94 replies to this topic

[Kaputnik]

Crappy crap posting bug double postet my post

Edited by Kaputnik, 01 November 2012 - 04:07 PM.

[Kaputnik]

i created a nice workaround topic for this wich automatically deletes the folder on every start, so that will smoothen your gameplay again

take a look:

http://mwomercs.com/...udden-fps-drop/

if your problem is not because of damaged shading files try this workaround:

http://mwomercs.com/...g-certain-ways/

Edited by Kaputnik, 30 October 2012 - 02:27 PM.

[XManBG]

No, it did not worked for me. The problem is independent from the Shaders folder.
Both first and second games are fluently playable with 50-60 FPS!! Every 3th game is at 4 FPS. I have to close the game and start over... It's annoying.

And this nightmare happened just after the Patch before the last one.

Cheers

rgreat, on 29 October 2012 - 04:14 PM, said:

Workaround:

Delete MechWarrior Online\USER\Shaders\ folder.

It will help for a while.

Edited by XManBG, 30 October 2012 - 10:35 AM.

[Mech Wrench]

Im having the same problem. I submitted a support ticket and was told: "this issue was resolved." Leaves me wondering why i had to submit a support ticket then.. ?¿ huh... anyway, hope someone finds a fix, as im still suffering from it regularly.

Processor Intel® Core™ i5-2500K CPU @ 3.30GHz 7.5 Memory (RAM) 4.00 GB 7.5 Graphics AMD Radeon HD 6870 Series 7.8 Gaming graphics 2806 MB Total available graphics memory 7.8

Windows 7 home premium 64bit

[Gwarsbane]

I find I have the 4fps on the forest map. Every time I play on that one map i have really bad fps.

[Chronojam]

I have created a workaround that actually works. Please see http://mwomercs.com/...g-certain-ways/

It is true that there is a chance for the shader cache to cause trouble over time, but it is not the chronic problem that most people think it is. The true issue is that your video card goes all ********, and it's relatively simple to tell it to remain in full gaming form.

If the linked fix works for you, please also confirm that you are using it in the thread I've linked.

[Lufos]

yep, its still on. got an answer from a dev today. he said they are working on black screen and 4 fps bug right now. with more dual core support later.

[CompproB237]

No luck reproducing this yet. Although, I've now twice had the game completely close itself with no error message.

[Tarl Cabot]

Colonel Pada Vinson, on 29 October 2012 - 07:45 PM, said:

This is not being caused by massive memory consumption on my end. my .exe file is fairly small...not the 1.5GB it was before.

This is my understanding of what is occurring.

Need to view the Commit Charge amount, not the normal memory usage.

The commit charge increases when any program is opened and used, and goes down when a program is closed. It will also change when already-running programs allocate or free private virtual memory.

The game or a game utilized driver/file is not freeing up the virtual memory for certain effects/items, and it continues to build til it max out and the game crashes.

And for 32-bit systems this can happen quicker than with 64-bit, and the smaller amount of memory the sooner it can happen.

The coders and Devs for SOE's Everquest had run into this issue several months ago. They have and continue to update the code to force the release of resources but some people tend to run into this issue more than others.

They had a very long thread going gather information, and people keeping an eye on their Commit Charge, what zones it was occurring in more, classes, spells, OS, 32 or 64bit, amount of memory, etc.

Again, have either the task manager up, Performance tab. Using Process Explorer makes it easier to read though.

[Jace11]

Similar, if not same issue. Seems to occur after 7 to 10 games. Definately mech model related as at start of game if I look towards enemy spawn location slow down occurs if I look away, it clear and FPS return to normal. The fact that happens at start of games, when no mechs have died suggests something else model related. Game is perfectly normal for first few games. GFX 460

Edited by Jace11, 30 October 2012 - 05:18 PM.

[Buzz313th]

Ok an update..

I tried every fix posted on multiple threads here. Unfortunately the developers have not taken the time to merge all 6 threads for the same bug.

1. I tried deleting shader cache and rebuilding. Made no difference.

2. Tried installing MSI afterburner and creating a profile that forces stock Vid card voltage, clock speed and throttles the card to a constant rate for both 2d and 3d programs.. Didn't work..

3. I did lower my ingame graphics settings to 50% of what I was normally running which was full boar.. And I noticed that I did not get the bug until 5 games instead of 3. I lowered my setting to minimum ingame and did not get the bug until 9 games... I can repeat this everytime...

Here's my hunch.... and it would be expected if the devs stepped up to the plate ...

