Pages: 1 [2] :: one page |
|
Author |
Thread Statistics | Show CCP posts - 2 post(s) |
Aldir Rundal
Gallente The Order of Chivalry Nex Eternus
|
Posted - 2007.07.19 00:50:00 -
[31]
Originally by: Ask Unbeatable
Originally by: CCP Explorer
Erlendur
So thats how you came up with the system names, aye
I wondered when somebody else was going to notice that
Recruiting |
Nnoz
Caldari Sturmgrenadier Inc R i s e
|
Posted - 2007.07.19 01:00:00 -
[32]
aye , its been repeated over and over the last month this issue. but no resolve. canx both my accounts i wont accept any more rebilling. I really love Eve as a game but END Game is fleet battles and even small ones are no longer dependent on the better pilots , only the one who gets the least lag during the encounter.
This issue needs your resolve EVE, not the walking in stations or the other accessories, please fix your game while i am away so i may consider rejoining.
Good luck to anyone remaining and hopefully you will be on the winning side of that lag crap.
Fly safe
|
Kirell
|
Posted - 2007.07.19 23:29:00 -
[33]
Yes I've cancelled my accounts.
Yes my ship loss has been replaced.
No you can't have my stuff.
I have little faith that thier "upcoming patch" will fix the issue.
CCP loses my sub money for a little while now.
Fighters V = 54 days. maybe close to the end of that I will see what's going on.
Who knows maybe I will get into Warhammer online beta. then I'll give me stuff away in Jita
|
An Anarchyyt
Gallente Federal Navy Academy
|
Posted - 2007.07.19 23:40:00 -
[34]
I'm sure the entire EVE populace and everyone at CCP is crying over such a loss.
|
Haffrage
Sharks With Frickin' Laser Beams
|
Posted - 2007.07.20 00:06:00 -
[35]
Originally by: Kirell I think the only way CCP is going to fix this issue is if we hit it where it hurt them the most.
Quote: We now have a confirmed fix for this issue and we are planning to release it as a part of Revelations 2.1 next week. ... In addition we're fixing a few other defects, in particular performance related issues. More details later in the patch notes.
Quote: Yes I've cancelled my accounts.
Yes my ship loss has been replaced.
No you can't have my stuff.
I have little faith that thier "upcoming patch" will fix the issue.
I...err...
lol?
So you say cancel until they fix the bug...they say they fix the bug. Before the patch is even deployed you say you still won't uncancel because it *probably* won't work. Yeah, I'm gonna go with "lol" on this one -----
Tech 2 Tier 2 Battlecruisers Eve GUI Tweaks |
Swedish Bob
GoonFleet GoonSwarm
|
Posted - 2007.07.20 03:55:00 -
[36]
Originally by: CCP Explorer Profile Guided Optimization (PGO) feature of Visual Studio 2008.
I hope you mean VS2005 and not 2008. VS2008 only had its first beta out in the last couple months. And I can only imagine the hidden bugs its cranking out.
|
Cade Burey
Amarr Underworld Protection Agency The Crimson Federation
|
Posted - 2007.07.20 09:21:00 -
[37]
Originally by: Drasked
Everyone i know pvps with 2 accounts or more a the same time, myself i can't understand how to play with only 1 account.
I must be in the ultra minor minority then, I dont even have an alt, never mind another account. I'm too mean to share my skill training time with another character
|
Shenko Minara
|
Posted - 2007.07.20 09:35:00 -
[38]
Originally by: MotherMoon and yet the OP still quit... I want to find them and give them a good smack. this was a VERY VERY fast fix for a bug this big.
I'm sorry, but you're wrong. This bug has been known for what, a month? I forget when this one appeared. It's a 'show stopper' class bug that affects people in huge ways.
Don't forget CCP's rather questionable procedure of only re-imbursing your ship and stuff that didn't survive in the wreck; this bug has cost people loads in time, effort and ISK and it's completely CCP's fault.
