Anyone else NEVER lag, never DC, and never crash?
By
Aggelakis,
Posted on: 2012-01-05 13:03 in CoH and CoV General Discussions
I think I've only crashed once or twice since getting my new video card last year. DC'd a few times, mostly due to the fact that I am on wireless and occasionally my router needs to be power cycled (working on getting a new one; no, a hard line isn't possible atm). I never really have any network lag unless others in the house are doing things online that take up a lot of bandwidth (someone watching Netflix, someone else watching YouTube, someone else using XBL). The only real issue I have is right when I load up the game, I get some graphical lag for a few seconds, but then it goes away. I'd been on a break for a few months, and it never did that before, but when I came back a couple of months ago I noticed it. I've tried some suggestions in the Technical Issues forum, but it still does it so I figure it's just AMD and CoH aren't getting along right now.
Keep in mind though, I'm mostly a soloer, and I'm not an Incarnate, so I don't really know how things will hold up in iTrials and such. Based on past experience in TFs, I'd assume I wouldn't have any issues.
Freedom
Blueside: Knight'Hawk, lvl 50, Scrapper
Yellowside: Dark'Falcon (Loyalist), lvl 20, Blaster
That Stinging Sensation #482183
Quote:
I think I've pinpointed when this happened and fixed it today. It was doing its intermittent slowdown thing when something occurred to me. So I went into the launcher and added the -renderthread 1 to the launch properties since I recall using on the direct desktop shortcut before we were forced to use the launcher, and when I started the game back up and played it for a good while longer it was no longer doing the intermittent slowdown thing.
|
Originally Posted by Tenzhi
Since issue 20 or 21, I don't exactly recall when it started, I've been running into intermittent instances where the game starts running like its in virtual molasses for 10-30 secs. It does it in the game, during character creation, and even at the character select screen.
|
Goodbye may seem forever
Farewell is like the end
But in my heart's the memory
And there you'll always be
-- The Fox and the Hound
My computer is a couple of years old and no, I never really have issues with CoH. Mind you, I never really have issues with a lot of things. I even completed skyrim without an issue. I think people just like to keep a lot of software running on their systems and it interfeers with everything.
My system.. (home built, I never budget for the best of the best, but I will do my homework about stuff)
Win 7 x64 Ultimate
8Gb ddr 3 ram
amd quad core at 3.8Ghz
2x 1TB HDD and 1x 2TB HDD
nVidia GTX460 1Gb
Combat Kangaroos, Justice Server. First 50's
Jirra Roo Plant/Storm/Stone/Musculature Controller
Combat Kangaroo Rifle/Energy/Mace/Spiritual Blaster
Kung Fu Kangaroo Martial Arts/Reflexes/Body/Spiritual Scrapper
Tribal Arc Shield/Elec/Mu/Spiritual Tanker
Quote:
This is simply not true. Memory leaks are usually bugs in the game's code, and are not the fault of your computer. A memory leak is caused by the allocation of an "object" in the client software that is never freed in order for the memory to be used for other things. These objects may describe things such as the special effects that are tossed around in a trial, or the characters appearing on your screen, the pieces of the costumes they're wearing, and so on. (Occasionally drivers do cause memory leaks, but wasn't the case this time around.)|
Originally Posted by Lothic
I think about 95% of these crash/lag problems people have with this game are based on the computer they are using to play it, not the game code or servers.
|
If even a few of those memory objects describing the thousands of explosions occurring in a trial are not properly freed, over time these "lost" objects will pile up and your computer will simply run out of memory. It will take longer for a computer with more memory to crash, but since RAM is finite, any computer running a program with a memory leak will eventually crash if you run that program long enough.
If the return value of a failed memory allocation is not properly checked, any kind of error can occur, from a corrupt pigg warning, to a bad instruction trap, to a glitched display, to an audio loop, to a well-controlled crash that solicits a question about what you were doing, to a direct and immediate exit to Windows with no warning whatsoever.
I was experiencing a wide variety of crashes in Lambdas and BAFs after one or two runs, with no clue as to why they were happening. I guessed memory leaks were behind it, but didn't know. After a few months and a few patches the crashes became less frequent and were no longer random, but were limited either to corrupt pigg file crashes (i.e., memory leaks) or explicit out of memory crashes. After another couple of patches the crashes stopped.
