These forums have been archived and are now read-only.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.
123Next pageLast page
 

TiDi

First post
Author
Lord Molly
The Tuskers
The Tuskers Co.
#1 - 2015-12-29 10:32:23 UTC
This video documents some real bad tidi issues

If you look closely, you might also notice some glitching, such as smart bombs doing damage in warp.......

All of this is caused by server node overload when to many people are in one system activating too many mod.


To combat this very frustrating trait of gameplay, CCP made a jolly form that you fill in and effectively tell them there is gonig to be a big fight and etc etc, so that they may reinforce a node and divert resources to said systems node and hopefully make it all ok.

Only issue is, that's not really the case is it. Have you ever been to a public event or a large pre-arranged scrap where tidi is down to 10%? where you fire a volley, then go and make dinner and come back to find you are still on the same cycle?

Id be interested to hear from CCP about how they intend to improve the cancer that is TiDi.
FarosWarrior
NED-Clan
Goonswarm Federation
#2 - 2015-12-29 10:39:43 UTC
Let's go back to the time when you had a big battle with 600 guys in local and the following would happen:

- Modules not even activating
- You being desynced for an hour without you knowing
- Warping on grid and waking up in a clone bay. An hour later...

And all the other crap that happened which are too numerous to write down. Ever heard the story about the battle in which 1 side jumped through a gate, and a 100-man fleet just woke up in clone bays because the other side had so many people on grid, that the jumping party couldn't even load grid?

TiDi was implemented for a reason, while I must admit I don't know the exact reason but I think it had something to do with giving the server more time to do what it has to do so all the bad stuff didn't happen.
Top Guac
Doomheim
#3 - 2015-12-29 10:40:39 UTC  |  Edited by: Top Guac
Aegis sov.

Has worked a lot for the most part. Tidi happens only occasionally now. Used to be weekly just 18 months ago.

Aegis Sov: less conflict, fewer fights, smaller fleets = no server load.

Back in the original devblog by Greyscale about fatigue, this was all predicted:

http://community.eveonline.com/news/dev-blogs/long-distance-travel-changes-inbound/

CCP wanted to make the game boring.
Lord Molly
The Tuskers
The Tuskers Co.
#4 - 2015-12-29 10:46:45 UTC
It just seems like the threat of tidi puts people off doing anything large scale, especially if its short notice.
Solecist Project
#5 - 2015-12-29 10:47:00 UTC
Not sure smartbombs doing damage in warp is a bug.

Depends on where that happens.


Also, shouldn't you ve called Molle? :P

That ringing in your ears you're experiencing right now is the last gasping breathe of a dying inner ear as it got thoroughly PULVERISED by the point roaring over your head at supersonic speeds. - Tippia

Gregor Parud
Imperial Academy
#6 - 2015-12-29 11:03:50 UTC
"Hilariously large fleets created to deter actual fighting cause crippling lag" and "tinfoil is just easier to use than realising you have spies amongst your ranks"
Barrogh Habalu
Imperial Shipment
Amarr Empire
#7 - 2015-12-29 11:05:51 UTC
Lord Molly wrote:
To combat this very frustrating trait of gameplay, CCP made a jolly form that you fill in and effectively tell them there is gonig to be a big fight and etc etc, so that they may reinforce a node and divert resources to said systems node and hopefully make it all ok.

Only issue is, that's not really the case is it.

Resources, even re-routed, are still finite.
Lucas Kell
Solitude Trading
S.N.O.T.
#8 - 2015-12-29 11:18:36 UTC
FarosWarrior wrote:
Let's go back to the time when you had a big battle with 600 guys in local and the following would happen:

- Modules not even activating
- You being desynced for an hour without you knowing
- Warping on grid and waking up in a clone bay. An hour later...

And all the other crap that happened which are too numerous to write down. Ever heard the story about the battle in which 1 side jumped through a gate, and a 100-man fleet just woke up in clone bays because the other side had so many people on grid, that the jumping party couldn't even load grid?

TiDi was implemented for a reason, while I must admit I don't know the exact reason but I think it had something to do with giving the server more time to do what it has to do so all the bad stuff didn't happen.
Alternatively they could just do what any normal person would do and multithread their server code. I know it's legacy code and will be a large investment, but it's already a joke that their solar system servers have to run on a single core (because back in the day, CPU power grew primarily on clock speed whereas now it grows on additional cores), and it's only going to get funnier as time goes on.