And it took them this long to come out with a fix? Most software companies would've jumped on this and had a fix out inside a week. Credit where credit is due though, for CCP this is actually a lightning response.
Personally I'll give them the benefit of the doubt, since their press release is "We have confirmed a fix" rather than "we believe we've found the cause". I will not, however, be surprised by post Rev 2.1 threads titled, "I'm still de-synching!!!!!!". -- 99% of Eve-o posters should stop posting. This probably includes me, but definitely includes you. |
Kharadran Sullath
Caldari IntoXication Inc PioneerX Syndicate
|
Posted - 2007.07.20 09:35:00 -
[39]
But at least send the message your willing to stand against CCP until they have this fixed.
Sure. CCP, I AM STANDING AGAINST YOU UNTIL TUESDAY! GG on quitting a few days before the fix just because you don't believe that it will work. ------ --Don't get saucy with me Bernaise!-- |
Valandril
Caldari Resurrection R i s e
|
Posted - 2007.07.20 10:05:00 -
[40]
Originally by: CCP Explorer We now have a confirmed fix for this issue and we are planning to release it as a part of Revelations 2.1 next week.
Revelations 2.1 is a "Need for Speed" upgrade; we're upgrading to Stackless Python 2.5 and we're compiling the Python interpreter using the new Profile Guided Optimization (PGO) feature of Visual Studio 2008.
There is a clear reduction in memory and CPU usage and we'll have numbers to publish at the time of release.
In addition we're fixing a few other defects, in particular performance related issues. More details later in the patch notes.
No way, desync fix that work ? Pix or it never happened :P --- Nanananana, u can't moderate my sig ! |
|
Flashheart
Caldari Dangermouse Inc.
|
Posted - 2007.07.20 10:41:00 -
[41]
I took part in yesterdays fleet test in the test server and to be fair it looks good.
It was about 50 on 50 with everything from cruisers to Motherships and I experienced no desync issues at all, response lag was almost unnoticeable and after several warp out/in I got mauled by a dread There were no reports of anyone having issues that I could see in the chats.
Place was littered with wrecks, hundreds of drones/fighters and everything was smooth and clean with no module or targetting delay.
Admittidly it was on test and it may not be so nice on the live server, but compared to previous tests it was like a new game.
|
Kinsy
Celestial Apocalypse Insurgency
|
Posted - 2007.07.20 10:45:00 -
[42]
Theres a patch next week to fix it, noob.
Read the news maybe next time.
Plim: I was driven to absolute idiocy by CCP |
|
CCP Explorer
|
Posted - 2007.07.20 10:54:00 -
[43]
Edited by: CCP Explorer on 20/07/2007 10:56:27
Originally by: Swedish Bob
Originally by: CCP Explorer Profile Guided Optimization (PGO) feature of Visual Studio 2008.
I hope you mean VS2005 and not 2008. VS2008 only had its first beta out in the last couple months. And I can only imagine the hidden bugs its cranking out.
Yes, was thinking VS8 and intended to write VS2005, thanks for the correction. We only upgraded to VS2005 when SP1 was out to allow for the worst bugs to be corrected.
Erlendur S. Thorsteinsson Software Director EVE Online, CCP Games |
|
Sakura Nihil
Tabula Rasa Systems The Star Fraction
|
Posted - 2007.07.20 11:05:00 -
[44]
Originally by: MotherMoon and yet the OP still quit... I want to find them and give them a good smack. this was a VERY VERY fast fix for a bug this big.
EVE bugs...
Weeding out the weak-willed since 2003 .
Screw Sovereignty! |
Imhothar Xarodit
Minmatar Wolverine Solutions Interstellar Alcohol Conglomerate
|
Posted - 2007.07.20 12:34:00 -
[45]
Edited by: Imhothar Xarodit on 20/07/2007 12:40:35 Edited by: Imhothar Xarodit on 20/07/2007 12:36:03 Edited by: Imhothar Xarodit on 20/07/2007 12:34:56
Originally by: CCP Explorer Edited by: CCP Explorer on 20/07/2007 10:58:48 We now have a confirmed fix for this issue and we are planning to release it as a part of Revelations 2.1 next week.