I haven't experienced a crash in a couple of months now, and I changed absolutely nothing about my PC: same amount of memory (4 gig), same drivers, same video card.
A month or so after the crashes stopped one of the devs asked in a thread about crashes whether people had seen an improvement. I had and said so.
I haven't seen memory leaks directly mentioned in the patch notes, but things improved after several cryptic "performance improvements" were mentioned.
I'm a programmer and I know how hard memory leaks can be to track down in a live system, where you don't have access to clients that are crashing in the field. So I'm not accusing the devs of shoddy programming; these things happen, especially in programs that have been around for years and have been worked on by dozens of different programmers.
I'm just glad they fixed the problems, because I was getting pretty fed up with crashing in the last 30 seconds of every Lambda I ran...
I have to note something here...
Quote:
I've always heard it to be more of a "wauom" than a "woob" sound.|
Originally Posted by UberGuy
...The clearest indicator of this is when our powers do not recharge even when our client thinks they should have. We click on our power, which is now full-sized and no longer translucent, and we get the "woob" sound with a "recharging" message. Why does this happen?...
|
/Freem.
//carry on.
August 31, 2012. A Day that will Live in Infamy. Or Information. Possibly Influence. Well, Inf, anyway. Thank you, Paragon Studios, for what you did, and the enjoyment and camaraderie you brought.
This is houtex, aka Mike, signing off the forums. G'night all. - 10/26/2012
Well... perhaps I was premature about that whole 'signing off' thing... - 11-9-2012
Quote:
I've been a software engineer for almost 20 years so why don't you tell -me- how computers work... |
Originally Posted by Rodion
This is simply not true. Memory leaks are usually bugs in the game's code, and are not the fault of your computer. A memory leak is caused by the allocation of an "object" in the client software that is never freed in order for the memory to be used for other things. These objects may describe things such as the special effects that are tossed around in a trial, or the characters appearing on your screen, the pieces of the costumes they're wearing, and so on. (Occasionally drivers do cause memory leaks, but wasn't the case this time around.)
If even a few of those memory objects describing the thousands of explosions occurring in a trial are not properly freed, over time these "lost" objects will pile up and your computer will simply run out of memory. It will take longer for a computer with more memory to crash, but since RAM is finite, any computer running a program with a memory leak will eventually crash if you run that program long enough. If the return value of a failed memory allocation is not properly checked, any kind of error can occur, from a corrupt pigg warning, to a bad instruction trap, to a glitched display, to an audio loop, to a well-controlled crash that solicits a question about what you were doing, to a direct and immediate exit to Windows with no warning whatsoever. I was experiencing a wide variety of crashes in Lambdas and BAFs after one or two runs, with no clue as to why they were happening. I guessed memory leaks were behind it, but didn't know. After a few months and a few patches the crashes became less frequent and were no longer random, but were limited either to corrupt pigg file crashes (i.e., memory leaks) or explicit out of memory crashes. After another couple of patches the crashes stopped. I haven't experienced a crash in a couple of months now, and I changed absolutely nothing about my PC: same amount of memory (4 gig), same drivers, same video card. A month or so after the crashes stopped one of the devs asked in a thread about crashes whether people had seen an improvement. I had and said so. I haven't seen memory leaks directly mentioned in the patch notes, but things improved after several cryptic "performance improvements" were mentioned. I'm a programmer and I know how hard memory leaks can be to track down in a live system, where you don't have access to clients that are crashing in the field. So I'm not accusing the devs of shoddy programming; these things happen, especially in programs that have been around for years and have been worked on by dozens of different programmers. I'm just glad they fixed the problems, because I was getting pretty fed up with crashing in the last 30 seconds of every Lambda I ran... |
Having said that obviously any application as large as this game will have some inefficiencies represented by memory leaks or optimizations that are not used that should be. These will show up on ANY computer, but if a computer is capable enough then these potential problems will likely not be noticeable by a user to any significant degree.
The Devs of this game have naturally tried to make this game work reasonably well on a wide range of hardware, but that does not mean they they are going to spend that much time making it run on something relatively out of date. Sure with some amazing effort they might able to get this game running on a i386 machine with 2 Mb of memory. But since the overwhelming majority of people trying to play games on computers in 2012 are using computers that are much more capable than that there's no real reason for our Devs to spend the effort supporting such a platform.
