Jump to content


Photo

General instability


27 replies to this topic

#1 Aizen Teppa

Aizen Teppa
  • Members
  • 206 posts

Posted 08 April 2012 - 05:52 PM

Hello again. I must admit. I was drolling like mad when 1.2 was out. And then slowly I started to turn on 1.2. I have many problems which were non-existent with version 1.1.

Primary 3 problems:

a. unresponsive ships
b. freezing of the game when loaded.
c. lag when AI moving units

Ad. a.

When I was running 1.1 I had no problems at all with 250 vs 250 ships space battles. Now after certain point I get crap load of problems with 80 vs 80. And all of that was running on an old 8800 GTS640. Never, not a single time I had problem with unresponsive ship(s) under 1.1. Played GFFFA, played smaller scenarios. All the same. Galactic Lag was immense, but space battles were very smooth. Probably my PC is too good for 1.2. Ugh...

Ad. b.

Today played a lot, many hours straight (Op. Skyhook). At some stage again I've got unresponsive ships bug. But more importantly saved games freeze instantly or couple seconds after reload. :angry: Need to kill process.

Ad. c.

Guessing that b. is the result of AI moving units around. Lag is probably even worse than in 1.1 (bigger fleets) with that respect. Nothing changed (at best). If I save game while AI is moving units I'm 99% certain that after reload game won't work

Edited by Aizen Teppa, 08 April 2012 - 05:53 PM.


#2 evilbobthebob

evilbobthebob

    evilbobthemapper

  • Project Team
  • 2,304 posts
  • Location:USA
  • Projects:Phoenix Rising Maps
  •  Phoenix Rising Mapping Lead

Posted 08 April 2012 - 06:05 PM

There's not much we can do about most of these issues. I'm personally aware of the galactic map lag when the AI is making decisions, but there isn't anything we can do about that unless you want really stupid AI. I've never experienced the bug where ships become completely unresponsive (I assume you mean fleet movements?) but we think that galactic fleet population costs may alleviate the issue.

Phoenix Rising, head of mapping. Thanks to everyone who got us to the position below!
Posted Image


#3 Phoenix Rising

Phoenix Rising

    Beyond the Impossible

  • Petrolution Staff
  • 6,509 posts
  • Projects:Phoenix Rising
  •  Mod Leader
  • Division:Petrolution
  • Job:Mod Specialist

Posted 08 April 2012 - 11:31 PM

I've never had the unresponsive ship bug on any version of the mod :unsure:. It must be system-related - I'm still on XP.

I'm further perplexed by your issues with space tactical: we didn't add much new content to space in v1.2 at all :sad:. In fact, the pop cap was lowered by 16 points for space. There's more optimization to do still, so it should improve, even if I can't explain it.

Galactic AI perceptions should not be significantly more CPU-intensive than vanilla - still O(n). If I understand it right, the infrastructure equations should only be evaluated once every 30 seconds or so (depending on what is built) per AI planet.

Edited by Phoenix Rising, 08 April 2012 - 11:32 PM.


#4 Aizen Teppa

Aizen Teppa
  • Members
  • 206 posts

Posted 09 April 2012 - 11:20 AM

Very sad with that. EaW/FoC engine is really too primitive for PR. Have you asked Lucas Arts/Petroglyph for some .exe improvements (like 2GBs+ memory support for starters), maybe another patch to beef-up a bit EaW/FoC environment? I know that scripting isn't the cure every time. Often fault lies in the exe/dll/whatever other file. Really gutted. Had to forfeit 4th campaign - this time Operation Skyhook. Returned to earlier saves, and instead unresponsive ships during tactical battles I've got unresponsive ships in the galactic view after conquering all planets I did previously - existing fleets are movable and OK, everything newly built is stuck (from probes upwards). Really my heart is bleeding. PR is by galaxy length best SW mod for EaW/FoC. I want to play, but every game ends exactly the same. :running away crying:

I tend to agree that too much ships in the galaxy may cause problems. I have problems with AI spamming just absurdly ludicrous amounts of bombers. Me personally, I rarely building more than 30 fighters, but AI is hitting me regularly with 200, 300, once nearly 500 :blink: ATR mark 7 (+plethora of Mon Calamari cumbers, Bulwarks, frigates and corvettes). For that kind assault I need at least 150 Tie-Defenders which is simply impossible to have at every planet.

While in tactical view unresponsive ships bug hit me even if I have one SSD (Sovereign) + 10 fighters for protection. At the beginning there is no problem, but when AI deploy 100-300 ATR bombers doh, with every second I lose ships which are controllable.

