These forums have been archived and are now read-only.

The new forums are live and can be found at https://forums.eveonline.com/

EVE General Discussion

 
  • Topic is locked indefinitely.
 

Time Dilation feedback

First post First post
Author
Nova Fox
Novafox Shipyards
#121 - 2012-01-23 19:47:05 UTC
Well here is an update I found eve searching around

CCP is already considering DynR (Dynamic Remaps is this the offical name for it now?) from the devs so far they commented that shuffeling server loads with curent technology is 'destructive' or in lamen terms forces players to log offs, and they're currenlty working on a way to not cause that so that the load shuffel can happen without the disconnects.

Either way should be interesting I rather be a small lag spike if it means nobody gets booted off then tidi dials a bit down as things speed back up with the load shuffeled off.

Dust 514's CPM 1 Iron Wolf Saber Eve mail me about Dust 514 issues.

Ranger 1
Ranger Corp
Vae. Victis.
#122 - 2012-01-23 20:24:26 UTC  |  Edited by: Ranger 1
Quote:
experienced TiDi over the weekend and was in a system, maybe more than 1, and it was just out gang of 9 people. It sucked


Again, you were likely on a node that a large battle was happening in. The battle may have been no where near you on the map, yet you could have been on the same node. If you think about it, that is actually wise programming.

Would you prefer a slow down, or for the game to inexplicably (from your point of view) crash and not let you back in for awhile...

View the latest EVE Online developments and other game related news and gameplay by visiting Ranger 1 Presents: Virtual Realms.

Rasz Lin
#123 - 2012-01-23 20:52:42 UTC  |  Edited by: Rasz Lin
-Every SOL is running IN ONE THREAD ON ONE CORE (usually more SOLs on one core, unless its reinforced system).

-You cant partition load on few cores, you cant even divide SOL into grids so if there are 10 battles in one SOL they ALL run in ONE THREAD ON ONE CORE.
-There is no mechanism for live migration, CCP cant just freeze SOL and migrate it to another core :/ It should be easy now that they figured out time dilation - freezing is just slowing time to infinity :)


Sarina Rhoda wrote:
>Correct me if i'm wrong but time dilation wasn't created to fix lag..... Time dilation was introduced to make fights fairer under
>extreme lag conditions. Previously the server would just randomly accept some peoples commands while completely
>ignoring other peoples. This meant some people were able to fight as if there was nothing wrong whilst other people were
>completely locked out.

>Time dilation was simply a method of slowing everything down so that the server has a fair chance and processing
>everyone's commands.


not server, one THREAD bound to ONE cpu core :((( Instead of letting server compute more of the battle they are throttling the battle to "fit" on one CPU.


Zleon Leigh wrote:
>TiDi is implemented because there does not exist a hardware solution that will handle the computation load and also the
>time of flight problem across the net.


Not true, There are FPS games (for example Mount And Blade Warband) currently that run smooth with >200 players bashing each other with melee/thrown/shooting weapons in the face. With server tick time 0.1sec (means steps per second stuff is updated server side). Eve has server tick time 1sec so by extrapolating those FPS games could handle 2000 FPS players at 1sec tick, all unning around, changing directions, waving melee weapons, hitting each other, shooting at each other, all on one map.
EVE doesnt really have directions, when it comes to calculating tracking it doesnt matter where your ship is pointing. EVE is EASY compared to FPS engines, all you need to worry is position and status of incoming/outgoing damage.



+There was an example of optimisation on dev blog ~year ago that showcased some piece of code that was O(n2) for every missile shot (or just object in space, dont rememer). There are still monsters like that in the code.
Stupid client side example - EVERY TIME session changes (jump/undock) your ASSET window is drawn at least 3 times.
-First it is initialized from some dataset in the memory, drawn with default alphabetical order of items.
- then it is REINITIALIZED and sorted according to your selection.
- then it is AGAIN reinitialized and drawn with remembered positions and station item trees open/closed.