This is why I'll easily stand by what I said. I'm not suggesting that the Devs are purposely trying to screw people with old computers. But by the same token anyone trying to run this game on an increasingly marginal/older machine can hardly reasonably expect to run as problem-free as it would on a bleeding-edge over-capable computer.
Now obviously the Devs are always proactively doing their best to fix memory leaks. There will even be cases like yours where certain areas of game play for given players will actually improve over time due to the corrective coding efforts of the Devs. Despite these efforts the overwhelming percentage of players who play this game will experience problems that tend to get worse, not better. This is a simple function of the fact that peoples' computers are always getting older relative to the ever increasing complexity of the game itself. The game is always going to get bigger and more complex so there is a certain responsibility on the part of the player to realize that their machines will not be able to keep up forever without being maintained and updated as well.
So ultimately good game play is the responsibility of both the Devs AND the players. The Devs do their best to deliver the most bullet-proofed, bug-free code they can. But knowing that no application is ever perfect there is a certain responsibility on the part of the players to use the best hardware they can to make the game run as seamlessly and error-free as possible. Sadly there are a lot of people out there who have unreasonable expectations and get upset when their low end machines fail to handle a game like this well. They incorrectly assign too much of the blame on the Devs when in reality a lion's share of the fault lies with trying to play this game on an inadequate machine. *shrugs*
Loth 50 Fire/Rad Controller [1392 Badges] [300 non-AE Souvenirs]
Ryver 50 Ele� Blaster [1392 Badges]
Silandra 50 Peacebringer [1138 Badges] [No Redside Badges]
--{=====> Virtue ♀
I do lag, rarely ever crash.
Though...got a few new computer pieces, I'm able to run Ultra Mode now 
I have recently had an issue with crashing, once a night, for about 2 weeks. prior to that, never had any technical issues beyond what could be expected.i expect some lag on a full-league rikti mothership raid, for instance, or an itf with multiple mm's, that's just simply being reasonable. but normal lag, not since i got my new comp 2 years ago.
Actually, I should comment again, this time, reguarding my wifes netbook. win 7 , Intel graphics chip, atom processor, 2Gb ram, Acer 5320 (i think), neat little device once you remove the incredible amount of bloatware off the thing. Get this, this little netbook will happily run CoH for hours on end (even on battery, love the battery life of that thing)
Now granted, since its running a gfx chip barely capable of anything 3D, putting the graphics up higher than safemode will cause you to look at a slide show (its basically set to minimum, then up the resolution to its native), yet despite its lack of resources, it too will happily run the game. CoH really isn't all that bad a piece of software 
Combat Kangaroos, Justice Server. First 50's
Jirra Roo Plant/Storm/Stone/Musculature Controller
Combat Kangaroo Rifle/Energy/Mace/Spiritual Blaster
Kung Fu Kangaroo Martial Arts/Reflexes/Body/Spiritual Scrapper
Tribal Arc Shield/Elec/Mu/Spiritual Tanker
Quote:
Except true memory leaks have been uncommon, and platform-specific. That implies they aren't in the core game code, and more likely in areas like the graphics subsystem. Also, I've never seen the game runaway leak and crash by exhausting heap on a system with more than 2 gigabytes of RAM in all the years I've observed it.|
Originally Posted by Rodion
This is simply not true. Memory leaks are usually bugs in the game's code, and are not the fault of your computer. A memory leak is caused by the allocation of an "object" in the client software that is never freed in order for the memory to be used for other things. These objects may describe things such as the special effects that are tossed around in a trial, or the characters appearing on your screen, the pieces of the costumes they're wearing, and so on. (Occasionally drivers do cause memory leaks, but wasn't the case this time around.)