I will try to run PR in compatibility mode XPSP2 or W2000 with all themes turned off. I know it may reduce performance, but I'm not concerned about performance. I'm concerned about (at some stage) broken gameplay and enjoyment. Will bump also tactical fleet size from 80 to 150/200/250. Maybe more ships will actually help?! The more the merrier, I wish! (...or wishful thinking :sad2: )

Hmm I must obtain WinXP 64 bit OEM somewhere if that is even possible. I won't install it now, because I have too much work related things, but it is good to have an option. Yes, I'm that desperate for PR1.2 to work! :evgr:

Edited by Aizen Teppa, 09 April 2012 - 11:59 AM.


#5 astrofan

astrofan
  • Members
  • 5 posts

Posted 09 April 2012 - 12:43 PM

I've never encountered problem A; I am running 7 not xp.

The game did freeze when loading, but after I had removed the few files which were mentioned in another thread, it worked fine. However, I do get some random exceptions now and then in the middle of the game (about two or so, so far)

There is some lag in galactic mode, which I expected, but I definitely see at least a small improvement in speed over 1.1; performance seems to fluctuate greatly, sometimes faster, sometimes slower. One hiccup in galactic mode was that suddenly Rothana wasn't able to research anything in the middle of the outer rim campaign; when I click on something to research, it goes into the queue and immidiately returns back to the bottom bar.

My tactical battles (for 1.1 and 1.2) have been smooth from the start. However, something happened in one space tactical battle to make my bottom bar disappear for its entire duration, which caused some headache when trying to manage my forces.

So overall, the mod for me is stable, but just has some hiccups every now and then.

#6 evilbobthebob

evilbobthebob

    evilbobthemapper

  • Project Team
  • 2,304 posts
  • Location:USA
  • Projects:Phoenix Rising Maps
  •  Phoenix Rising Mapping Lead

Posted 09 April 2012 - 12:48 PM

The research bug is something we know about.

The tactical interface can be disabled or enabled with the 'i' key. That can be useful for the zoomed-out overview mode because it allows you to turn the bar back on.

Phoenix Rising, head of mapping. Thanks to everyone who got us to the position below!
Posted Image


#7 Phoenix Rising

Phoenix Rising

    Beyond the Impossible

  • Petrolution Staff
  • 6,509 posts
  • Projects:Phoenix Rising
  •  Mod Leader
  • Division:Petrolution
  • Job:Mod Specialist

Posted 09 April 2012 - 05:42 PM

Aizen, I suspect we can only continue to make improvements and hope it clears up in a future version. Strategic population for ships, given this round of bug reports, is likely inevitable.

If you could report the circumstances of any exceptions, that would be most helpful.

#8 Aizen Teppa

Aizen Teppa
  • Members
  • 206 posts

Posted 10 April 2012 - 04:03 PM

Well two things which I noticed when errors hit.

Tactical battle: when if I starting to lose control over my fleet (or no control from the start which is result of bugged units in Galactic View) at some stage (longer the battle higher probability) I always get exception error like this :


Exception in thread 624 - Main Thread
Exception code EXCEPTION_ACCESS_VIOLATION
	The thread tried to read from or write to a virtual address for which it does not have the appropriate access.
	Attempt to read from address 000000A0
Exception occurred at 008E4880 - Unknown code pointer
Stack walk:
  0069B357 : Unknown
Register dump:
  Eip:008E4880 Esp:0028F45C Ebp:1A483190
  Eax:00000001 Ebx:1A5A7768 Ecx:2546B0D0
  Edx:4854F8C4 Esi:42BA7040 Edi:00000000
  EFlags:00210202
  CS:0023  SS:002b  DS:002b  ES:002b  FS:0053  GS:002b
Bytes at CS:EIP (008E4880): 8B 8F A0 00 00 00 85 C9 74 09 E8 C1 37 DB FF 84 C0 75 6F 8B 85 38 01 00 00 85 C0 74 1A 50 6A 01 8B CE E8 39 FD D3 FF 8B 8D 38 01 00 00 51 6A 01 8B CF E8 29 FD D3 FF 6A 00 57 56 8B CB E8 EE F5

------
+ plenty of gibberish below

Guessing that overwhelming majority of exceptions is result of swfoc.exe inability to access more than 2GB of memory. At some stage swfoc.exe (even when using well below 2GB of RAM) hit the buffers with some system process which refuse access and GTFO! :p In one sentence: very poor memory management and mapping.