Whole GUI was refactored, yet it is still inefficient and is redrawing parts of the screen that dont change at all, it seems nothing is being kept in buffers, compositioning just doesnt exist. You open Market screen and it will call GPU EVERY frame drawing every single font one by one instead of generating that window ONCE into a bitmap buffer and just using it untill something changes. Even Android finally uses proper compositioning and only redraws part of the screen that change - GUI in your phone is smarter than GUI in this game :)




They are working on those things, just toooo slow. There was a time Market was running on same cpu core as the rest of the system, afaik now there are dedicated market nodes. Next step would be enabling separate grids to run on different cores and live migration between CPU cores, preferably over the network so they could migrate them between servers and not only between cpus on same machine.
Of course best case scenario would be rewritting "battle code" into something MULTITHREADED so one big battle can use more than one CPU on a Node (nodes are multi CPU, CPUs have HT) dividing load. Calculating distances, misses, damages, movements, it could all be multithreaded and done at the same time instead of doing it in sequence.



In simple terms tidi SUCKS MONKEY BALLS and is a clear evidence you rather ducttape crap together like minmatar battleship than fix fundamental problems. DROP python server side already.
Rasz Lin
#124 - 2012-01-23 21:05:21 UTC
Michael Banki wrote:
I was in a big fight, there was TIDI but lag was BAD. Might be that tidi was juming on and off alot or that I had lag on my client.

So could tidi change SLOWER if there are 1000 ppl in system. And another thing:

Woulndnt it be great if there was a client lag indicator?

So that we know when to get a better computer...


its simple - when your framerate drops its your computer, or more precisely its lame Python code trying to fill all the brackets info that gets displayed on your screen.
Ranger 1
Ranger Corp
Vae. Victis.
#125 - 2012-01-23 21:46:07 UTC
No, you are correct.

It would be a much better idea to shut down EVE completely for a few years while it is rewritten to function as you describe.

Except, of course, nobody but you appears to want that... and your proposed solutions are problematic at best in light of the information CCP has already provided as to the "hows' and "why's" of the EVE code base.

View the latest EVE Online developments and other game related news and gameplay by visiting Ranger 1 Presents: Virtual Realms.

Kwa Zulu
The Graduates
The Initiative.
#126 - 2012-01-24 11:16:55 UTC
CCP I just want to say Time Dilation is awesome Big smile

Large fights are possible again, EVE does has a future
Florestan Bronstein
Ministry of War
Amarr Empire
#127 - 2012-01-24 11:22:02 UTC  |  Edited by: Florestan Bronstein
Rasz Lin wrote:

Not true, There are FPS games (for example Mount And Blade Warband) currently that run smooth with >200 players bashing each other with melee/thrown/shooting weapons in the face. With server tick time 0.1sec (means steps per second stuff is updated server side). Eve has server tick time 1sec so by extrapolating those FPS games could handle 2000 FPS players at 1sec tick, all unning around, changing directions, waving melee weapons, hitting each other, shooting at each other, all on one map.


if I have to update 200 players over each other player's status every 0.1 sec how many operations per second is that?

if I have to update 2000 players over each other player's status every 1 sec how many operations per second is that?

edit: as for the actual topic - experienced TiDi when undocking in a ~50 man fleet from a station.
Cleared within 20-30s after everybody had finished undocking but still annoying.
Of course I might just not have noticed the lag created of undocking such a small fleet before and the TiDi indicator only makes it obvious but still - would never have expected that this creates issues with server load.
DarkAegix
Center for Advanced Studies
Gallente Federation
#128 - 2012-01-24 11:23:28 UTC
Lyron-Baktos wrote:
experienced TiDi over the weekend and was in a system, maybe more than 1, and it was just out gang of 9 people. It sucked

I'm sorry, but there was probably a 500+ blob a few systems away.
Without tidi you ship would have exploded in lag, and client crashed to desktop.
Tallian Saotome
Nuclear Arms Exchange Inc.
#129 - 2012-01-24 11:34:55 UTC
Florestan Bronstein wrote:
Rasz Lin wrote:

Not true, There are FPS games (for example Mount And Blade Warband) currently that run smooth with >200 players bashing each other with melee/thrown/shooting weapons in the face. With server tick time 0.1sec (means steps per second stuff is updated server side). Eve has server tick time 1sec so by extrapolating those FPS games could handle 2000 FPS players at 1sec tick, all unning around, changing directions, waving melee weapons, hitting each other, shooting at each other, all on one map.


if I have to update 200 players over each other player's status every 0.1 sec how many operations per second is that?

if I have to update 2000 players over each other player's status every 1 sec how many operations per second is that?

edit: as for the actual topic - experienced TiDi when undocking in a ~50 man fleet from a station.
Cleared within 20-30s after everybody had finished undocking but still annoying.
Of course I might just not have noticed the lag created of undocking such a small fleet before and the TiDi indicator only makes it obvious but still - would never have expected that this creates issues with server load.