I am running a GTX560TI with 1024 mb's of mem. I believe that Tarl Cabot is right. That there are certain effects that are saturating the memory and not getting refreshed... but I believe it is the vid card memory.

At full ingame graphics I can run two matches no problem. The third will bug out to less than 5 fps when I am looking at alot of mechs or alot of effects. A restart of the client fixes it and I can run another two matches no problem. I did not have this bug before the last patch.

I believe that the reason I only last 2 matches before I bug out is because of only 1 gig of vid memory.. I am gonna take a stab at this and say that other people who have this bug will be able to run longer than 2-3 matches at full graphics if they have more than 1 gig of vid memory.

Can some of you who have this bug, please run a few tests.. Set all your graphics at full ingame and see how many matches it takes for the bug to show up.. Report your findings here with the amount of onboard vid memory you have... I hope that my hunch is correct that the ones with more vid memory will be able to run more trouble free matches before the client bugs out..

Thanks

PS... I also submitted a ticket yesterday and have yet to recieve anything other than the automated reply.

Im ok with my 4 mechs from the founders pack, but honestly, I aint gonna even think about buying anything more until this is fixed...

Edited by Buzz313th, 30 October 2012 - 06:32 PM.

[Kawai]

Core 2 Duo E6750 2.66GHz
NVIDIA GeForce 8800 GT, latest drivers
4 GB DDR2 RAM
Win XP SP3 x32
DX9
Screen resolution 1600x900
closest system settings you are using in MWO Options LOW

First two rounds the game is just nice and smooth. 4fps bug shows up every 3-rd match. Well I know my systems specs are low and I could understand the low fps if it would be every time, but no - I can play 2 rounds with high fps and then I get problems. Need to restart the client every 3-rd match. I guess this Commit Charge effect causing this, as I can't see a decent load on my system while bug occurs (30-50% or so).

And yes, if I max out the graphics, I could get the bug problem at 2nd match already.

Edited by Kawai, 30 October 2012 - 06:55 PM.

[Magic Murder Bag]

can't tell if getting this bug or not, mostly cause my FPS is usually glued to the 20s but here's my specs (i think i don goofed on the motherboard though):

motherboard: ASUS P8Z68-V pro/GEN 3(overclocked mode enabled at the moment)
CPU: intel quad core processor sandybridge 27k
16 gig RAM
128 gig SATA SSD card
OS: windows 7 premium 64 bit
DX9
NVidia GeForce 660 Ti

Edited by Magic Murder Bag, 30 October 2012 - 07:20 PM.

[GL Chozo4]

To confirm what 'Buzz313th' said - it is definitely a graphics memory issue.

I find that the bug is also related to graphics memory not being cleared between matches... I sent in a bug report as well with details but here is the clip of my submittal in case it helps others. I normally get 20-30fps at the first match but crawl on the next unless it is the same map as the first match. It is not exactly a system ram issue as that stays nominal if shared video ram is limited but rather a graphics ram issue so for those using high amounts of shared memory for video however, it will become a system memory problem as well. This will effectively exacerbate those with 2gb or less of system ram and make things even more unbearable (resulting in out of memory/memory allocation errors, crash to desktop, etc).


AMD Opteron 1389 [2.9ghz @ 3.65ghz)
ASRock 960gm U3S3
XFX GTX260 Black Edition (2 729/1458/1242)
2x2gb DDR3 OCZ Obsidian Ram
Windows Server 2003 Enterprise SP2/RC2
19" CRT Hyundai ImageQuest P910+ @ 2560x1920 Resolution
Ingame Settings: High/Ultra with PostAA set to OFF

Quote

Issue:
Between matches in Mechwarrior Online, Graphics Memory is not cleared of any cached textures.

Notes:
This is evident whereas on first launch the consumed GFX ram is around ~120mb used and then the needed GFX ram gets used on the first match and there is not a single slowdown or hitch. After a match is completed, the video memory consumed remains at 700-800mb and will not clear. Upon launching the next match, this needed ram for the next map gets used on top of what's already there pushing the used GFX ram consumed to roughly 1200mb (dedicated+shared for overflow) and beyond. Using an 896mb graphics card, this becomes very crippling after only one match as anything beyond caps out the additional system ram reserved for video memory overflow and thus cripples fps down to 3fps or even less. This will keep expanding to consume even larger video-memory allocations on newer cards judging by the upward trend as it consumes all of the graphics ram eventually forcing all GFX memory moves/writes to system memory instead.

