Guild Wars 2

Cactus.2710:

Zephyrus.9680:

Just priorities.

Skill lag is exclusive to t1 and t2. Culling is much more widespread.

No it isn’t. I (as well as most of my guildies) get ability lag fairly often during NA prime time and we’re in Tier 5.

Get it in T4 too. And it is far worse an issue than culling. You can play with culling. You can’t play when you can’t activate your skills.

Vena.8436:

Cactus.2710:

In any case, none of us really know whether the real limitation at ANet’s end is game engine, server hardware, or in/out bandwidth. It could be all three, but I’m putting my money on game engine being #1 and network bandwidth being #2. The reason I say that is that recoding the game engine takes LOTS of resources and is fraught with opportunity to screw something up, server hardware isn’t particularly expensive (although the software to manage if might be), and network bandwidth costs a lot of money each and every month.

I don’t see why you don’t believe the answer provided when it is logical. Its not like he’s saying the problem with skill lag is related to magical faeries or kittenroaches in the cogs. Its a bottleneck on the processor for skill lag (processing in-game events), and its bandwidth for culling (sending info about said processed in-game events… most of it, hence the culling). Makes sense… and they need to upgrade their hardware.

Culling is a bandwidth issue, it has little to no bearing on the data that was already all being processed and bottlenecked to begin with (the game did not not process data it wasn’t sending… it just wasn’t sending it :: your character is still going to die to culled player X spamming his aoe on your culled body without the two of you ever seeing one another). Removing it will certainly add another layer of stress on the servers but it will be much more an issue of bandwidth than an issue of stressing the server -> all the data is already processed or trying to be processed (before bottlenecked into lag), sending it isn’t going to blow out the wall sockets.

I don’t doubt their engine is poorly optimized but I’m pretty sure they have a solid flow of income from the store to fund servers/bandwidth. If they can appease the WvW crowd with some good developments, and throw some fun stuff into the shop for us too… who knows, maybe they’d get even more money.

We agree exactly on the cause of the problems and what needs to be done to fix it. Where we apparently disagree is on how we interpret what Habib said. He stated that the factors required to fix culling were different than the factors required to fix lag. His exact quote was:

“They are fundamentally different problems that will need different solutions.”

In my opinion, that was pure deflection. Culling would never have been necessary if the game engine and server-side network bandwidth had been sufficient to handle the traffic. Five months ago Habib gave this assessment of culling.

https://forum-en.gw2archive.eu/forum/pvp/wuvwuv/The-real-problem-here-is-invisible-enemies-Give-their-algorithms-time-to-match-servers-properly/page/4#post356817

In my opinion, that post was candid where his recent one was not. It wasn’t long ago that Habib posted a promise (https://forum-en.gw2archive.eu/forum/pvp/wuvwuv/Update-on-Culling/first) that Anet would buy whatever server hardware and bandwidth they needed to properly run the game, but it seems that really isn’t the case. As we both agree, lag is caused by server-side bottlenecks so I totally do not understand how Habib can now claim that culling and lag are independent issues when his previous posts have explained that they are not.

Cactus.2710:

As we both agree, lag is caused by server-side bottlenecks so I totally do not understand how Habib can now claim that culling and lag are independent issues when his previous posts have explained that they are not.

Ultimately, I do think we are reading this differently, perhaps specifically on where one puts emphasis and where one reads what is being presented in layman terms: ie. the word “fundamental”.

I do not see the issue in his general claim when they are different problems in terms of what needs to be changed (ie. where the bottleneck occurs): bandwidth (the normally understood use of bandwidth) vs. throughput (cpu bottleneck) → (loosely) connection vs. hardware. They both tie into engine efficiencies (processing data and how well compressed said data is for subsequent trafficking) but fixing the problems requires different approaches. Yes, at the very root, they are an extension of engine optimization but they lead to different issues that don’t in particular impact one another (and I feel like focusing on the word “fundamentally” outside of layman usage is playing at semantics). Likely, the reason culling is the first to be addressed is because outside of re-working engine controls and giving client side power over rendering, it is an “easier” (big quotation marks on that word) problem to solve than changing the hardware.

