Aller au contenu

Photo

memory leak


  • Veuillez vous connecter pour répondre
39 réponses à ce sujet

#26
JironGhrad

JironGhrad
  • Members
  • 1 657 messages

Arpl wrote...

http://en.wikipedia....iki/System_call


All I'm going to say is this: citing a source without sufficient citations to meet wikipedia's limited requirements and lacking any reference material related to the operating system in current discussion proves nothing.  You could have written that article yourself (whether you did or not isn't even really the point).

If this

Wikipedia article: A system call is a request made by any program to the operating system for performing tasks—picked from a predefined set—which the said program does not have required permissions to execute in its own flow of execution.

is correct, then to bypass it all that has to happen is the OS (Windows can and sometimes does) grant the application full permissions. Whether that would be a good thing or not isn't the point of the discussion. It could happen and that might even explain a lot of the oddities associated with the way Dragon Age seems to work on some machines.

#27
Apex Sammoth

Apex Sammoth
  • Members
  • 195 messages

JWT2k6 wrote...

This looks like a good thread to keep up with.
*grabs some popcorn*



I was really enjoying the Tech Banter. Awe....

Anways with all the help on these threads and I have been saying this more than enough on this forum. It is a Bioware /EA problem PERIOD. Nothing more can be said to justify it. The Facts are there are way too many people with problems for it to be there systems. It's just plain logic take away all the Tech savy talk and it still comes down to the developer.

#28
Deathstalker2324

Deathstalker2324
  • Members
  • 9 messages

Apex Sammoth wrote...

JWT2k6 wrote...

This looks like a good thread to keep up with.
*grabs some popcorn*



I was really enjoying the Tech Banter. Awe....

Anways with all the help on these threads and I have been saying this more than enough on this forum. It is a Bioware /EA problem PERIOD. Nothing more can be said to justify it. The Facts are there are way too many people with problems for it to be there systems. It's just plain logic take away all the Tech savy talk and it still comes down to the developer.


Someone give him a cookie.

#29
B33ker

B33ker
  • Members
  • 825 messages
The funny thing is, way back when this whole "longer loading times" thing was posted on the forums, one of the devs even said it was a "memory leak".



Right, wrong or indifferent, if they call it that, thats fine with me as long as it gets fixed.



Semantics, sheesh.

#30
whtnyte-raernst

whtnyte-raernst
  • Members
  • 549 messages

B33ker wrote...

The funny thing is, way back when this whole "longer loading times" thing was posted on the forums, one of the devs even said it was a "memory leak".

Right, wrong or indifferent, if they call it that, thats fine with me as long as it gets fixed.

Semantics, sheesh.


I agree, who cares what anyone wants to call it. We can speculate all day. Personally, I'm not a programmer, I'm a computer forensics/network security guy. But how many people would know the difference between a "memory leak", "buffer over flow" or "poor variable management" *shrugs. I know even my system slows down dramatically after the 4-6 hour play time mark.
I built this computer from scratch, I know everything that's in it and on it! Personally, I think it's setting up a lot of long variables and then not releasing them when their no longer needed. Maybe some programmer thought it might speed up things when you return to an area you just visited and it back fired, who knows.

#31
SobrietyTest

SobrietyTest
  • Members
  • 42 messages

B33ker wrote...

The funny thing is, way back when this whole "longer loading times" thing was posted on the forums, one of the devs even said it was a "memory leak".

Right, wrong or indifferent, if they call it that, thats fine with me as long as it gets fixed.

Semantics, sheesh.



AMEN brother,  but still watching the techies have a forum battle royal over symantics and whoes right and wrong is just to funny to ignore. For those of us non-techies... whatever the cause, the result is extreramly poor game performance for a great many people, we dont care what caused it, we only care that it gets FIXED. the devs tell us what info they need from us in order to figure out the problem and we give it to them. prety simple realy.

#32
Ryukahn

Ryukahn
  • Members
  • 159 messages
I have the same problems during extended periods of play. In the mean time I just don't play it much lately, thus I don't have to worry about the super slowdown and loading. :)

#33
Arpl

Arpl
  • Members
  • 15 messages

JironGhrad wrote...

