Pages: 1 [2] :: one page |
|
Author |
Thread Statistics | Show CCP posts - 0 post(s) |

Phrixus Zephyr
D00M. Triumvirate.
|
Posted - 2007.08.09 23:02:00 -
[31]
Originally by: Seiver D'amross Edited by: Seiver D''amross on 09/08/2007 13:12:34 anytime i am in a large fleet and hostile gang jumps into system i de-sync, it was how i lost my first carrer and CCP said it never happend.
...and how does that back up your claim? Thats not proof.
You don't know its the gang system at all, you're making an assumption with little or no evidence available to you.
Originally by: Lucian Sulla "I have sand in my vagina, so I can't pvp today"
|

S3dINSTBE
The JORG Corporation FATAL Alliance
|
Posted - 2007.08.09 23:14:00 -
[32]
Originally by: Phrixus Zephyr
Originally by: Seiver D'amross Edited by: Seiver D''amross on 09/08/2007 13:12:34 anytime i am in a large fleet and hostile gang jumps into system i de-sync, it was how i lost my first carrer and CCP said it never happend.
...and how does that back up your claim? Thats not proof.
You don't know its the gang system at all, you're making an assumption with little or no evidence available to you.
Wow you should work for CCP, you got that speech down to a tee
---------------------------------------- Shove it |

Alias11
GoonFleet GoonSwarm
|
Posted - 2007.08.09 23:20:00 -
[33]
Originally by: Seiver D'amross Edited by: Seiver D''amross on 09/08/2007 13:12:34 anytime i am in a large fleet and hostile gang jumps into system i de-sync, it was how i lost my first carrer and CCP said it never happend.
Large fleets exacerbate the Desync problem and make it more likely to happen, but they are not the cause of it.
|

Angor
The JORG Corporation FATAL Alliance
|
Posted - 2007.08.09 23:52:00 -
[34]
/signed....
I'll try anything to help fix this increasingly broken game _______________________________ WTS 50m sp PvP Minmitar Character including Dread and Carrier |

Dylan Rhade
Caldari UK Corp FATAL Alliance
|
Posted - 2007.08.10 06:25:00 -
[35]
I say the whole of EvE just carebears to get the isk back for the ships unfairly lost due to petitioning and boycotts pvp on large scales entirely.
Theres not much point of getting all of your uber-kick-ass-froob bashing alliance, all kitted out in 200 mil T2 sniping BS, camp a gate, and have one tenth of your fleets numbers jump in, and lay waste to everything they see, with no possibility to defend themselves because the defenders de-sync, drop and hey presto MORE ****ed of customers.
Is it SO hard to tell when 200+ people in the same system disconnect instantly, realise they have de-sync'd thus allowing petitions for lost ships on such a stupid scale for such an annoying reason.
Stability not flashier graphics
D
|

Jtbenns
Gallente QUANT Corp. Southern Connection
|
Posted - 2007.08.12 14:40:00 -
[36]
I'd agreewith the op and if you wanted me to do something in this just mail me and ill bring a few people ut it comes down to who really is truthfully going to not break the rules. other than that im game. --------------------------------- RWREEEEEEEEEEEEEEEEEEEEE
|

End Yourself
Core Domination Big Bang Quantum
|
Posted - 2007.08.12 14:50:00 -
[37]
Pretty funny to see the same delusional whines we were used to from ASCN, G... coming from BoB+Pets nowadays. 
--- Fighting for peace is like screwing for virginity. |

Beyond Horizon
Black Omega Security
|
Posted - 2007.08.12 15:21:00 -
[38]
Originally by: w0rmy
Originally by: End Yourself
Regarding desynch: I have desynched WITHOUT beeing in a fleet or gang at all, so have others i know.
Have desyncd warping into a belt, bouncing off a roid, in a system which I was the only one there.

lol wormy is in mcorp   -
BH |

Necronomicon
Caldari KIA Corp KIA Alliance
|
Posted - 2007.08.12 16:47:00 -
[39]
It really is quite simple, and the OP sort of has it right. CCP should invite the issuers of de-sync complaints to put aside their differences for an hour or so, and bring everything (MS/Titans, the whole shebang) to a pre-arranged system, gang up and meet at a location, devs can then insta move them all to a system in Jovs space. They do not need to engage, just deploy drones and wait....de-sync will come. Everyone can then be insta moved back to their original system, and get back to playing with CCP having all the info they need.
Yes, it would be a hectic couple of hours for CCP and the ppl involed, but it would be invaluable to their fixing this issue which is in EVERYONE'S best interests.
Carlsberg dont make Eve Pilots, but if they did, i wouldnt be one of them.
|