Tidi was a stopgap measure. It allowed them to give the server more time to get through the event queue so that they could still run everything on a single heavily overclocked core. At some point the promoted to "the fix" and now it's a burden. Things like BIAB help, but in the long run the solution has to be modernising their server code.

The Indecisive Noob - EVE fan blog.

Wholesale Trading - The new bulk trading mailing list.

Nevyn Auscent
Broke Sauce
#9 - 2015-12-29 11:19:50 UTC
Solecist Project wrote:
Not sure smartbombs doing damage in warp is a bug.

Depends on where that happens.


Also, shouldn't you ve called Molle? :P

Smart bombs always do damage even if you are in warp still. That's a long standing feature of smart bombs.
Remiel Pollard
Aliastra
Gallente Federation
#10 - 2015-12-29 11:22:15 UTC  |  Edited by: Remiel Pollard
Nevyn Auscent wrote:
Solecist Project wrote:
Not sure smartbombs doing damage in warp is a bug.

Depends on where that happens.


Also, shouldn't you ve called Molle? :P

Smart bombs always do damage even if you are in warp still. That's a long standing feature of smart bombs.


Yeah, this. First hand experience here, been blown up ship AND pod by smartbombs while about to come out of warp on a gate. It's kinda fun Lol.

EDIT: as a side note, also seen it done to others while in a smartbombing gatecamp but not smartbombing myself.

“Some capsuleers claim that ECM is 'dishonorable' and 'unfair'. Jam those ones first, and kill them last.” - Jirai 'Fatal' Laitanen, Pithum Nullifier Training Manual c. YC104

Neuntausend
Republic Military School
Minmatar Republic
#11 - 2015-12-29 12:57:32 UTC
Lucas Kell wrote:
Alternatively they could just do what any normal person would do and multithread their server code.



https://pbs.twimg.com/media/B4rVC4ICQAAr65m.jpg

It's not that easy.
Nafensoriel
Brutor Tribe
Minmatar Republic
#12 - 2015-12-29 13:06:49 UTC
BUWHAHAHAHAHAHA!
"Just multithread it"
Yeah. No. Id rather light myself on fire and let a thousand rabid chihuahuas feast upon my still living charred corpse.


TiDi is an efficient and much loved replacement to the old method.. The black screen of death.
Once upon a time there was no controlled slowdown. You either got in and lagged like a dead monkey trying to play tennis or you got treated to the black screen of "go make dinner while your client crashes and you die".

I know I know.. saying TiDi is good even though it harvests your soul to sit in -10% TiDi seems weird... but trust me the alternative is far worse.
Solecist Project
#13 - 2015-12-29 13:09:15 UTC
Nevyn Auscent wrote:
Solecist Project wrote:
Not sure smartbombs doing damage in warp is a bug.

Depends on where that happens.


Also, shouldn't you ve called Molle? :P

Smart bombs always do damage even if you are in warp still. That's a long standing feature of smart bombs.

Yeah that was my thought as well.
Where's Santo Trafficante anyway... ^_^

That ringing in your ears you're experiencing right now is the last gasping breathe of a dying inner ear as it got thoroughly PULVERISED by the point roaring over your head at supersonic speeds. - Tippia

Always Shi
t Posting
#14 - 2015-12-29 13:19:06 UTC
Lord Molly wrote:
Id be interested to hear from CCP about how they intend to improve the cancer that is TiDi.



  • TiDi is a goddamn luxury compared to what came before it
  • Smartbombs have always damaged ships in warp, it's just less noticeable at 100% speed
  • The new TQ hardware coming in a couple of months time (plus the recent and ongoing software improvements like BitB) are all helping to make TiDi occur less often.


In summary, your post is bad and you should feel bad.
Lucas Kell
Solitude Trading
S.N.O.T.
#15 - 2015-12-29 13:40:05 UTC
Neuntausend wrote:
Lucas Kell wrote:
Alternatively they could just do what any normal person would do and multithread their server code.
https://pbs.twimg.com/media/B4rVC4ICQAAr65m.jpg

It's not that easy.
I didn't say it was easy, but good developers don't give up because a task is difficult. Tidi isn't a real solution.

The Indecisive Noob - EVE fan blog.

Wholesale Trading - The new bulk trading mailing list.