Arpl wrote...

http://en.wikipedia....iki/System_call


All I'm going to say is this: citing a source without sufficient citations to meet wikipedia's limited requirements and lacking any reference material related to the operating system in current discussion proves nothing.  You could have written that article yourself (whether you did or not isn't even really the point).

If this

Wikipedia article: A system call is a request made by any program to the operating system for performing tasks—picked from a predefined set—which the said program does not have required permissions to execute in its own flow of execution.

is correct, then to bypass it all that has to happen is the OS (Windows can and sometimes does) grant the application full permissions. Whether that would be a good thing or not isn't the point of the discussion. It could happen and that might even explain a lot of the oddities associated with the way Dragon Age seems to work on some machines.

I was only going for the definition of the term, but whatever, it seems you need a lecture on it. That article is talking about a different set of permissions than you're used to, please refer to the chapter on protection and privilege levels in Intel Architecture Software Developer's Manual Volume 3: System Programming. Windows does not (and "cannot" due to builtin security measures) grant these permissions to an application, only drivers can be granted anything above ring2 and they rarely get any more permissions than ring1. (Edit: Don't take my word for it, read the Windows DDK docs)

Applications cannot leave ring3 and must therefore use system calls to interface with the drivers(And by extension, the hardware) and the rest of the system. Since DAO does not run as a driver it obviously needs to use system calls to do things like transferring texture and geometry data to the graphics card, rotating the camera, playing music, and so on. As stated before the cache is written back and invalidated on each system call, so your theory about cache "pollution" from other processes being the main culprit falls flat; the cache is invalidated all the time(It's designed to be!) no matter how many other processes you have poking around.

Modifié par Arpl, 15 décembre 2009 - 12:32 .


#34
CDay09

CDay09
  • Members
  • 11 messages
Yeap, I give up. Being the only one to have posted a sensible, promising start on a solution that has been ignored more than enough...



*Grabs popcorn* Anyone need some white cheddar?

#35
whtnyte-raernst

whtnyte-raernst
  • Members
  • 549 messages

CDay09 wrote...

Yeap, I give up. Being the only one to have posted a sensible, promising start on a solution that has been ignored more than enough...

*Grabs popcorn* Anyone need some white cheddar?


It really is a silly argument unless one of us wants to get in there and rip apart code that probably had over a dozen different programmers fingers in the pie. Something we would probably get arrested for :ph34r:

#36
JironGhrad

JironGhrad
  • Members
  • 1 657 messages

Arpl wrote...

JironGhrad wrote...

Arpl wrote...

http://en.wikipedia....iki/System_call


All I'm going to say is this: citing a source without sufficient citations to meet wikipedia's limited requirements and lacking any reference material related to the operating system in current discussion proves nothing.  You could have written that article yourself (whether you did or not isn't even really the point).

If this

Wikipedia article: A system call is a request made by any program to the operating system for performing tasks—picked from a predefined set—which the said program does not have required permissions to execute in its own flow of execution.

is correct, then to bypass it all that has to happen is the OS (Windows can and sometimes does) grant the application full permissions. Whether that would be a good thing or not isn't the point of the discussion. It could happen and that might even explain a lot of the oddities associated with the way Dragon Age seems to work on some machines.

I was only going for the definition of the term, but whatever, it seems you need a lecture on it. That article is talking about a different set of permissions than you're used to, please refer to the chapter on protection and privilege levels in Intel Architecture Software Developer's Manual Volume 3: System Programming. Windows does not (and "cannot" due to builtin security measures) grant these permissions to an application, only drivers can be granted anything above ring2 and they rarely get any more permissions than ring1. (Edit: Don't take my word for it, read the Windows DDK docs)

Applications cannot leave ring3 and must therefore use system calls to interface with the drivers(And by extension, the hardware) and the rest of the system. Since DAO does not run as a driver it obviously needs to use system calls to do things like transferring texture and geometry data to the graphics card, rotating the camera, playing music, and so on. As stated before the cache is written back and invalidated on each system call, so your theory about cache "pollution" from other processes being the main culprit falls flat; the cache is invalidated all the time(It's designed to be!) no matter how many other processes you have poking around.