Smakko
Ad Astra Vexillum Brutally Clever Empire
|
Posted - 2007.08.13 12:43:00 -
[40]
Some of the instances of desync people mention ITT seem related to the clock bug in windows that ClockMon solves. The system clock in windows is known to be buggy, which is why regularly syncing the system clock with the BIOS based real time clock (as Clockmon does) actually solves some crashes in EvE, and in other games. On the other hand, it's beyond me why all games written for Windows don't have special code designed to use the higher precision RTC clock exclusively. In game coding, one of the most basic things to nail is the timing function.
I would be truly mortified if the desync problem could be resolved by something as simple as running Clockmon on the server nodes... but I don't really believe it's that simple.
The system clock related timing bug in windows usually occurs due to heavy processor load on the system. For example, a program that heavily taxes system resources, such as a video game, will somehow delay the execution of the operating system's timing function, resulting in a decay in accuracy of the system clock.
One method I'm aware of in calculating precise time for game related software has something to do with counting the number of processor cycles and interpolating the precise time. For example, if the hardware clock reports the current time to the precise second, and the CPU runs 1000 operations per second, I can calculate the exact time to the 1/1000th of a second by counting the number of operations since the time of the last second boundary reported by the hardware clock. (Eh, sort of... by trying to explain this in laymen's terms I'm resorting to an unrealistic example)
What irks me is that the desync problem seems to occur most regularly with large fleet battles, which would, of course, be a scenario of heavy processor load. Client-server desync during heavy processor load sounds like a classic case of buggy timing code. I really feel like CCP would know all this, so their failure to resolve the desync issue in a timely fashion baffles me. People were installing Clockmon for EvE client crashes in 2004 ffs.
My guess is that CCP is aware that the bug is timing related, but they have not yet located all the places in the codebase that need to be altered. It is possible that their use of stackless Python obfuscates the relationship between their code and low level timing functions, making diagnosis even more difficult. Also, the sheer number of calculations performed server-side that rely on precise time measurement may be a significant barrier to resolving the desync problem in any elegant way. Barring an elegant solution, one has to fall back to a series of hacks that can be akin to a house of cards. My only comfort is that I love EvE despite its serious warts, and I sincerely hope CCP solves the problem before it's too late. Also, a little perspective helps assuage me, it's not nearly as bad as Asheron's Call or Ultima Online was. Those games crashed if you sneezed on the keyboard.
Or, alternately, desync has nothing to do with timing whatsoever, CCP has been over their timing function with a fine toothed comb, and they're scratching their heads over what the hell is going on.
In any case, this seems like the wrong forum to be discussing this.
Originally by: NSA DAKILLA bruce the cream of noobs
|
|

Neena Valdi
Art of War Cult of War
|
Posted - 2007.08.13 16:02:00 -
[41]
Interesting idea, but I doubt that the actual EVE desync problem have any relation to bios - system clock's small 'shifts'. Also the described by you problem happens very rarely nowadays as Windows NT+ do manage the system resources quite good.
P.S. The 'desync' happens even during low CPU load, so it's not related to lack of system resources or clock anyhow.
|

Thundirr
Mercurialis Inc. Interstellar Alcohol Conglomerate
|
Posted - 2007.08.13 16:19:00 -
[42]
Edited by: Thundirr on 13/08/2007 16:19:04 Recently IAC, FIX, and MC were involved in a very large fleet fight in 49-. As has been posted MC lost a MS reportedly due to 'de=sync'.
If this actually gets a GM respone I might be able to talk to a rep from FIX and MC and get IAC to agree to a.. Test fight. We could easily bring a 100+ gang to do this and im sure FIX and MC wouldnt mind fixing this game so we can get to what we love doing most.. Killing each other fairly.
Thundirr out.
|

Smakko
Ad Astra Vexillum Brutally Clever Empire
|
Posted - 2007.08.13 19:45:00 -
[43]
Originally by: Neena Valdi Interesting idea, but I doubt that the actual EVE desync problem have any relation to bios - system clock's small 'shifts'. Also the described by you problem happens very rarely nowadays as Windows NT+ do manage the system resources quite good.
P.S. The 'desync' happens even during low CPU load, so it's not related to lack of system resources or clock anyhow.
I haven't used NT in a while, but the system clock problem I referred to still is an issue on my XP system (updated with all patches). In any case, you point out that the desync happens during low CPU load, but fail to differentiate between client and server load. Even in the case where both client and server load are relatively low, it is possible that somewhere in either the client or server code there is something that is highly sensitive to even slight lag, such as that which might occur when trying to pull data from the database - a separate system.
I'm of the mind that the desync that people occasionally get when initiating warp, or when bumping into asteroids in empty systems, are separate but related issues to the mass desync / login queue issues that result from large-ish fleet battles.
It's all just speculation, but my main point is that the desync related problems on the client, that are solved for many by regularly resyncing the system clock to the bios rtc clock, are potentially a separate issue from any server-side related problems, and is definitely not solved by any recent improvements to the Windows NT/2000/XP OS code.
Furthermore, I really don't understand how minor timing issues or lag resulting from communication delays between various cluster subsystems could result in catastrophic failure of the game session. It seems like the software should be able to recover, at least after 30 secs or so, from whatever timing issues the devs have reported in other threads... but if bug solving were easy we'd always have perfect code.
Anyhow, I wasn't being so arrogant as to presume the desync problem could be simply resolved simply by running Clockmon on the server - that suggestion was made tongue in cheek. But I still maintain that many people that complain of some desync related issues are experiencing something different than what happens in fleet battles - where pilots are unable to activate modules or navigate, enter a login queue when trying to relogin to the system, but a small number of pilots are able to function nearly 100%. The fact is Clockmon works for many people, at least in resolving certain client-side issues, and has worked since about 2004 when it was first recommended to pilots on eve-o.
Originally by: NSA DAKILLA bruce the cream of noobs
|