If they cannot figure out how to ease the load on the servers from engine optimization (which they may well be doing and testing, I cannot say) they would need likely need an extensive down time to replace the servers. (I don’t think a gradual shutdown/shift of operations would work well in this case because the WvW servers are already taxed and resource shifts just to keep the game running while servers are replaced, is likely not going to work out or the game will simply be unplayable.) Such a change, however, is likely costly and they need to take these steps one at a time.

If he wanted to be disingenuous or hide the facts, there are far better ways of addressing us than to actually even admit that there is a problem and that it needs to be addressed, or that it differs from the culling problem.

I’m pretty sure that like most developers he wasn’t thinking at that wide of a grain. Sure both problems are caused by insufficient server side resources. That covers a wide variety of potential areas. Certainly bandwidth/CPU/Memory/Disk are all server side resources that could potentially lead to problems. He also clearly stated the bandwidth was only part of the reason for culling in that same post. Which is why we are being given options around rendering.

I hate to use analogies on the internet, since there’s always someone out there who is can only think literally and points out irrelevancies in how the analogy doesn’t apply… But I will move forward.

There could be two sets of problems in the steering assembly of a car. And as a user, we basically just car that the cars move the right direction when we turn. But a problem with the electronics and a problem with the suspension are fundamentally different systems, even if they both are problems.

Going back to the details of this case:

Culling was not a problem caused by bandwidth. Culling was a solution implemented to keep bandwidth usage lower. So now they’re removing the solution and ensuring that they are purchasing enough bandwidth to cover the issue. So the changes that will be in place are:

1. Remove code that implemented culling (in the “objects in range sub-system”)
2. Add additional bandwidth
3. Provide the client application a way to manage the additional models without flooding the GPU

For skill lag, the problem may be CPU binding based on an inefficient algorithm in the damage allocation subsystem. So the fixes may be:

1. Fix the inefficient algorithm.
2. Add extra CPU.

It’s also possible the the problem is resource contention against something that isn’t hardware. The perf problem I’ve been wrestling with for the past few months at work has been realated to database locking and deadlocks. The user was experiencing something that would have felt similar to skill lag. However all of the hardware had ample resources to spare. So for this, the solution would be:

1. Restructure resource allocation scheme (locking framework) to eliminate contention problems.

Habib Loew.6239:

I should clarify here again that the culling fix will have minimal impact on skill lag. Skill lag happens because the server too busy to process everything that’s happening (e.g. it’s at 100% CPU). The removal of culling addresses issues experienced by the client but those changes don’t really impact the skill lag situation. They are fundamentally different problems that will need different solutions.

Skill lag is a much more crippling issue than culling.

Especially since you guys continue to promote the zerg mentality with every change you make in the game. Latest news to support zerging: All kills/captures/etc. give the same amount of World XP regardless of how many people were involved.

The correct way to give out World XP would be some kind of curve which benefits smaller groups and punishes people for zerging in 50+ blobs. Stonemist and Keep takes can continue to work on the old principle, then you do not punish people for using large numbers to take such objectives.

Come on, Habib. Program GW2 to run on a massively parallel system and be supercool. Individual CPUs won’t matter when all the CPUs in a single datacenter work together.

Cactus.2710:

Zephyrus.9680:

Just priorities.

Skill lag is exclusive to t1 and t2. Culling is much more widespread.

No it isn’t. I (as well as most of my guildies) get ability lag fairly often during NA prime time and we’re in Tier 5.

Well that’s interesting. If that’s true I cautiously retract my statement and am curious if any other lower tiers experience this.

In the 100s of hours and 100s of large battles I’ve played in WvW in T3 and T4 NA prime time I’ve only noticed skill lag 2 or 3 times and it was almost insignificant.

My PC was mid range several years ago and runs big fights with no problems so it’s not that. I would guess it’s either server or geography (location relative to servers). I’m on US East coast Maguuma server. Maybe your server is sharing resources with a T1 or T2 server while we’re not.