Revelations 2.1 is a "Need for Speed" upgrade; we're upgrading to Stackless Python 2.5 and we're compiling the Python interpreter using the new Profile Guided Optimization (PGO) feature of Visual Studio 2005.
There is a clear reduction in memory and CPU usage and we'll have numbers to publish at the time of release.
In addition we're fixing a few other defects, in particular performance related issues. More details later in the patch notes.
I was always wondering why EVE is coded in Python. I mean, there is no doubt about the performace gains of C++ vs Python. I know, that Python as an interpreter makes developing a bit easier as you can test new things without recompiling, but I think that the performance loss by using Python is so huge that it might be considered to port the code (or at least the very very very performance intensive parts) to C++ or a comparable compiled environment or even assembly optimized code.
Are you actually using low-level libraries for such tasks or is the entire server written in Python? And what about the Client? Those are actually technical facts about EVE that I would be interest alot in (like what parts of the system are wirtten in what language, what subsystems have the highest load etc). Are there any chances you could provide such a document? Might even be an interesting subject for a scientific analysis
Sure EVE is in this way very special and something unique but it would interest me alot what the performance benefits of using another system could be. And I bet there are means to let the server go much better than it is now...
|
Matthew
Caldari BloodStar Technologies
|
Posted - 2007.07.20 12:41:00 -
[46]
Originally by: Imhothar Xarodit I was always wondering why EVE is coded in Python. I mean, there is no doubt about the performace gains of C++ vs Python. I know, that Python as an interpreter makes developing a bit easier as you can test new things without recompiling, but I think that the performance loss by using Python is so huge that it might be considered to port the code (or at least the very very very performance intensive parts) to C++ or a comparable compiled environment or even assembly optimized code.
From memory, I'm sure they've mentioned before that they use python specifically because it allows for faster development of new stuff, but all the performance-critical stuff is in C, with newer stuff being optimized over to C from python as the code becomes more mature. ------- There is no magic Wand of Fixing, and it is not powered by forum whines. |
Jenna Shame
|
Posted - 2007.07.20 14:02:00 -
[47]
I'm not sure how anyone can call this a fast fix. This has been a known issue for several weeks now, since the rev 2 patch.
I can understand new features not working right, but when you break something that wasn't broken before AND it totally screws up your game for a huge section of the player base a fix two months later or so is NOT fast.
I love EvE and over all I think CCP does a good job, but I'm not a fan of their patching. Even their own volunteer support tells you to set a long train for a reason.
|
Braaage
Laborius Chapter
|
Posted - 2007.07.20 14:13:00 -
[48]
If you were part of the TEST server croud you'd know:
1) CCP have been trying to fix this since it was found 2) We (the people who frequent the test server) have been assisting CCP as much as possible in tracking down the cause and testing a number of solutions (the last of which was successful) 3) That CCP care a LOT. --------------------------------------------------------------------------------------- www.eve-guides.com POS, Outpost and Sovereignty info |
Haks'he Lirky
8th
|
Posted - 2007.07.20 14:19:00 -
[49]
Originally by: Jenna Shame
I'm not sure how anyone can call this a fast fix. This has been a known issue for several weeks now, since the rev 2 patch.
I can understand new features not working right, but when you break something that wasn't broken before AND it totally screws up your game for a huge section of the player base a fix two months later or so is NOT fast.
I love EvE and over all I think CCP does a good job, but I'm not a fan of their patching. Even their own volunteer support tells you to set a long train for a reason.
Ive never understood customers who think that just because a bug or malfunction is important that it automatically means it can be fixed yesterday, I understand that you want it to be fixed yesterday but reality is not dictated by the customer nor is it dicated by the developer, Fixing stuff takes the time it takes.