If even a few of those memory objects describing the thousands of explosions occurring in a trial are not properly freed, over time these "lost" objects will pile up and your computer will simply run out of memory. It will take longer for a computer with more memory to crash, but since RAM is finite, any computer running a program with a memory leak will eventually crash if you run that program long enough. If the return value of a failed memory allocation is not properly checked, any kind of error can occur, from a corrupt pigg warning, to a bad instruction trap, to a glitched display, to an audio loop, to a well-controlled crash that solicits a question about what you were doing, to a direct and immediate exit to Windows with no warning whatsoever. I was experiencing a wide variety of crashes in Lambdas and BAFs after one or two runs, with no clue as to why they were happening. I guessed memory leaks were behind it, but didn't know. After a few months and a few patches the crashes became less frequent and were no longer random, but were limited either to corrupt pigg file crashes (i.e., memory leaks) or explicit out of memory crashes. After another couple of patches the crashes stopped. I haven't experienced a crash in a couple of months now, and I changed absolutely nothing about my PC: same amount of memory (4 gig), same drivers, same video card. A month or so after the crashes stopped one of the devs asked in a thread about crashes whether people had seen an improvement. I had and said so. I haven't seen memory leaks directly mentioned in the patch notes, but things improved after several cryptic "performance improvements" were mentioned. I'm a programmer and I know how hard memory leaks can be to track down in a live system, where you don't have access to clients that are crashing in the field. So I'm not accusing the devs of shoddy programming; these things happen, especially in programs that have been around for years and have been worked on by dozens of different programmers. I'm just glad they fixed the problems, because I was getting pretty fed up with crashing in the last 30 seconds of every Lambda I ran... |
I don't think the game has ever had a lot of true allocation leaks, at least not straight-forward ones. I have a feeling many of the "out of memory" problems were actually more subtle heap management issues. I should also point out that when the game client was having "out of memory" crashes more often I observed two things. One: /unloadgfx periodically would forestall the crash indefinitely. A true memory leak related to unfreed handles is unlikely to be addressed in that way. Two: CoH would sometimes crash in a manner that suggested it was out of memory when my system still had lots of physical memory free. That suggested the problem was an allocator problem and not a literal out of memory problem.
[Guide to Defense] [Scrapper Secondaries Comparison] [Archetype Popularity Analysis]
In one little corner of the universe, there's nothing more irritating than a misfile...
(Please support the best webcomic about a cosmic universal realignment by impaired angelic interference resulting in identity crisis angst. Or I release the pigmy water thieves.)
Besides one of Windows blue-moon inexplicable 'Error report!' crashes which have happened since forever, and from time to time whatever the net feels like doing...not in a loooong time. No crashes I can remember, no real lag, except in stuff where everyone is (many GM events etc)
Quote:
|
Originally Posted by Zwillinger
GG, I would tell you that "I am killing you with my mind", but I couldn't find an emoticon to properly express my sentiment.
|
Quote:
|
Originally Posted by Captain_Photon
NOTE: The Incarnate System is basically farming for IOs on a larger scale, and with more obtrusive lore.
|
Notably, "lag" was something of a misnomer for those problems. The term "lag" does usually apply to delayed response times, and the usual culprit is high internet latency. While things like the "valley of lag" did cause delayed responses from the game, what was actually being called "lag" was more accuratly described as "server-side time dialation".
The server was getting bogged down with too much to do, and it wasn't updating things in real time any more as a result. The clearest indicator of this is when our powers do not recharge even when our client thinks they should have. We click on our power, which is now full-sized and no longer translucent, and we get the "woob" sound with a "recharging" message. Why does this happen?
It's the server that determines when our powers should recharge, but it has to let the client know when that is to show that on our power icons. It has to update our client if something causes it to change, such as when we get recharge buffs or debuffs. When time dialation is going on, it lets the client know, and unless something requires it to send an update (like we get debuffed or activate Hasten), that's the end of the conversation until the power does recharge. But the actual elapsed time on the client and the server don't line up, because time is "taking longer" on the server.
This was what we experienced in places like the "valley of lag". The presence of elongated power recharge time was a core element of the phenomena, which means our clients and/or our network bandwidth were not the dominant factors.
Presently, changes made in the run up to CoH:F immensely improved the CoH server side's performance in these areas, making "time dialation" far less common. Things like the "valley of lag" appear to largely be a thing of the past.
Blue
American Steele: 50 BS/Inv
Nightfall: 50 DDD
Sable Slayer: 50 DM/Rgn
Fortune's Shadow: 50 Dark/Psi
WinterStrike: 47 Ice/Dev
Quantum Well: 43 Inv/EM
Twilit Destiny: 43 MA/DA
Red
Shadowslip: 50 DDC
Final Rest: 50 MA/Rgn
Abyssal Frost: 50 Ice/Dark
Golden Ember: 50 SM/FA