Wait til your 800 man blob hits a jump gate.....


Basically, session changes lag the hell out of the system, and tidi makes it screamingly obvious.

Inappropriate signature removed, CCP Phantom.

Iniquita
Perkone
Caldari State
#130 - 2012-01-24 12:02:37 UTC
Warping seems to be an issue, TiDI warps can take forever and it seems impossible to cancel warps with a control space anymore.
Holy One
Privat Party
#131 - 2012-01-24 13:28:19 UTC
Better than black screen. Still ****.

Piloting anything that has to actually move or kite is almost impossible.

I personally would prefer hard limits and proper fights.

:)

mirel yirrin
Ore Mongers
#132 - 2012-01-24 14:03:42 UTC
Time Dilation was activated in BWF a few weeks ago. The 500 man combined fleets three way brawl that was going on at the time went for laggy as usual to completely lagless and highly responsive. The frame rate suffered a bit for me, but that is because I have everything turned up to max + sound during fleet battles, for extra immersion. Was down from usual 60FPS to about 25 - 30.

To sum;

zero module lag
zero ship command i.e warp commands, movement commands lag
maximum awesome.

Darius III for CSM! Or your money back!

Tallian Saotome
Nuclear Arms Exchange Inc.
#133 - 2012-01-24 14:06:41 UTC
Holy One wrote:
Better than black screen. Still ****.

Piloting anything that has to actually move or kite is almost impossible.

I personally would prefer hard limits and proper fights.

I had no problems... Once I realized that it was taking 10x as long to accelerate/turn/etc, just like it was taking them 10x as long to lock/fire/etc.

At least its fair.

Inappropriate signature removed, CCP Phantom.

Tres Farmer
Gallente Federation Intelligence Service
#134 - 2012-01-24 16:23:06 UTC  |  Edited by: Tres Farmer
Vegare wrote:
I participated in two ~1400 man fights last weekend and have to say: TiDi is awesome!

Everyone who was there knows fights like those would have not been possible not long ago.

To all those whiners: obviously there are lots of gameplay issues to being dilated down to 0.1 but its IS gameplay contrary to being blackscreened.

If I read all those comments correctly this will happen:

- less reinforced fleet fights (because 1400+ is possible on a node with 75 systems and the TiDi is similar to a reinforced node.. plus you can fight anytime, anywhere, no need to fill out forms and stick to it)
- more fleet fights than before .. anytime, anywhere, as it works now
- more bystanders being pulled into TiDi in other sol-threads on the same node
- no way for bystanders to avoid TiDi nodes (Are they supposed to move to enjoy TiDi = 1 ratting? Where? Is there a map option taht shows systems TiDi?)
Arrow bystanders become really, really mad


@CCP Veritas or Lemur
How do you intend to tackle this scenario?


Is there any possibility for the thread causing the whole node to incur a TiDi < 1 to be restricted to the maximum load possible so that all other nodes that are peacefull and quiet* can keep processing their stuff at TiDi = 1?
Can your technology do that?

Example of what I'm talking about:
Node_XYZ running 50 sol threads (CPU usage 30%, each thread is being monitored for it's usage, breathing room would be 70%)
49 threads handle the usual stuff with a low population.
Thread 50 is becoming overloaded and causes the CPU load to go up.
The other threads stay within their power envelope
Thread 50 should now be restricted to increase it's power envelope so that the total load of the node doesn't exceed 95%.
The other 49 threads stay at TiDi = 1 and have a 'breathing room' of 5% if any of the threads needs a bigger power envelope.

The point is.. thread 50 will have a TiDi <<< 1 anyway, but why should all the other threads on that node suffer the same fate?


Asking the same the other way round:
Do you have any metrics that would show the benefit of the whole node being at a TiDi <<< 1
versus 49 threads at TiDi = 1 and 1 thread at TiDi <<<< 1 ?


