutter (or very limited that does not hinder the simulation experience). When you run this mission in DCS 1.5.7, you'll experience stutter. Actually the mission will be unplayable. Not all missions have lots of AI, so this appears only at specific missions.

 

2. Let us move now the clarification up to Multi Player.

 

Multi Player missions that ran fine in DCS 1.5.6 will suddenly not perform at all in DCS 1.5.7. Again due to lot of AI Activity, and there must be (lots of) AI with radar capability.

You have two components, client and sever. The server runs the mission, the client connects to the server and synchronizes the mission.

The 'smoothness' experience however is something different here.

The client is doing no mission processing. The client is simply taking the synchronized inputs from the server, that is running the mission, and render the received information in the simulation. Now if the server becomes CPU overloaded (= does not perform anymore), you'll see lag, but no "stutter" on the client. The symptoms you see is "lag", but this is very subjective, and really not a measurement to judge this issue at all. Measurements need to be done on the server, in terms of CPU load and "FPS" (on the server).

 

FC

 

 

I think you're right but the "lots of AI" is relative. Look at the first mission in the SU27 Mozdok campaign or the first mission in the F15 campaign, there aren't really a lot of things going on there and yet those run terribly.

[CoBlue]

I can confirm the terrible lag issue in multiplayer for our squad.

Usually we are 20-30 players & it's been lagging for quite some time now, making squadron activity & mission building very frustrating.

[Ghostrider142]

I just read about 7 pages of this thread and it describes exactly the problems I incur on Normandy and occasionally in my favorite missions in NTTR. In a way good to know it affects 1.5.7 as well since a fix for one will be a fix for all. I haven't fired up the game in a month because of it, hope there is a fix really soon.

[Ventus_Clu]

I'll throw my 2 pence in and say I have these very same issues Minus the poor FPS (still get stutters) when I fly large community driven missions (lots of clients, lots of AI, fair amount of SAMs)

 

I know there is a lot of work going on (merge, F18 etc) but hopefully this will be fixed soon!

 

My sys specs are

I7 4960x (@4.5Ghz)

32 GB Ram

GTX 1080ti

windows 10

[NineLine]

I'll throw my 2 pence in and say I have these very same issues Minus the poor FPS (still get stutters) when I fly large community driven missions (lots of clients, lots of AI, fair amount of SAMs)

 

I know there is a lot of work going on (merge, F18 etc) but hopefully this will be fixed soon!

 

My sys specs are

I7 4960x (@4.5Ghz)

32 GB Ram

GTX 1080ti

windows 10

 

Dont worry, its being looked at for sure now. It may just take time to get it nailed down, but they will be working to get it ASAP.

[twistking]

sure, maybe it is not a new bug, but an shortcoming of the engine. Maybe i just now experienced the problem, because it was the first time, i populated an airport with more than 4 aircrafts.

 

However to me it still feels like buggy code, that puts a unnecessarily high burden on cpu.

To recap:

- Ground vehicles and "static (aircraft) objects" have only small, reasonable impact on fps

- HQ aircrafts have big fps impact even when being "uncontrolled". Impact on fps stacks up in linear (!) way for every asset on screen

- LQ (lower quality/older) aircrafts have lower, yet noticble impact on fps, that also stacks up for every asset on screen

- system ram and vram seem not to be the problem, as there is no lag or stutter when switching units

- running on "old" i7-860 @3,8ghz (limited single thread performance compared to newer cpus)

 

-> the weird thing is, that fps drops only appears when unit is on screen, while fps drop for the same aircraft as "static object" is insiginificant, maybe pointing towards some unholy relationship between physics simulation and renderer.

...

 

I did some more testing. First i lowered graphics quality to get some more headroom, but the problem still persists, so the fps still drops linearly for every aircraft in sight, only that with lower graphics i end up at a slightly higher fps. Impact is still unreasonable.

Also it seems, that aircraft that have ended their mission don't have such a big impact on cpu, while aircraft that are still "uncontrolled" do!!!

 

Aircraft that have ended their mission (pilot left or ejected) are still affected by the physics, so maybe the problem is AI-related. Still weird though, that it makes a difference wether the unit is in sight or not.

 

To replicate:

Populate airport with 20 aircrafts, parked on ramp and "uncontrolled".

Notice performance impact when aircrafts are on screen, also notice impact on cpu cycle while in f-10-map and accelerating time.

Restart, but now have aircrafts start/land or spawn them mid-air and have them land on same airport. Wait until all aircrafts have landed and are parked.

Notice performance impact with units on screen (should be lower now), also notice impact on cpu in f-10-map by accelerating time (should be less laggy now).

As mentioned multiple times, my cpu is not the fastest anymore, so on modern systems it may not be as obvious. Still it should be evident, that the loss in performance is not reasonable.

 

*steam, stable branch 1.5x