Zephyrus, I experienced this terribly between 5:30 and 7:20 PM EST yesterday, but not again when I logged in at 8:45-10:PM EST. ON maguuma as well.

At one point, I was moving around in the world and most of the other players seemed stuck, except for 2 who were moving in a straight line. It was quite surreal and I didn’t pick up on what was going on for a bit.

Oh… to clarify, the worst skill lag I was seeing was not ikittenerg situation at all.

Stop the bullkitten new culling system and change back to december.

JaironKalach.4938:

Oh… to clarify, the worst skill lag I was seeing was not ikittenerg situation at all.

That sounds like a latency issue… unless it was everyone, then there was a big fight somewhere (though be aware that WvW tiers currently share server resources over what I believe are two clusters (Borlis Pass and Yak’s Bend?)). It may well not even have been in your tier but on a map that was sharing resources with your map and 80×3 people all found themselves in Garrison.

The simplest thing they can do is to just put T1/T2 of EU/NA on in their own separate server configuration (per region, naturally), and split the remaining tiers as they have been doing. T1/T2 are the biggest hogs and their current re-allocation method just isn’t fast enough to keep up with the sudden demand spikes. Put those tiers in their own resource rich environment and many of the problems may well solve themselves.

Well I can say I have seen it in tier 7. Use to be only in sm in the lords room. The other day I got it when 2 fair sized zergs collided at umber of all places. Yes some how we lagged out umber.

And I have to agree skill lag to me is worse then culling. Culling affected everyone. Sure I can’t see them they can’t see me, I got smashed this time, next time it will be them.

But lag doesn’t affect everyone the same. On my ele skill lag all but cripples me, can’t change attunements dog crap auto attack. Running a sword on my thief auto attack is near my strongest attack. I’m happy if everyone is stuck in auto.

Cactus.2710:

Skill lag is what happens when the server cannot respond to what the client is asking it to do because the server and the server side network are overloaded … i.e., trying to do too many things in a short period of time. Skill lag is NOT … I repeat, NOT … a client side CPU issue. My computer is a water cooled 3770K/GTX680/SSD/16gig-1866MHz box running Win7-64 and I can get ability lag even in Tier 5. Wanna know how? ANet uses a merged server system to share server capacity for WvW (if you search about you’ll find a video where they explain that). When a big fight breaks out in T1 it can take some time for the merged servers to dynamically redistribute capacity to compensate, and when it does that there is a ripple effect elsewhere.

It’s pure BS when Habib says that skill lag and culling are independent. Culling was put in place to minimize skill lag in high demand situations (lots of stuff for the servers to process). It IS true that culling can be turned off like a switch (where skill lag cannot) and it IS true that client side rendering can be improved via the measures that ANet has described so that when culling is removed toasters can still give decent results, but the root cause of skill lag is lack of server side data handling capacity … and that’s the real reason culling was instituted in the first place. They are NOT totally separate topics.

I did not mean to say skill lag was because of a client CPU overload, and neither did Habib. I meant skill lag was a server CPU load issue, exactly like Habib said. Culling was meant to solve a server network bandwidth issue, not a server computation bandwidth issue. So yes, they absolutely require fundamentally different solutions. You can throw more bandwidth to solve the network capacity essentially without a problem, but you can’t necessarily increase server compute capacity without a problem.

I really don’t understand why people just like to assume the devs are all malicious beings that lie at will just to sound better. Really? I’m pretty sure they know more about the internal workings of the game than you do. But maybe I’m wrong.