Now see, that's EXACTLY what I wanted.  And honestly, I do actually understand it but nothing is more irritating than having someone come along and say "you're wrong" without any evidence to prove that they're anything other than the insane patient sneaking out and rediagnosing every patient in the doctor's office. Thank you for "proving" that my attempt at explaining it in a way that made more sense to the average user was flawed.  Its quite unfortunate (but true nevertheless) that some things can't be explained without a sufficient working knowledge of certain technical ideas (which holds true for a number of fields).  I applaud your efforts, Arpl. Image IPB

#37
Axe_Murderer

Axe_Murderer
  • Members
  • 279 messages
I may as well throw a monkey wrench into this argument. I'm running mine on a single core CPU and am seeing the slowdown problem so I doubt it's strictly a multi-core hardware issue. I realize I'm underspec for the game, but it runs surprisingly well anyway. My frame rates are easily above 20, until the performance degradation problem starts to manifest. The area load time is just one of the issues I am noticing when it occurs and it is the most severe of the symptoms by far. My frame rate drops significantly (maybe gets cut in half down to about 10) when it happens as well, but the severity of it varies wildly and is definitely dependent on the area I'm in, and my input devices become a little sluggish too (response time drops slightly yet noticably).

Intel Celeron 2.93 GHz single core
3 MB RAM
Intel Mobo
nVidia 7800 GSOC 256 MB
Win XP Pro SP3

What I've seen about it that surprised me was that while in combat everything runs smoothly. With my rig I expected all those opponent AIs and party tactics and VFXs to slow things to a crawl -- but not the case. In fact performance drops hardly at all during combat I have much more problems outside of combat just moving around and running convos. Post-combat is where I see the lag starting most often.

The slowdown can happen gradually over a long period of time (like maybe 1-2 hours) of light action play, with load time and frame rates eventually becoming unplayable until I restart. Or (most often) right after a particularly populated combat event with maybe 10-20 opponents it BAM starts lagging severely directly after the combat mode ends and my party stows their weapons, but only after playing for at least 15 minutes or so. If I save just before a big combat event, when I reload it and fight, no problem...if I've been playing awhile and run thru the big battle, as soon as its over the performance degrades sharply and stays that way until I reset. Although it doesn't happen every time a big battle ends, when it does happen, it usually starts right after combat and especially big combats. Those travelling encounters you get with about a dozen darkspawn are a major culprit for triggering it. It can take 2 hours to start happening or it can take 20 minutes, I've not seen much correlation between the time I've been playing and when it happens really...never as quick as like 5 minutes, but for me certainly it's started happened in less than 30 minutes before and other times as long as maybe 3 hours or more.
 
Unexpected part to me is that the combat itself with all that action runs fine even when the degradation is happening before combat starts, the slowdown doesn't show until everything has quieted down...performance actually gets better when the combat starts and returns to the lag when it ends. Convos are hit by the slowdown pretty severely, not during all the fancy animations and lip synching and VOs, no not there...rather, waiting for the list of responses to appear once the dialog lines have been delivered, and waiting for the convo to shut down at the end is what takes forever.

Another thing I noticed is that I must completely shut down the game all the way back to Windows to fix it. If I just quit and reload, it will still be lagging. This is more indicative of a memory leak than a hardware issue, but I didn't investigate enough to see if memory was actually leaking -- it sure behaves like a memory leak IMO.

Of course it could very well be different for me considering my rig is below recommended specs. I will point out again however that when the lag isn't happening my machine runs the game at a perfectly acceptable rate even in every single place where I experienced heavily lagging. There is also something about the Denerim market area right around the market tent in the center of it that makes my frame rate drop by about 30-40% every time I go there no matter when I do, but it is only around that one place -- once I run by it things pick up again.