Kagura Nikon
Native Freshfood
Minmatar Republic
#16 - 2015-12-29 13:40:43 UTC
Lord Molly wrote:
This video documents some real bad tidi issues

If you look closely, you might also notice some glitching, such as smart bombs doing damage in warp.......

All of this is caused by server node overload when to many people are in one system activating too many mod.


To combat this very frustrating trait of gameplay, CCP made a jolly form that you fill in and effectively tell them there is gonig to be a big fight and etc etc, so that they may reinforce a node and divert resources to said systems node and hopefully make it all ok.

Only issue is, that's not really the case is it. Have you ever been to a public event or a large pre-arranged scrap where tidi is down to 10%? where you fire a volley, then go and make dinner and come back to find you are still on the same cycle?

Id be interested to hear from CCP about how they intend to improve the cancer that is TiDi.



TIDi cancer? spoiled kid. before Tidi when we jumped into a large fight what hapepned is that you would wait between 30 minutes to 2 hours with a black screen. then 30% of people woudl disconnect, other 33% would load grid and be unable to use any controls of the ship. Other 13% would be able to laod and press fire a ingle time per hour. The rest would wait another hour and try their luck again...

"If brute force does not solve your problem....  then you are  surely not using enough!"

Nat Silverguard
Aideron Robotics
Aideron Robotics.
#17 - 2015-12-29 13:47:39 UTC
ah, ignorance is indeed a bliss.

never change, op, never change. o7

Just Add Water

Kagura Nikon
Native Freshfood
Minmatar Republic
#18 - 2015-12-29 13:49:38 UTC
Lucas Kell wrote:
Neuntausend wrote:
Lucas Kell wrote:
Alternatively they could just do what any normal person would do and multithread their server code.
https://pbs.twimg.com/media/B4rVC4ICQAAr65m.jpg

It's not that easy.
I didn't say it was easy, but good developers don't give up because a task is difficult. Tidi isn't a real solution.



Introducing paralelism on a task whose bottleneck is I/O with the database and sincronization is not always the best thing to do. If you cannot change the model to remove data dependency you are not going to solve anything. Usually it is better to remodel the system. The thigns that are needed for an eve fleet battle are NOT huge and do nto justify even a single modern CPU being bottlenecked.

Physics simulation of particles is a classic example of problem that in theory is a prime candidate for paralelization, but when you do it you incur in the need of a more complex iteration model and syncronization mechanisms that downgrade the performance massively> The result is that you need then to use several CPU to match what you could earlier do with a single CPU. Of course, if you can just throw even more CPUS at the problem you have a gain, but the cost involved is high, therefore it is not always a good solution.

"If brute force does not solve your problem....  then you are  surely not using enough!"

Kagura Nikon
Native Freshfood
Minmatar Republic
#19 - 2015-12-29 13:52:14 UTC
Lucas Kell wrote:


Tidi was a stopgap measure. It allowed them to give the server more time to get through the event queue so that they could still run everything on a single heavily overclocked core. At some point the promoted to "the fix" and now it's a burden. Things like BIAB help, but in the long run the solution has to be modernising their server code.


Sorry, I might be mistaken, but you sound as someone that has no clue what multithreading is. What do you think BIAB is? It IS ALREADY paralelization (external) of the most consuming task of the node.

"If brute force does not solve your problem....  then you are  surely not using enough!"

Lucas Kell
Solitude Trading
S.N.O.T.
#20 - 2015-12-29 13:55:06 UTC
Kagura Nikon wrote:
Introducing paralelism on a task whose bottleneck is I/O with the database and sincronization is not always the best thing to do. If you cannot change the model to remove data dependency you are not going to solve anything. Usually it is better to remodel the system. The thigns that are needed for an eve fleet battle are NOT huge and do nto justify even a single modern CPU being bottlenecked.
But we already know that's not the bottleneck. The bottleneck is CPU. This is why when you fill in that form they use a reinforced server, which is one where most of the cores are shut off on the CPU and used as a heat spreader while the core in use is overclocked like crazy. A single modern CPU has multiple cores, all (or most) of which will be used by most CPU heavy game servers. EVE doesn't. A single core has to process all of the information coming in from the thousands of players in the system. Even on a modern CPU, that is too much.

The Indecisive Noob - EVE fan blog.

Wholesale Trading - The new bulk trading mailing list.

123Next pageLast page