7 months I’ve been playing the game, I never really experienced “unplayable” lag until recently. In WvW, Even if there weren’t fights in my field of vision, players would just walk around and “teleport” and my FPS was around 65-70, and when there are mid-sized skirmishes between 10-20 players on-screen my FPS goes down to around 50ish, and big zergs ball battles FPS goes down to 25-30, my FPS never went below 20 ever. System is fine. Given this performance, lag is insane, people are teleporting everywhere…my skills aren’t executing…when I hit something I’m not even sure if I’m damaging it because it just dies with just a few numbers popping out of its head. What’s worse is even if there are no fights at all happening on-screen with me having 60+ FPS, these things still happen, people running and teleport, skills not executing even simple auto-attacks takes like 2-3 seconds to occur, it’s disheartening. First thing I did was to call my ISP, just verified that there’s no maintenance or technical issue going on, my connection is fine as it was 7 months ago. In short, I’ve been playing the game for 7 months with the same system (a decent one as I’m also using it to play other more demanding PC games with no problems) and same ISP never had this problem until recently. So now I know that the cause is not coming from my side. Can anyone explain in more understandable terms how this LAG is being caused from their servers? I’m not IT savvy especially on servers, just the process of elimination led me to believe that the problem is not coming from my side. How will this LAG be fixed if at all? This is a very crucial point for me if I will continue playing the game, I like the game very much, but it’s just frustrating and stressing dealing with this.

EDIT: I’m from South East Asia, I know this is a big factor in latency, but like I said, been playing for 7 months with minimal issues with lag, and I also asked my ISP if there were any technical issues that may cause latency problem from SEA to NA,Texas to cause a negative effect, they said none from their side, in fact, they informed me that they recently boosted their speed (which is somewhat true as I use Skype a lot and felt the increase in quality there), so they were very emphatic that it wasn’t on their side. If it was on their side, then I should have had this issue from the beginning. That’s not the case, something happened from ANET’s side midway that started causing this. Thanks.

the lag is unplayable, haven’t logged unless we are GvG’ing.
You can’t do anything when the lag hits, you can’t dodge roll, you can’t heal, you can’t even counter the zerg running right over you with sanctuary or line of warding CUS YOU CANT DO kitten.

Why teh kitten do I want to see enemies when a new issue is obviously worse YOU CANT FIGHT BACK.

Just go back to the original culling

PredatoR.5247:

Just go back to the original culling

This is unrelated to culling (as you would quickly learn if you read the handful of large posts in this thread).

Vena.8436:

PredatoR.5247:

Just go back to the original culling

This is unrelated to culling (as you would quickly learn if you read the handful of large posts in this thread).

I don’t believe that. Because I don’t believe in coincidences. The serious crippling lag started recently.

Johje Holan.4607:

I don’t believe that. Because I don’t believe in coincidences. The serious crippling lag started recently.

The problems are caused by two separate things, again this has been elaborated upon by multiple people here, but you’re free to believe that Habib is lying to you for kittens and giggles.

Not to mention that, on the majority, the issues of these past few weeks are a bit far removed from two months ago when affinity culling was added. Its a matter of coincidence if they actually happened together, not close to or over a month apart.

Ok, well let me put it this way. Something they did recently caused it.

Johje Holan.4607:

Ok, well let me put it this way. Something they did recently caused it.

Oh no doubt, and it likely came in the Feb patch. It could be anything, though, for example: a change to skill tracking & calculating (to make sure its done more properly) that, on a glance, seems harmless but when multiplied 100x, starts to hammer the system. Or, since this is programming, someone forgot to carry the 1 somewhere. :p

But as it stands, the statements of re-rolling culling are uninformed and (almost assuredly) wrong.

Skill lag is BAD!

Anet, Please fix ASAP.

So, the lagging problem is on my CPU or on arenanet servers?

Because I get INSANE lag when playing WvW (and I didn’t get this lag a few months ago), and I play every other game on max and it’s fine… even on low graphic it lags

Game is broken and unplayable due to skill lag. I hit a skill and it goes off 5-10 seconds later!…. It was not like this before, and I have been playing since launch. Please fix your game! Thanks!

Really?! So we will still have to deal with skill lagg that makes you unable to move for 10 seconds, doesn’t allow you to heal, doesn’t allow to fight…? How in any way are we supposed to play a game like that?
Does anet really think it is better to get completely rampaged WHILE seeing who’s doing it, rather than getting completely rampaged WHILE not seeing who’s doing it?! What kind of logic is this?