Currently, for example - I'm sitting in the lobby after a second match with 873/896mb video memory currently consumed prior to a match leaving only 23mb remaining to handle the next match (and 512mb of shared memory). Needless to say, the last match was completely unplayable compared to ~30fps on the first.

Edited by GL Chozo4, 30 October 2012 - 08:23 PM.

[Buzz313th]

GL Chozo4, on 30 October 2012 - 08:01 PM, said:

I find that the bug is also related to graphics memory not being cleared between matches... I sent in a bug report as well with details but here is the clip of my submittal in case it helps others. I normally get 20-30fps at the first match but crawl on the next unless it is the same map as the first match.


AMD Opteron 1389 [2.9ghz @ 3.65ghz)
ASRock 960gm U3S3
XFX GTX260 Black Edition (2 729/1458/1242)
2x2gb DDR3 OCZ Obsidian Ram
Windows Server 2003 Enterprise SP2/RC2
19" CRT Hyundai ImageQuest P910+ @ 2560x1920 Resolution
Ingame Settings: High/Ultra with PostAA set to OFF

I think you are 100% correct Chozo.

Lets see if the devs can comment on this at all..

It would be nice if others can confirm this.

[Sturmwind]

The constant drop in FPS the longer you play, or the sudden lock-up of the game after playing for about 1hr is at least known to me quite well. However, since it went OpenBeta, as soon as log into the game my FPS go down to 4. Doesn't matter if it's the mechlab or a map, it's completely unplayable now for me. Even when I reduce ALL settings to lowest possible there's NO difference.

I donwloaded, deinstalled and reinstalled the game completely, emptied the shaders cache...you name it. Nothing helps. For whatever reason I can't play anymore and I know for certain that it's not a problem on my end, as I didn't change ANYTHING on my machine and could play fine still last week...

Let's just hope this gets fixed ASAP...

[Akhael]

Exact same problem for me. Before Open Beta I had a 25/30 fps rate, game was running fine, after monday evening (France here), mechlab is 4 fps, game is unplayable.

I have tried almost everything to get things back to normal (except re-installing Win7, which I am considering at the moment).

However, I noticed that when the game is on even Windows is laggy. I mean every effect (buttons lights, windows opening,…) is ssloooww. So I came to the conclusion that somehow the graphic card is involved, it's as if the system is partially ignoring Catalyst (which runs my AMD HD6630M) and trying to run things with the basic default driver (just a guess here as I am not really comfortable with PC architecture so maybe thats not possible and I am mistaken). I still have to see if I can tweak the settings of the card out of Catalyst.

I am running win7 on a mac mini (don't laugh) with an i7 and 4 gb of ram. The game was running just fine before Open beta. Funny thing is I was about to throw some money into the game… well this is not gonna happen until they fix this issue.

Can anyone check that slow Windows issue while under the 4 fps mechlab bug ?

[TigaShark]

I see the issue after several hours and maybe 10-12 matches, after the bug kicks in I see intermittent slideshow//unplayable framerates when enemy mechs are on screen, this happens on maps which were previously fine....restarting the game tends to fix it..



System spec

Windows 8 64bit enterprise edition
Intel Core2 Quad 9550 @ 4ghz
8Gb DDR2 1100
1gb ATI Radeon 5850

Everythings water cooled (CPU/Chipset/VGA) so no issues with temps whatsoever.



Game settings
“Very High” preset
1920x1200 res
Motion Blur OFF
Vsync “ON”

Edited by TigaShark, 31 October 2012 - 02:01 AM.

[Tarl Cabot]

For those who want to track it better, use GPU-Z to track the memory usage for both system and GPU. The drop down boxes allows you to view either the current/max/min usage for both GPU and system memory. It will also log the information.

Here is another one, very small footprint but only works with XP OS
http://www.nuclearplayground.com/



I myself had not looked at the GPU usage/max when the issue pops up for me, as it rarely does unless I have been playing for a long period of time, with over 15+ matches.

[Akhael]

Ok so I followed your advice and installed GPU-Z and guess what ? As soon as MWO client is up and ready (profile loaded, trial mechs ready, etc.) my GPU usage reaches 100 % and my whole VRAM is stuck. And with nothing happening on screen, I mean I am just standing there in the mechlab with the Yen-Lo-Wang staring at me in the central ad and my graphic card is already overloaded !!! Something's wrong here…

Edited by Akhael, 31 October 2012 - 07:59 AM.