In Galactic View Mode when new ships are unmovable I did noticed that AI do NOT move units at all. While existing (old) fleets working fine, if you put a stack of "old" ships on new ones which are unmovable, and try to move units to different location, only "old" units will move. And in time (if I decide to play with bugged units) game just simply hangs. No warning, no exception error message. Just everything stops. Ctrl+Alt+Del> Kill Task

Like I said I never had something like that with 1.1 even during epic space battles (250 vs 250 or 300 vs 300). I was running XP 32b and -mostly- Vista 32b at that time. Need to start yet again in Compatibility Mode.

#9 evilbobthebob

evilbobthebob

    evilbobthemapper

  • Project Team
  • 2,304 posts
  • Location:USA
  • Projects:Phoenix Rising Maps
  •  Phoenix Rising Mapping Lead

Posted 10 April 2012 - 04:49 PM

You say that the exception is always the same...is it always the same hex code of 008E4880 - Unknown code pointer?

Phoenix Rising, head of mapping. Thanks to everyone who got us to the position below!
Posted Image


#10 Aizen Teppa

Aizen Teppa
  • Members
  • 206 posts

Posted 11 April 2012 - 05:45 AM

Hmm if I had to guess I will say yes (except hex code) . Frankly I never paid much attention after reading fragment about access violation. But I will check next one. Saved copy of last one, now need to wait for another, oh well....

BTW: just want to ask if immense galactic lag s part of the Operaton Shadowhand package? ;)

#11 Phoenix Rising

Phoenix Rising

    Beyond the Impossible

  • Petrolution Staff
  • 6,509 posts
  • Projects:Phoenix Rising
  •  Mod Leader
  • Division:Petrolution
  • Job:Mod Specialist

Posted 11 April 2012 - 06:03 AM

If the game crashes twice on the same line number, then it's the same bug that's the cause. Even if we can't access the source, the top of the error report, as you have it, is useful for organizational purposes. If more people run into 0x008E4880, we could at least rule out or bolster the operating system theory.

I'd postulate that OSH is the biggest campaign we'll ever tackle.

#12 Ghostrider

Ghostrider

    Sith Lord of Campaigns

  • Project Team
  • 2,035 posts
  •  Phoenix Rising QA Lead; Manual Editor

Posted 11 April 2012 - 05:50 PM

I can't imagine designing fleets that are a) larger or b) more complex that those seen in OSH, where the Deep Core fleets had been constructed in secret for over 30 years culminating with those behemoths that form the flagships of all the Imperial Core fleets.

In fact it was really challenging to give an impression of the scale and variety of fleet units, and still keep within the EAW limit of about 60 different fleet slots. If you have a fleet with more than 60 different unit Types, the engine doesn't like it.

#13 Aizen Teppa

Aizen Teppa
  • Members
  • 206 posts

Posted 12 April 2012 - 06:03 AM

Wait a minute, that's maybe it! 60 slots. I never disband warships, no matter what kind of garbage they are, one laser cannon is more useful than none [and in 1.2 AI whoop my backside pretty much every time until I can form some logical, easy to support/defend, perimeter]. Disbanding only useless freighters and cargo ships because I don't need them on the front lines. I will have to pick last properly working save, disband few obsolete fighters of frigates and see what happens later. On the other hand, its maybe also AI fault because what they throw at me sometimes (numbers!) is simply mind boggling..

We have saying/proverb whatever you like to call it. Hope is mother of stupid ... [and aunt of those in need]. Well... I'm in desperate need, and I don't care how stupid I'm. :evgr:

#14 Phoenix Rising

Phoenix Rising

    Beyond the Impossible

  • Petrolution Staff
  • 6,509 posts
  • Projects:Phoenix Rising
  •  Mod Leader
  • Division:Petrolution
  • Job:Mod Specialist

Posted 12 April 2012 - 09:47 AM

From what we can gather, 1 Dreadnaught in a fleet and 100 Dreadnaughts in a fleet is going to be treated the same by the engine. It's many different types of ships that slow things down.

#15 Aizen Teppa

Aizen Teppa
  • Members
  • 206 posts

Posted 13 April 2012 - 04:07 PM

Here is one little snippet from list of problems I discovered + last save from Op. Skyhook which is bugged to hell.

http://www.filefacto...17-09-15-16_avi [10MB]

[file will be available for 7 days - guest upload]

That particular fleet is rag tag combination of 3 fleets. 1 fleet arrived at Bandomeer from Axilla*. I conquered 4 planets over there. Next two are remnants of defeated fleets over Pradosts and Toprawa. Discovered this completely by accident. I wanted to check what troops i have at Bandomeer before i realized that game "bugged-out". And bang. Total mess.