I’d love to know what is causing the rise in skill delay… the wvw update is going to be for nothing if it remains as bad as it is, playing in t1 in eu.. any fight around garrision is just a nice 30 second delay to any skills used, it is a joke.

Habib Loew.6239:

The low-res models all use standard skins and are persisted in memory so they only need to be loaded once. By setting the WvW Character Limit and WvW Character Quality correctly it is possible to have all player characters (except yours) displayed using the low-res models (and thus the standard skins).

Can anyone tell me where the WvW Character Limit and WvW Character Quality options really are? I was staring at the options menu and I could not see anything related to WvW settings.

Lexvia.4087:

Habib Loew.6239:

The low-res models all use standard skins and are persisted in memory so they only need to be loaded once. By setting the WvW Character Limit and WvW Character Quality correctly it is possible to have all player characters (except yours) displayed using the low-res models (and thus the standard skins).

Can anyone tell me where the WvW Character Limit and WvW Character Quality options really are? I was staring at the options menu and I could not see anything related to WvW settings.

They are not patched in yet. He is talking about an upcoming patch.

Vena.8436:

Johje Holan.4607:

I don’t believe that. Because I don’t believe in coincidences. The serious crippling lag started recently.

The problems are caused by two separate things, again this has been elaborated upon by multiple people here, but you’re free to believe that Habib is lying to you for kittens and giggles.

Not to mention that, on the majority, the issues of these past few weeks are a bit far removed from two months ago when affinity culling was added. Its a matter of coincidence if they actually happened together, not close to or over a month apart.

You’ve said that several times and you’ve been wrong each time. Culling and ability lag are not unrelated, and in fact culling was originally instituted to minimize lag … although it is true that there are things that affect client side rendering (an aspect of culling) that do not strongly affect server side lag.

Here’s a test for you. Please make a list of the things that affect lag, and then make a separate list of the the things that affect culling.

natanfk.1649:

So, the lagging problem is on my CPU or on arenanet servers?

Because I get INSANE lag when playing WvW (and I didn’t get this lag a few months ago), and I play every other game on max and it’s fine… even on low graphic it lags

Ability lag (aka skill lag) is strictly a server-side problem caused by more load than the engine/servers/network can handle.

Culling is BOTH a server-side (load reducer) and a client-side (toaster rendering) issue, but without culling the load is heavier on the engine/servers/network than it would be with culling.

ANet believes that they have fixed the client-side aspects of culling to the point that they are willing to remove it in the next patch, but they have said that lag will not improve. They have coded some things differently and given players some rendering options so that a wider range of computers can function without culling, but they haven’t done anything to upgrade anything server-side. Load mismatches are still going to give us lots of ability lag.

It went away for a week, and now it’s back. One guildmate whom it had been affecting was not having issues today, but I was having issues the moment I stepped into WvW. Worsened as I approached the zerg.

morrolan.9608:

Habib Loew.6239:

I should clarify here again that the culling fix will have minimal impact on skill lag. Skill lag happens because the server too busy to process everything that’s happening (e.g. it’s at 100% CPU). The removal of culling addresses issues experienced by the client but those changes don’t really impact the skill lag situation. They are fundamentally different problems that will need different solutions.

So why has skill lag increased?

my guess is because either:
a. they downgraded the number of servers in each cluster due to cost
b. they put more workload on each server cluster due to cost
c. a + b

Habib Loew.6239:

I should clarify here again that the culling fix will have minimal impact on skill lag.

Thanks for the info Habib. Like others, I’d like to express my desire to see skill lag addressed.

I have also been seeing some of the worst skill lag I have ever seen in the game. My previous gameplay in WvW has never seen anything as bad as what I am now seeing. The only time I ever saw anything close was the Karka event (and that was PvE so the issues are different).

I’m glad to know I’m not the only one seeing this issue of crippling skill lag. At least that will insure it will get attention. Frankly, I would welcome the return of culling if it would make WvW playable for me again but that’s not likely since the patch seems to have worked for other players.