Looking at the seriousness of this issue and at the same time it's vague nature I am frankly quite impressed that CCP has a fix planned allready (I wouldnt be surprised if it wouldnt be fixed in whole with the patch).
|
Shenko Minara
|
Posted - 2007.07.20 14:31:00 -
[50]
Originally by: Haks'he Lirky
Ive never understood customers who think that just because a bug or malfunction is important that it automatically means it can be fixed yesterday, I understand that you want it to be fixed yesterday but reality is not dictated by the customer nor is it dicated by the developer, Fixing stuff takes the time it takes.
As someone that has not only worked in the software industry, but also as a small team on a project whose budget is in the millions, your priorities and procedures are different.
When you meet a 'show stopper' in this environment, you drop everything to fix it. Since Eve is by its very nature an incremental development (allowing better documentation, code tracability and simpler test matrices), and the fact that they have a live test environment, this has been a very slow fix.
-- 99% of Eve-o posters should stop posting. This probably includes me, but definitely includes you. |
|
Haks'he Lirky
8th
|
Posted - 2007.07.20 14:41:00 -
[51]
Originally by: Shenko Minara
Originally by: Haks'he Lirky
Ive never understood customers who think that just because a bug or malfunction is important that it automatically means it can be fixed yesterday, I understand that you want it to be fixed yesterday but reality is not dictated by the customer nor is it dicated by the developer, Fixing stuff takes the time it takes.
As someone that has not only worked in the software industry, but also as a small team on a project whose budget is in the millions, your priorities and procedures are different.
When you meet a 'show stopper' in this environment, you drop everything to fix it. Since Eve is by its very nature an incremental development (allowing better documentation, code tracability and simpler test matrices), and the fact that they have a live test environment, this has been a very slow fix.
So you have never met with an issue that no matter how many developers you throw into the fray it still takes sweet frigging time to fix?
Do you also have a single version software that gets pushed to all your customers at the same time thus making it extremely hard to ask people to downgrade while the issue is being addressed?
I agree that show stoppers should be addressed, but I also know that while EVE is developed incrementally it still has a severly buggy client server interface and I have heard through some live dev blog that a complete overhaul is in the works. This suggests that while they have a nice development environment they also have faulty old code that is bogging down the process.
I have and do work in the software industry aswell with a budget ranging in some dollar value that wouldnt impress you, neither should it, my argument is simple. Issues can take their time to fix no matter how many resources you throw at it.
|
Isyel
Minmatar Queens of the Stone Age Anarchy Empire
|
Posted - 2007.07.20 14:45:00 -
[52]
Originally by: Shenko Minara
Originally by: Haks'he Lirky
Ive never understood customers who think that just because a bug or malfunction is important that it automatically means it can be fixed yesterday, I understand that you want it to be fixed yesterday but reality is not dictated by the customer nor is it dicated by the developer, Fixing stuff takes the time it takes.
As someone that has not only worked in the software industry, but also as a small team on a project whose budget is in the millions, your priorities and procedures are different.
When you meet a 'show stopper' in this environment, you drop everything to fix it. Since Eve is by its very nature an incremental development (allowing better documentation, code tracability and simpler test matrices), and the fact that they have a live test environment, this has been a very slow fix.
I R 1337 psychiatrist with multiple degrees in i know it all on teh intrawebz and say you now send me isk and we do some counseling, ok? ok! ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Originally by: Wrangler We don't want to discriminate anyone! We want *both* anti-social *and* social players to grief each other!
|
Tarminic
Black Flame Industries
|
Posted - 2007.07.20 14:52:00 -
[53]
Edited by: Tarminic on 20/07/2007 14:52:20
Originally by: Kirell Who knows maybe I will get into Warhammer online beta. then I'll give me stuff away in Jita
Enjoy your Wowhammer online. I also have another recommendation.
Originally by: Imhothar Xarodit I was always wondering why EVE is coded in Python. I mean, there is no doubt about the performace gains of C++ vs Python. I know, that Python as an interpreter makes developing a bit easier as you can test new things without recompiling, but I think that the performance loss by using Python is so huge that it might be considered to port the code (or at least the very very very performance intensive parts) to C++ or a comparable compiled environment or even assembly optimized code.