*) ...must obviously be some kind of metric to decide if a sol thread should be time dilated or not.
For example take the average population over the last 24 hours as a value and if that is being overshoot by 100% is the flag that this sol thread can become time dilated
Or another one.. all sol threads on that node together cause a load of B% with the average population of the last 24 hours, if all were to increase their population by 10%.. how much load would that cause? 1.1 x B%? Any sol thread that exceeds that power envelope (110%) is open to time dilation.
Avo Lina
State War Academy
Caldari State
#135 - 2012-01-24 21:55:40 UTC
Zleon Leigh wrote:
Jarnis McPieksu wrote:
Zleon Leigh wrote:
CCP Lemur wrote:
Zleon Leigh wrote:
So - we're seeing TiDi engage in empty systems and in Jita. Fail? Oh wait!! By engaging it in empty systems you can lower how much hardware you need to field. Guess returning nodes to vendor is next step after layoffs....


I'm not a Veritas but I assume that since some nodes host a bunch of low load systems at once TiDi would kick in for all of these when one gets a sudden spike of activity. Or would you rather have the ones just screwed without any indicators? And I'm sure we are not giving any nodes back rather getting better ones like we have always been doing in the past and will continue to be doing in the future.



I'm in a 0.0 system with 17 pilots in it, several docked in station, NO FIGHTING, peaceful as can be - TiDI kicked in. LAME.




I think you fail to understand that single physical server can run multiple solar systems.

In the case of this huge fight, reportedly the same node (server computer) ran 75 solar systems. One of them has a huge fight and all 75 get time dilated because they are all on the same server.


LOL Trust me... I understand. There is no reason for TiDi to kick in for all systems on a node - that would just be stupid. You dilate the stressed systems and fork off the low end systems to other nodes. You don't turn the clock down on the entire CPU stack.





You are so FFing dumb, you don't even realize how stupid the nonsense is you are spewing forth
LethalGeek
Sebiestor Tribe
Minmatar Republic
#136 - 2012-01-24 22:00:19 UTC
Zleon Leigh wrote:
Apparently so... I used to do CPU hardware design for supercomputers... :)

You're either lying or purposefully ignoring what people are telling you about how the current server software is written.
Othran
Route One
#137 - 2012-01-25 11:35:07 UTC
Florestan Bronstein wrote:
Of course I might just not have noticed the lag created of undocking such a small fleet before and the TiDi indicator only makes it obvious but still - would never have expected that this creates issues with server load.


Is it a fleet where individual members are likely to have a lot of bookmarks?
Tallian Saotome
Nuclear Arms Exchange Inc.
#138 - 2012-01-25 11:44:43 UTC
LethalGeek wrote:
Zleon Leigh wrote:
Apparently so... I used to do CPU hardware design for supercomputers... :)

You're either lying or purposefully ignoring what people are telling you about how the current server software is written.

Given how long its been since he posted, I'm gonna say he didn't realize it was an artificial software limitation, instead thinking we were saying it was hardware.

Seems to understand the issue now ;)

Inappropriate signature removed, CCP Phantom.

Holy One
Privat Party
#139 - 2012-01-26 13:47:43 UTC
Tallian Saotome wrote:
Holy One wrote:
Better than black screen. Still ****.

Piloting anything that has to actually move or kite is almost impossible.

I personally would prefer hard limits and proper fights.

I had no problems... Once I realized that it was taking 10x as long to accelerate/turn/etc, just like it was taking them 10x as long to lock/fire/etc.

At least its fair.


*shrug

I don't have any desire to spend 4-5 hrs doing what normally takes 2.

:)

Zleon Leigh
#140 - 2012-01-26 22:44:43 UTC
Tallian Saotome wrote:
LethalGeek wrote:
Zleon Leigh wrote:
Apparently so... I used to do CPU hardware design for supercomputers... :)

You're either lying or purposefully ignoring what people are telling you about how the current server software is written.

Given how long its been since he posted, I'm gonna say he didn't realize it was an artificial software limitation, instead thinking we were saying it was hardware.

Seems to understand the issue now ;)



Sorry - been busy.

Correct. Didn't realize that EVE software had been implemented so poorly. Thought all the 70's programmers had retired before they had a chance to work on EVE.

Incarna - Newest business example of mismanaged capital. CCP - Continuing to gank independent PI producers every day

PvP's latest  incentive program ** Unified Inventory **  'Cause you gotta kill something after trying to use it