Cactus.2710:

Habib Loew.6239:

I should clarify here again that the culling fix will have minimal impact on skill lag. Skill lag happens because the server is too busy to process everything that’s happening (e.g. it’s at 100% CPU). The removal of culling addresses issues experienced by the client but those changes don’t really impact the skill lag situation. They are fundamentally different problems that will need different solutions.

Sorry, but this just doesn’t sound right. One of your earlier posts even states that culling was instituted to avoid server-side overload, and that would make sense. I don’t believe for a second that lag and culling are independent issues. Culling was only partially a client side crutch for underpowered computers … it was primarily put in place to regulate how much data the server had to process and force feed through the pipeline. The server doesn’t even really know whether something has rendered yet client side, but it does know when it is getting bogged down trying to process too much data internally.

If lag still exists after you drop culling it’s either because your game engine is too limited and you don’t have the resources to fix it, or you backed off on your earlier promise to buy whatever bandwidth you needed to fix the problem.

In any case, it’s pretty clear that we’re being BS’d here.

culling was there because their internet connection was ‘excrement’ and unable to upload enough data to all clients. it seems they upgraded their analogue modem to adsl (ok, bit of a joke but you should get the idea) and were able to switch of culling as a result.

skill lag is there because their servers are ‘excrement’ too. now they need to upgrade their C64 to something more recent and skill lag will be a thing of the past too.

Culling not affecting skill lag makes perfect sense if you have a bit of game programming experience. They’re two different problems.

Culling is – how can we not overwhelm our clients’ computers with drawing too much stuff? They tried limiting it on the server side and it didn’t work out too well because it made WvW feel less epic and you were constantly getting killed by invisible stuff. So now they moved the limiting to the client side and let you sort out the details. Bandwidth was never an issue – they have the same servers now as they did three months ago, after all, and in terms of showing me everyone’s locations they are doing just fine.

Skill lag is different. What is likely happening is that when you use a skill, it goes into a queue on the server. Every game tick, the server goes through the entire queue and processes all the different effects, damages and heals and applies conditions to the right people, and so on. Skill lag happens when this queue gets too full. Since you can press the button for a skill and nothing happens (as opposed to it just taking a while to happen) it’s likely that if the queue gets too full they just throw away the overflowing ones because they don’t want to have to drop frames on the server.

So you get skill lag from too many people spamming skills no matter where they are on the map, but in big fights you still see their positions and health and such just fine.

(This is all speculation, by the way, don’t start complaining that Anet needs to make their queues bigger or something dumb.)

Johje Holan.4607:

Cactus.2710:

Zephyrus.9680:

Just priorities.

Skill lag is exclusive to t1 and t2. Culling is much more widespread.

No it isn’t. I (as well as most of my guildies) get ability lag fairly often during NA prime time and we’re in Tier 5.

Get it in T4 too. And it is far worse an issue than culling. You can play with culling. You can’t play when you can’t activate your skills.

good to know, at least I dont have to waste gem to transfer hoping skill lag would go away. Anet dev team just continue to show the lack of effort and planning when they designed the WvW environment, it just seems like they didnt do any QA on that part of the game.

Buy better hardware Anet!

Skill lag is different. What is likely happening is that when you use a skill, it goes into a queue on the server. Every game tick, the server goes through the entire queue and processes all the different effects, damages and heals and applies conditions to the right people, and so on. Skill lag happens when this queue gets too full. Since you can press the button for a skill and nothing happens (as opposed to it just taking a while to happen) it’s likely that if the queue gets too full they just throw away the overflowing ones because they don’t want to have to drop frames on the server.
So you get skill lag from too many people spamming skills no matter where they are on the map, but in big fights you still see their positions and health and such just fine.

This. People need to stop equating skill lag to lack of culling.

Also better hardware may not solve it. WoW has skill lag during patch weeks and big events, yet you’d think the millions Blizzard/Activision have they’d be able to solve it with better hardware.