I will start new campaign (Thrawn Of. or Op. Skyhook) on Sat in compatibility mode and see what happens.

* - sorry for any typos. I don't recite SW names too well, it is not my primary fix. ;)

Attached Files



#16 evilbobthebob

evilbobthebob

    evilbobthemapper

  • Project Team
  • 2,304 posts
  • Location:USA
  • Projects:Phoenix Rising Maps
  •  Phoenix Rising Mapping Lead

Posted 13 April 2012 - 04:38 PM

I've watched that, and I can see what problem you're having. It's definitely a case of too many unit types in one fleet. As PR says, implementing fleet population cost is almost certainly the solution.

Phoenix Rising, head of mapping. Thanks to everyone who got us to the position below!
Posted Image


#17 Phoenix Rising

Phoenix Rising

    Beyond the Impossible

  • Petrolution Staff
  • 6,509 posts
  • Projects:Phoenix Rising
  •  Mod Leader
  • Division:Petrolution
  • Job:Mod Specialist

Posted 13 April 2012 - 09:50 PM

I'm downloading now, but if units disappear or start flashing in a fleet slot, you've got too many variants in a fleet and need to break it up. Yeah, watching now - that's definitely the issue. Unfortunately, we can't force <35 variants to a fleet.

Edited by Phoenix Rising, 13 April 2012 - 10:02 PM.


#18 Aizen Teppa

Aizen Teppa
  • Members
  • 206 posts

Posted 15 April 2012 - 11:05 AM

OK, have fun analyzing. 5 exception with same code. Plus screenshot of research bug. When I researched upgrades at Hockaleg, upgrade ended in the orbit (I can sell it just fine) but research of the prototype working fine :blink: . Abandoned researching there. I don't know if it is important, but after last crash reloaded battle at Toprawa. My Interdictor couldn't enable Gravity well generator. Really weird.

Reached a point where I can't save game because it ends like this (exceptions!). Bloody frustrating...

Running in normal mode, not compatibility. With Compatibility enabled it crashed 50x more.

OK 6 exceptions. Can't save, can't invade. Game as such isn't bugged (unresponsive ships, missing heroes etc)

Attached Thumbnails

  • EAW_20120414_12405676.JPG

Attached Files



#19 evilbobthebob

evilbobthebob

    evilbobthemapper

  • Project Team
  • 2,304 posts
  • Location:USA
  • Projects:Phoenix Rising Maps
  •  Phoenix Rising Mapping Lead

Posted 15 April 2012 - 11:24 AM

That error code is one that has popped up in testing before. PR got it on Operation Skyhook as well, I believe. The odd tech research error suggests that maybe it's an issue with the tech tree, but I can't be sure. The game may possibly be having issues with saving the tech tree status (effectively the story mode status) in that campaign.

Phoenix Rising, head of mapping. Thanks to everyone who got us to the position below!
Posted Image


#20 Aizen Teppa

Aizen Teppa
  • Members
  • 206 posts

Posted 16 April 2012 - 04:06 PM

OK, I solved one mystery at least. I think I know why researched upgrades move into orbit.

Fault is in GC settings. In this example, Op. Skyhook default setting is to start at tech level 2. But I selected it to start at level 1. All units which are researched already (like e.g. Victory SD II) have also prototype research icon next to it - by default at lvl 1 Empire can only build Venators. Researching prototype is not a problem, but research of upgrades is not possible or if it is possible only first upgrade is available. When research is finished upgrade ends in the orbit. This issue affect all space and land units which are already researched (with default setup) when player forcing lower tech level at the start of GC.

Research of other units is not affected - e.g. existing Victory SD II is upgradeable to level 4 without problems.

Anyway... shhh I don't want to jinx it. Loaded earlier save - before that Toprawa battle. And all is fine. One noteworthy change I made is that I added swfoc.exe to exclusion list for ffdshow for audio and video codecs (using K-Lite pack). I had that idea because of being long time X3 fan. X3/TC Game has audio (mostly) and video (sometimes) issues with ffdshow which somewhat/greatly reduce performance. From what I noticed there is plenty of SW titles on the list already. I don't know if Foc or basic EaW using ffdshow, but I have to say stability greatly improved. Played into late night without single problem... Now everything probably will disintegrate into oblivion. I shouldn't say this really... :ninja:

Attached Thumbnails

  • ffdshow audio decoder configuration_2012-04-16_17-57-09.png




Reply to this topic



  


0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users