Micro freezing
Ghosthree3
Join Date: 2010-02-13 Member: 70557Members, Reinforced - Supporter
Join Date: 2010-02-13 Member: 70557Members, Reinforced - Supporter
There's been many threads about this, but none were on the first page so let's make another one, just for me.
I've uploaded a video showing what I'm talking about here, www.youtube.com/watch?v=XBS8xsDkctU
It seems to be happening more and more frequently but doesn't happen on the same maps/spots or w/e to point to a map issue. Has happened in a lot of the recent patches, but has only become seriously bad since 246 or so.
The game is quickly become unplayable.
I've uploaded a video showing what I'm talking about here, www.youtube.com/watch?v=XBS8xsDkctU
It seems to be happening more and more frequently but doesn't happen on the same maps/spots or w/e to point to a map issue. Has happened in a lot of the recent patches, but has only become seriously bad since 246 or so.
The game is quickly become unplayable.
Comments
Id say it just freezes for a few seconds and then runs again.
More into detail. Any clues what r_stats, net_stats and profile show when this happens?
Go to C:\Users\<username>\AppData\Roaming\Natural Selection 2 and delete the cache folder.
The full details or something in particular?
EDIT: Also I will try to delete the shader cache.
Also, paging @IronHorse to this thread! Help me bro.
Yea the second sets of screenshots are typical of spawning. It was optimized but never past the typical "hitch" point. But its universal and more of a bug report..
The first set of profile shots are really interesting, and your video is great representation of symptoms.
First off: i'd delete everything shaders and recompile all those bad boys. Folders can be found in the hidden %appdata%\Natural Selection 2\ folder under "cache" - as well as in your NS2 install directory in steam, \Steam\SteamApps\common\Natural Selection 2\ns2\ you will find a "shaders" folder. Delete both this shader folder and that cache folder.
Secondly: Install the DXSETUP.exe found in \Steam\SteamApps\common\Natural Selection 2\redist\DirectX
Thirdly: Ensure latest GPU drivers are installed (duh)
Fourthly: Disable any antivirus or overlay program, like teamspeak or msi afterburner, any performance or monitoring software.*
Fifthly: *Ensure they these will remain off from a reboot by going to start > Run > type "msconfig" without quotes > Startup tab > disable any and all things that are not manufactured by microsoft or are audio /cpu/ display drivers. Do not leave things like logitech or anti virus checked. - DO THIS STEP AT YOUR OWN RISK
Sixly: Restart your computer and test.
Seventhly: When you attempt to reproduce after the 10 minute compiling of new shaders, and are in the ready room, i want you to type r_loading true (could be 1 instead) in the console as well as loadtimes true.
If you dont experience the hitch, awesome that was it - but if you experience it:
As soon as you reproduce that hitch, type something in chat like "got it" quickly, and now exit ns2. (do not re run it)
Check your log.txt in your hidden appdata folder and post it in here.
What i believe is occuring is: Your files are being rewritten and those profile shots show you compiling and managing shaders on the fly.. causing hitching. So i believe the two are connected. Re compiling new shaders may only work for a while, as a result.
Second step done.
Third step already done.
Fourth step, don't have any of that except the antivirus (which DOES do on the fly scanning (relevant?), I'll turn it off next time I play.
Fifth step, don't need to touch msconfig as I know what I have running, and it ain't crap like that.
Six, ugh, I hate rebooting, it's been 8 days (no it isn't because I haven't rebooted, it happened before it, and before the one before that). I'll do it though.
Seventh, I will do, but these are for log purposes only right and aren't actually a "fix"?
If the problem is they are being rewritten, can I read only all the files o_O?
Also, I have a friend that experiences the exact same issue (at least visually it seems that way), reduces chance of hardware cause?
PS. IronHorse delivers.
EDIT: Also I want to get a profile image of a better freeze like the one in the video where I kill the skulk.
EDIT: I should add that as a commander (I'm a commander for my team so I play it a lot) I don't seem to get the freezes, or they are so small and so few that I don't seem to get them.
You shouldn't be compiling shaders while playing - and given your HD issues of files being corrupted, i'd wager that this is what is occuring.
But the other steps like AV realtime scanning could also be causing these sorts of hitches.. and thats what the log will tell me.