It’s a combination of software design and hardware.

Cactus.2710:

Ability lag (aka skill lag) is strictly a server-side problem caused by more load than the engine/servers/network can handle.

while generally true, other things CAN affect skill lag like a ISP, traffic shaping by ISP, overloaded internet connections between 2 different ISP that rout your packats via alternating routes due to load balancing etc…

i used to play an other mmo where the servers were in france and i’m in Belgium. at times when there was cripeling lag, i saw that my packets didn’t go from my place to brussel to paris to the server but that one packet went from home to brussel to frankfurt to paris, next went from home to brussel to london to paris, the one after that went from home to brussel to amsterkitteno lyon to paris, … this had as an effect that the messages i typed sometimes didn’t come on screen in the order i typed them, people rubberbanded allover the place and there was horrible skill lag.

…

just saying that there are other sources of skill lag than just sub-par servers even if it might not apply to the current situation.

lettucemode.3789:

Culling is – how can we not overwhelm our clients’ computers with drawing too much stuff?

that would be graphic culling which is something different (confusing isn’t it ?) graphic culling is the client that doesn’t calculate or render things that are behind something that is closer and obstructs the view. there is simply no point in calculating and rendering players/mobs/objects behind, say, a wall that would completely block the view.

the culling used by Anet was something where the server had a maximum number of players that the server would tell each individual client about to cut down on their outbound network traffic.

lettucemode.3789:

Bandwidth was never an issue – they have the same servers now as they did three months ago, after all, and in terms of showing me everyone’s locations they are doing just fine.

i think you are wrong again. Anet’s outbound/upload capacity was the main issue that was responsible for them using culling. as far as i can tell they were able to disable culling after they upgraded their internet connection.

muylaetrix.2096:

lettucemode.3789:

Culling is – how can we not overwhelm our clients’ computers with drawing too much stuff?

that would be graphic culling which is something different (confusing isn’t it ?)

No, not really, I’ve done some graphics work and so I know what the term means :P Whether I meant the technical term or the Arenanet-WvW term, either way you are drawing less stuff, I chose simple (admittedly ambiguous) words because I’m lazy.

muylaetrix.2096:

Anet’s outbound/upload capacity was the main issue that was responsible for them using culling. as far as i can tell they were able to disable culling after they upgraded their internet connection.

Source for this? Otherwise I’m sticking with my theory. Habib made a post about the changes they were going to make to fix it a few months ago, and the only network change he mentioned was getting the higher-ups to approve using more. They were never anywhere near the cap.

How long will lag solution to the problem? As much as the players get tired of playing in such business conditions? Everything was ok until the entry last patch (26.03). I think that this is not the best solution.

A proposed solution would be to limit the map size.

The current map is large with a large overhead.

Might want to consider breaking the borderlands apart into separate games entirely.

Maybe keep the Eternal Battlegrounds as weekly, but making the others as daily with it’s own reward system.

This would solve a lot of issues but would require making smaller maps.

lettucemode.3789:

No, not really, I’ve done some graphics work and so I know what the term means :P Whether I meant the technical term or the Arenanet-WvW term, either way you are drawing less stuff, I chose simple (admittedly ambiguous) words because I’m lazy.

server side culling (the server only tells you about xx objects because if we try send all relevant info to everyone, we might not have enough bandwidth to send all data and some clients might get no data at all) and client culling (the client doesn’t render stuff which you can’t see) are two different things.

lettucemode.3789:

Source for this? Otherwise I’m sticking with my theory. Habib made a post about the changes they were going to make to fix it a few months ago, and the only network change he mentioned was getting the higher-ups to approve using more. They were never anywhere near the cap.

i think we are speaking about the same post. i read it as “we don’t have enough upload bandwidth. to send all data to everyone we need more bandwidth. this will cost money as we need to upgrade our connection”

the culling used by anet was a cost saving measure to compensate for their insufficient bandwidth.

