As this issue seems to have been resolved with new drivers I'll mark this thread as solved now.
in my oppinion it is the same issue, people just have different outcomes. That soundloop (caused because game crashed due to memory getting filled and last audio gets looping) actually comes with some kind of error, i've never seen it, but pressing ENTER closes bf4. So if they fix THIS, whats most likely caused by the leak then we should be ok 🙂=
Logical? no ?
@fetsnage wrote:
in my oppinion it is the same issue, people just have different outcomes. That soundloop (caused because game crashed due to memory getting filled and last audio gets looping) actually comes with some kind of error, i've never seen it, but pressing ENTER closes bf4. So if they fix THIS, whats most likely caused by the leak then we should be ok 🙂=
Logical? no ?
Actually not.
If you look at your memory when soundloop crash occours it's not using more memory. Not in the normal memory and not in the non paged pool which this thread is about.
Also, the non paged pool area fills up your memory in a way which makes it not able to release it which means you have to restart your computer. All you need to do with the soundloop crash is press ctrl+alt+delete and close BF4 which makes you able to play again.
The Soundloop crash can happen right after you start the game, the NPP memoryleak mainly happens after mapchange.
The NNP memoryleak also has been there since beta and this new soundloop crash has only been there since Feb. 13.
The NNP memoryleak has specific hardware on which it doesn't work, the soundloop crash has not.
Lastly, the soundloop crash only affects the maps: Dawnbreaker, Parcel Storm, Operation Locker and Metro - the bug in this thread affects all maps.
So, to sum up:
No, it's not logical that it's the same, it's illogical.
The soundloop crash and the NPP memory leak are in almost any way totally different as bugs go. The only thing they really have in common is that they both affect BF4
Yup, it's a false equivalency. Just because two things have similar "effects" does not mean they have the same cause. That's the same way homeopathy pretends works. Besides, as Carbonic nicely summarised, in this case the two bugs are entirely different in both cause and effect so there is no logical reason to make a connection between the two, they don't even have the same "effect" other than preventing the game working.
As I've seen people reporting getting the leaks on that or that browser, I've done some testing and I can confirm that for me the mmemory still leaks on all of these:
- Maxthon v4.2
- IE v11
- Chrome v32
Also, I'm not really sure what Alt+Tab-ing does really, as people say that they get the leak if they don't alt+tab, I actually get the leak a couple of minutes after I alt+tab out of the game.
Any thoughts?
I had this problem before the soundloop crash but not from day one. I could play about 1 map untill crash time. Switching to Mantle solved the soundloop crash, but instead I got the NP pool Memoryleak again.
I tested the game on 3 systems on the soundloop and NP pool memoryleak (my work system was just to test it, so I still can't play the game, since my employer doesn't like it when it's used for games ..it's at the office 😉 )
Soundloop:
- Desktop 1(private): Windows 8, (and 8.1 after update) intel I5, 12 gb ram @1600 (or @ auto), ssd HD, Ati HD 7800. Runs perfectly for about 10 minutes then freezes. need to terminate in taskmanager.
- Desktop 2 (Work): Windows 8.1, Intel I7, 18 GB ram @ auto ssd HD, geforce gtx 770. Runs perfectly for about 2 minutes then freezes. need to terminate in taskmanager.
- Laptop takes forever to load a level and that on LOW grafics and after 10-15 minutes the entire level (and opponents) is visible.(all OK since it's just a Laptop) left it running for a few hours and ran arround a bit because playing isn't an option with those grafics. But it never crashed.
Np Pool:
- Desktop 1(private):
Windows 8, (and 8.1 after update)
intel I5
12 gb ram @1600 (or @ auto),
ssd HD,
Ati HD 7800.
Runs perfectly for about 20-30 minutes then stutters and crashes to desktop.
- Desktop 2 (Work):
Windows 8.1,
Intel I7,
18 GB ram @ auto,
ssd HD,
geforce gtx 770.
Runs perfectly
- Laptop Dell
Windows 8.1
Intel I5,
8GB Ram,
ati graphicscard 2 GB (don't know wich one)
Takes forever to load a level and that on LOW grafics and after 10-15 minutes the entire level (and opponents) is visible. ( all OK since it's just a Laptop) left it running for a few hours and ran arround a bit because playing isn't an option with those grafics. But it never crashed.
Seems that my "logic" still was correct 😉. New (todays) patch seems to have fixed the leak...
+ new catalyst (14.2) rolling out. Featuring:
*Catalyst 14.2 Beta resolves hangs and stuttering for the Mantle codepath in BF4.
*It enables multi-GPU frame pacing in BF3 and BF4 on non-XDMA configurations.
@fetsnage wrote:
Seems that my "logic" still was correct 😉. New (todays) patch seems to have fixed the leak...
+ new catalyst (14.2) rolling out. Featuring:
*Catalyst 14.2 Beta resolves hangs and stuttering for the Mantle codepath in BF4.
*It enables multi-GPU frame pacing in BF3 and BF4 on non-XDMA configurations.
1. Those are NOT the patch notes. These are the patchnotes:
Feature Highlights of The AMD Catalyst™ 14.2 Beta V1.3 Driver for Windows®
Known Issues
2. Personally, my leak dissapeared/not appeared since after 10 days after it appeared due to the patch which means your statement was not true in my case.
The AMD driver update didn't fix anything here. HD6950 still same probs. Only the sound loop freeze disappeared through the latest 25.02.2014 patch for bf4. At first I thought the leak was fixed then a round Dawnbreaker and Siege of Shanghai came after map change and my NP pool was at 5 gb when stuttering occured and killed my fun.
"Seems that my "logic" still was correct "
You keep using that word "logic". I'm not so sure you understand what it means (insert Princess Bride reference here)
The memory leak is not specific to Mantle, and while it is possible that the introduction of some mantle code by DICE or AMD contributed to the issue, that has not been found or confirmed by anyone. This issue has been happening well before Mantle was even a thing, and it happens for cards that don't even support Mantle. Your "logic" was also trying to tie together the soundloop bug and the memory leak, yet there is no mention of that and you still claim to be correct? It's great the people in the community help out and contribute, but you are just confusing the issue.
Anyway, it will be great if the memory leak is actually fixed by the new AMD drivers. Personally I can't find out because the sound loop bug kills my game before the memory leak has a chance to raise its head.