I believe the only significant performance loss is the overhead of using the python runtime, though i can't be positive. The primary motivation behind using Stackless Python over C++ or something comparable is that stackless python allows the very fast and effective sharing of memory, which is practically required considering the massive concurrency issues EVE deals with. Doing this with standard C++ threads would not work nearly as well due to the sheer numbers.
Quote:
Are you actually using low-level libraries for such tasks or is the entire server written in Python? And what about the Client? Those are actually technical facts about EVE that I would be interest alot in (like what parts of the system are wirtten in what language, what subsystems have the highest load etc). Are there any chances you could provide such a document? Might even be an interesting subject for a scientific analysis
I can't answer any of those questions, and that would be very interesting to know for us nerds.
Originally by: Shenko Minara And it took them this long to come out with a fix? Most software companies would've jumped on this and had a fix out inside a week. Credit where credit is due though, for CCP this is actually a lightning response.
Proof or STFU.
------------ ULTIMATE LAG SOLUTION IBTL! IBDS/DC! IBTC! 1st in a BoB post! And other such forum tom-foolery. |
Xorlev
|
Posted - 2007.07.20 15:22:00 -
[54]
Yeah, right. Most companies would let the software rot for a few months while they released the new version which fixed the bugs and made you pay to upgrade.
I think some thanks are in order for CCP, being in the chaos channel and frequenting the test servers I could see all the discussions, tests, and the dedication that went into fixing this. Quit acting like spoiled children about it. You asked for a fix, you got one. Thank them, rather than saying any other software company would've fixed it, blah blah.
|
Shenko Minara
|
Posted - 2007.07.20 15:23:00 -
[55]
Edited by: Shenko Minara on 20/07/2007 15:24:05
Originally by: Isyel I R 1337 psychiatrist with multiple degrees in i know it all on teh intrawebz and say you now send me isk and we do some counseling, ok? ok!
Excellent. Thanks for contributing. Please see my sig. :)
---- "So you have never met with an issue that no matter how many developers you throw into the fray it still takes sweet frigging time to fix?"
Actually, no. Our testing is comprehensive enough to spot most show stoppers, and on the occaisions when we miss them, we offer workarounds or rollbacks.
"Do you also have a single version software that gets pushed to all your customers at the same time thus making it extremely hard to ask people to downgrade while the issue is being addressed?"
Yes. That's exactly what we do. And Eve's inherent nature and distrubution mean that rolling back is far easier for CCP than most companies.
"I have and do work in the software industry aswell with a budget ranging in some dollar value that wouldnt impress you, neither should it, my argument is simple. Issues can take their time to fix no matter how many resources you throw at it."
The dollar amount I mentioned earlier (pounds sterling actually ;p) was simply to offer a comparison in working practices. And you're right, there *is* a limit to how much resource you can assign a problem before you aren't getting any extra benefit (stacking penalty, heheh). However, in the meantime that problem still needs to be addressed, even if it can't be fixed.
There are too many far-reaching consequences to doing rollbacks as far as Eve is concerned, but there should at least be enough structure and future-proofing in place to support such a device. -- 99% of Eve-o posters should stop posting. This probably includes me, but definitely includes you. |
Imhothar Xarodit
Minmatar Wolverine Solutions Interstellar Alcohol Conglomerate
|
Posted - 2007.07.20 15:46:00 -
[56]
Originally by: Tarminic Edited by: Tarminic on 20/07/2007 14:52:20 I believe the only significant performance loss is the overhead of using the python runtime, though i can't be positive. The primary motivation behind using Stackless Python over C++ or something comparable is that stackless python allows the very fast and effective sharing of memory, which is practically required considering the massive concurrency issues EVE deals with. Doing this with standard C++ threads would not work nearly as well due to the sheer numbers.