Neena Valdi
Art of War Cult of War
|
Posted - 2007.08.13 20:20:00 -
[44]
Originally by: Smakko
Originally by: Neena Valdi Interesting idea, but I doubt that the actual EVE desync problem have any relation to bios - system clock's small 'shifts'. Also the described by you problem happens very rarely nowadays as Windows NT+ do manage the system resources quite good.
P.S. The 'desync' happens even during low CPU load, so it's not related to lack of system resources or clock anyhow.
I haven't used NT in a while, but the system clock problem I referred to still is an issue on my XP system (updated with all patches).
I don't have the issue you describing on any XP systems i've been working with over last years. Saw it once long ago on 98, but that's basically it.
Originally by: Smakko
In any case, you point out that the desync happens during low CPU load, but fail to differentiate between client and server load. Even in the case where both client and server load are relatively low, it is possible that somewhere in either the client or server code there is something that is highly sensitive to even slight lag, such as that which might occur when trying to pull data from the database - a separate system.
Most likely the problem lies somewhere in algorithms. As I understand client during the lag tries to 'forsee' the feature, on the other hand the packets between server and clients have a minimalistic view: as less data as possible... Anyways the desync problem is not simple, otherwise it'd be fixed long ago.
Originally by: Smakko
It's all just speculation, but my main point is that the desync related problems on the client, that are solved for many by regularly resyncing the system clock to the bios rtc clock, are potentially a separate issue from any server-side related problems, and is definitely not solved by any recent improvements to the Windows NT/2000/XP OS code.
No, no, it can't be a clock related problem. At least not on client side... - I am sure server have no idea about client side clock. The client it self is almost only an interface with minimal logic on client side. Client tells server what action you *want* to do, blocks you from doing 'impossible' actions like trying to activate some modules during the warp procedure etc.
Originally by: Smakko
Furthermore, I really don't understand how minor timing issues or lag resulting from communication delays between various cluster subsystems could result in catastrophic failure of the game session. It seems like the software should be able to recover, at least after 30 secs or so, from whatever timing issues the devs have reported in other threads... but if bug solving were easy we'd always have perfect code.
Because it's not about the lag (or not only about the lag). It's a bug somewhere in code.
Originally by: Smakko
Anyhow, I wasn't being so arrogant as to presume the desync problem could be simply resolved simply by running Clockmon on the server - that suggestion was made tongue in cheek. But I still maintain that many people that complain of some desync related issues are experiencing something different than what happens in fleet battles - where pilots are unable to activate modules or navigate, enter a login queue when trying to relogin to the system, but a small number of pilots are able to function nearly 100%. The fact is Clockmon works for many people, at least in resolving certain client-side issues, and has worked since about 2004 when it was first recommended to pilots on eve-o.
Clockmon helps with what? What kind if 'client side issues' does it fix?
|

Smakko
Ad Astra Vexillum Brutally Clever Empire
|
Posted - 2007.08.15 17:19:00 -
[45]
This discussion isn't going anywhere, it's devolved into he-said-she-said. Please refer to previous posts on Clockmon for more information.
http://oldforums.eveonline.com/?a=topic&threadID=111435 http://oldforums.eveonline.com/?a=topic&threadID=122025
Originally by: NSA DAKILLA bruce the cream of noobs
|

Iasius
Warp Angels Free Trade Zone.
|
Posted - 2007.08.15 17:46:00 -
[46]
MMO programmer really do push the netcode. Several years ago SOE (grr) tweaked Planetside to far so it dropped to often. I think ccp need to pull the lever back two notches. I have a couple of mates in a house who have done lots of testing and ccp's internet node keeps dropping them.
Anyway where is the Java command line version of eve  . Warp Angels - Acorns To Trees. Now Recruiting. |
|
|
|
Pages: 1 [2] :: one page |
First page | Previous page | Next page | Last page |