Modifié par Axe_Murderer, 15 décembre 2009 - 02:34 .


#38
CDay09

CDay09
  • Members
  • 11 messages
Axe, if your rig is below the spec, that leaves me reason to believe that it may be overheating that's causing your problem as it was causing mine. NVidia cards will try to keep your GPU below 120 degrees Celsius, and when your computer hits that temperature, MASSIVE drops in fps occur. Try lowering your GPU with the tips I gave on the first page and tell me if that works.



Also, in an other topic, REH1967 may have found an alternative cure if it is not the temperature that is causing the problem. http://social.biowar...433311/1#448778 Take a look at that as well.



Good luck to you friend.

#39
Foufoug

Foufoug
  • Members
  • 24 messages

whtnyte-raernst wrote...

You know guys, you have a lot of peoples eyes glazing over here.
I understand what your saying, but then, I have a degree in computer forensics. 99% of the people on here don't care if it's a memory leak (I know, it's not a memory leak is one programs virtual sandbox corrupting another programs memory space) or buffer overruns, or bad variable handling and not releasing, all they care is the game bogs down! And right now, the best solution is: If you've been playing for 6 hours and it's slow loading, shut the game down and restart!


+1 and stop the useless troll fight :s  we really dont care if it memory leak related or not, all we want is a fix -_-

My mind is the game badly manage memory.

#40
LeraG

LeraG
  • Members
  • 173 messages
Tried all the fixes suggested and none of them worked for me!



My system:



Alienware Aurora



Processor: 1

CPU_Name: AMD Athlon™ 64 X2 Dual Core Processor 5200+

CPU_Manufacturer: AuthenticAMD

CPU_Caption: x64 Family 15 Model 107 Stepping 2

CPU_Version: Model 11, Stepping 2

CPU_ProcessorId: 178BFBFF00060FB2

CPU_CurrentClockSpeed: 2700MHz

CPU_AddressWidth: 32Bits

CPU_DataWidth: 64Bits

CPU_SocketDesignation: CPU 1



System Memory



Memory Module: 1

Mem_Capacity: 1024MB

Mem_BankLabel: DIMM #1

Mem_Type: DDR2

Mem_Frequency: 772 MHz

Memory Module: 2

Mem_Capacity: 1024MB

Mem_BankLabel: DIMM #2

Mem_Type: DDR2

Mem_Frequency: 772 MHz

Memory Module: 3

Mem_Capacity: 1024MB

Mem_BankLabel: DIMM #3

Mem_Type: DDR2

Mem_Frequency: 772 MHz

Memory Module: 4

Mem_Capacity: 1024MB

Mem_BankLabel: DIMM #4

Mem_Type: DDR2

Mem_Frequency: 772 MHz



BIOS Information



BIOS_Manufacturer: American Megatrends Inc.

BIOS_Name: BIOS Date: 08/22/08 Ver: 08.00.14

BIOS_Version1: 1301

BIOS_Version2: ALWARE - 20080822



Motherboard Information



MB_Manufacturer: ASUSTeK Computer INC.

MB_Product: M3A32-MVP DELUXE



Video Adapter

Video_Caption: NVIDIA GeForce 9800 GT

Video_AdapterRAM: 512MB

Video_DriverVersion: 7.15.11.7500



Hard Drive(s)



ST3250410AS ATA Device

Disk_Model: ST3250410AS ATA Device

Disk_Size: 250GB

Disk_TotalHeads: 255

Disk_TotalCylinders: 30401

Disk_TotalTracks: 7752255

Disk_TracksPerCylinder: 255



Optical Drives / Other Storage



HL-DT-ST DVD-RAM GH22NP20 ATA Device



Operating System: Microsoft Windows Vista Home Premium





DirectX Version: 10



Just today I played for only 30 minutes and load times were 5 minutes. Nothing is running in the background at all. Defrag, affinity fix, performance set at high, no power saving is set either.



Probably a coincidence but after waiting for minutes I start clicking my mouse down by the word "Loading" and then hit the enter key...lol after a few seconds it loads.



Any other fixes to try?