The biggest impact on performance between Python and C is the fact, that Python is an interpreter. The code has to be translated to machine instructions on-the-fly and that is a very time consuming process. It sure gives the developer a very easy means of adjusting and quickly testing the code, no doubt. The stack issue you mention is solvable in C with some asemmbly code workaround giving you also "stackless" recursions Considering threads: To be honest, I never wrote a line of Python code, so I can't really argue on this, but threading in generall is handled by the Operating System's scheduler and I doubt the server creates a new thread for every user request ;) But I think that is quite obvious hence there are multiple nodes. My guess is a fixed number of threads per job per node going through user request by priority queue or some other sort of selection algorithm but only CCP can answer that.
Would love to get a draft or something of the EVE cluster structure, showing used algorithms, subsystems etc. Something to dig in and get a better understanding of CCP's work.
|
Tarminic
Black Flame Industries
|
Posted - 2007.07.20 16:50:00 -
[57]
Originally by: Imhothar Xarodit The biggest impact on performance between Python and C is the fact, that Python is an interpreter. The code has to be translated to machine instructions on-the-fly and that is a very time consuming process. It sure gives the developer a very easy means of adjusting and quickly testing the code, no doubt. The stack issue you mention is solvable in C with some asemmbly code workaround giving you also "stackless" recursions
The term "assembly code workaround" sounds very dangerous to me.
Quote: Considering threads: To be honest, I never wrote a line of Python code, so I can't really argue on this, but threading in generall is handled by the Operating System's scheduler and I doubt the server creates a new thread for every user request ;) But I think that is quite obvious hence there are multiple nodes. My guess is a fixed number of threads per job per node going through user request by priority queue or some other sort of selection algorithm but only CCP can answer that.
Hard to say, but it don't think it's impossible that there's a single thread per user per system/node, given that my computer is currently running 434 individual threads (though comparisons on that level are weak at best). It may also be possible that the server uses a thread pool divided across however many systems each node covers.
Quote:
Would love to get a draft or something of the EVE cluster structure, showing used algorithms, subsystems etc. Something to dig in and get a better understanding of CCP's work.
That would be amazing, actually. I would buy it if they sold it. ------------ ULTIMATE LAG SOLUTION IBTL! IBDS/DC! IBTC! 1st in a BoB post! And other such forum tom-foolery. |
Jenna Shame
|
Posted - 2007.07.20 17:58:00 -
[58]
Originally by: Xorlev Yeah, right. Most companies would let the software rot for a few months while they released the new version which fixed the bugs and made you pay to upgrade.
I think some thanks are in order for CCP, being in the chaos channel and frequenting the test servers I could see all the discussions, tests, and the dedication that went into fixing this. Quit acting like spoiled children about it. You asked for a fix, you got one. Thank them, rather than saying any other software company would've fixed it, blah blah.
Well technically its still not fixed until the next patch. If I want to pvp this weekend, which is the only time I can really devote to anything major in game, I will still suffer from desyncs. Last time I did it I had to relog 4 times in two hours, twice in one battle, due to this issue. We stopped forming gangs since we had to constantly reform. I assume that this is a possibility this weekend as well. So no they didn't 'fix' anything yet, they have a 'fix' but its not applied to the servers.
If any issue had 'hot fix' written all over it, I'd think something that makes the game experience frustrating at best, and grossly unfun at worst would be on that list.
I think the biggest issue with the software industry is we are all so desensitized to these issues that any fix feels like we should be great full. I just fired my bank for less.
Also since CCP is not reimbursing for desync issues, they should have a big 'fly at your own risk, game is unstable' warning when you log on.
|
Imhothar Xarodit
Minmatar Wolverine Solutions Interstellar Alcohol Conglomerate
|
Posted - 2007.07.21 14:11:00 -
[59]
Edited by: Imhothar Xarodit on 21/07/2007 14:10:23
Originally by: Tarminic The term "assembly code workaround" sounds very dangerous to me.
Not if you know what you're doing
|
|
|
|
Pages: 1 [2] :: one page |
First page | Previous page | Next page | Last page |