….

ofc, i could be wrong :p

skill lag kicks in now on eternal if 2 maxed zergs(2 worlds hit each other)

Possible way to get more ppl on the maps and get rid of skill lag:
Multi-Server running the WvW maps(overflow servers could be used for this)

The Server Cluster used Lan the exange data.
The Servers running the same WvW map run on the same data set.
Each Server handles only a limited count on Players(ca 150?)
Each server uses Lan to exange Calculated Player data(Sync Point)

Each server must only handle fixed(maxed) Player amount of calculation and gets the other calculated Player data via fast LAN. it would only use extra servers if they are realy needed(like overflow) or even better 1 server could handle 3-4 zones with a few ppl.

ServerInBoxThread()
->ReciveData for active Maps on this Server from other Servers
->ReciveRawData from players

ServerMainThread()
->SyncPoint -> Copy RecivedData(pointer switch)
->Calculate RawData from players on this server(got the other calculated player data from other server)
->Sends Calculate RawData to other servers who need it(fast via lan)
->Sends the players the needed data

so there will be a little delay for player vs player from other server(1sim frame min),but since the servers are able to exange data fast it should be not a prob. Even possible to dynamic cluster players who interact a lot on the same server

just an idea;)

muylaetrix.2096:

lettucemode.3789:

No, not really, I’ve done some graphics work and so I know what the term means :P Whether I meant the technical term or the Arenanet-WvW term, either way you are drawing less stuff, I chose simple (admittedly ambiguous) words because I’m lazy.

server side culling (the server only tells you about xx objects because if we try send all relevant info to everyone, we might not have enough bandwidth to send all data and some clients might get no data at all) and client culling (the client doesn’t render stuff which you can’t see) are two different things.

lettucemode.3789:

Source for this? Otherwise I’m sticking with my theory. Habib made a post about the changes they were going to make to fix it a few months ago, and the only network change he mentioned was getting the higher-ups to approve using more. They were never anywhere near the cap.

i think we are speaking about the same post. i read it as “we don’t have enough upload bandwidth. to send all data to everyone we need more bandwidth. this will cost money as we need to upgrade our connection”

the culling used by anet was a cost saving measure to compensate for their insufficient bandwidth.

….

ofc, i could be wrong :p

You are not wrong … Lettuce is. Habib’s post on culling clearly listed server-side network bandwidth as being one of the culprits, albeit one of the easiest to fix.

I was in WvW with guildies last night and it appears the culling issue was addressed, abeit a bit differently. We saw hundreds of red names running around (well it seemed like hundreds) but no bodies attached to them. ehhehehehe

But I did finally catch a screen shot of our small group and, guess what?, we have the same problem in a small group too. I think in this group we captured a point with maybe 20 or so guildies, plus a couple of other guild groups and all I saw closest to me were names, no bodies.

Unintended consequences I think.

Cactus.2710:

muylaetrix.2096:

lettucemode.3789:

No, not really, I’ve done some graphics work and so I know what the term means :P Whether I meant the technical term or the Arenanet-WvW term, either way you are drawing less stuff, I chose simple (admittedly ambiguous) words because I’m lazy.

server side culling (the server only tells you about xx objects because if we try send all relevant info to everyone, we might not have enough bandwidth to send all data and some clients might get no data at all) and client culling (the client doesn’t render stuff which you can’t see) are two different things.

lettucemode.3789:

Source for this? Otherwise I’m sticking with my theory. Habib made a post about the changes they were going to make to fix it a few months ago, and the only network change he mentioned was getting the higher-ups to approve using more. They were never anywhere near the cap.

i think we are speaking about the same post. i read it as “we don’t have enough upload bandwidth. to send all data to everyone we need more bandwidth. this will cost money as we need to upgrade our connection”

the culling used by anet was a cost saving measure to compensate for their insufficient bandwidth.

….

ofc, i could be wrong :p

You are not wrong … Lettuce is. Habib’s post on culling clearly listed server-side network bandwidth as being one of the culprits, albeit one of the easiest to fix.

I think they are making money off this game, but what I do not understand is why they are not upgrading their equipment and bandwidth.