Just narrowing it down
EDIT: I verified to get the shaders back, since they wouldn't get created.
Also mod please merge my posts, I feel bad but need to bump
Glad to see that fixed it and was the cause!
I would definitely say that your two problems are linked now..
https://dropbox.com/s/y4c2r07lr5va52c/log.txt
Would have posted directly but it was huge.
EDIT: I should add I'm posting this because removing the bad memory stick I ended up having didn't help, files could still be overwriting themselves but it isn't because of the RAM (or it better not be).
EDIT2: Well OC wasn't the issue for sure, because like stated in the other thread, I'm now at 4.6ghz, same problems, whatever it is it's getting worse, can't play for more than 2 seconds without a freeze. I might just reformat (also I don't seem to freeze as a commander, just as a first person player).
EDIT3: If it wasn't clear, by freezing that often, the game is now unplayable and I'm lucky I'm a commander for my team and not on field.
EDIT4: I've put the RAM that was reported bad (I'm not even sure if it's the RAM anymore, everything went to shit so I think the mobo has become damaged) back in to see if it improved system performance at all, because frankly 2 damaged addresses on one stick that may have been causing small file writing damages that occurred infrequently wasn't worth the huge loss in performance it may have caused. Will update within the next few minutes to say whether that put NS2 performance back where it was before (freezing but not every 2s).
EDIT5: Putting the damaged stick back in relieved the issue, played for 3 full minutes without a freeze (checking usage I was at 4.0gb used so it must have been an available memory issue). Unless something unbelievably lucky has happened however, the original problem must surely still exist.
EDIT6: Problem still exists, but hasn't occurred so often.
Of course you have to do the boring memtests for hours upon hours to truly test if its the module or the motherboard/slot.
It is quite possible the mobo itself received damage and I will throw in a oddbal here. I never truly overclock myself so I am very rusty on the basics, so correct me if im wrong.
But overclocking in various ways can increase heat on other components then just the cpu.
So havent other chips on your mobo received damage due to overheat? More expensive mobo's like mine dont have a mere fan on the other cheaps but whole cooling attachments to attach huge coolblocks, watercooling and stuff to prevent this problem.
Also, and I am very rusty on this one, I remember that certain pci lanes can share stuff with the cpu or clock or something. Basicly, raise one, also overclocks the pci-e lane sortof.
This would also mean your vidcard may be pushed harder then it can manage.
Its what I remember, but the more recent OC folk are free to prove me wrong.
Also it IS possible that damage has happened due to the OC, however since the chip ran at around 50-60 max during a prime95 stress test I find it unlikely that it over heated, also the timing just seems too weird, it was working fine, I wipe my bios, now I can't get it working? I dunno, something changed and it seems less likely to be hardware than software. However it is possible that during one of the many times I was swapping RAM sticks in and out that the board sustained damage (especially when pushing them in, makes the computer tip if I don't hold the other side lol, maybe bent the board?).
It is unlikely that the problem can be further diagnosed until I have new RAM unfortunately, however if anyone has an idea, feel free to post it.
You could try to reflash your bios. I severely advice to do this with 1 working ram module. Dont forget, it loads the bios into memory before flashing it, so far I know. Who says it didnt load it in the broken part?
You didnt really awnser my question about chips like the northbridge/southbridge becoming to hot, or the pci-e slot/card due to overclock. (although its more for you to test)
Unless you didnt ground yourself the board shouldnt just take damage from swapping memory.
I already flashed my bios, see why my OC is no longer in effect at 5.0ghz above. Again, need new RAM.
Yeah I don't know about the northbridge becoming hot or not (iirc the southbridge is deprecated in newer mobos and no longer exists).
I took great care to ground myself.
If so, yank it out and reflash.
I just went through 3 days of nightmare bios flashing and cmos resetting and bad memory sticks myself...
Ended up narrowing it down to a soldered bios chip that was corrupted. UGH.
Bought the replacements today.
I assume you did set all your bios settings after the reset to the ones you need?
And this @ironhorse, is why I have a hihabyte mobo as they have a grand total of 2 bios chips.
Yes im paranoid with that.
Thats impossible in Ghosthree3's case, as tests have already shown problems beyond the game. Like memtest reporting bad memory.