Long pauses during unit moves

[AlanH]

Originally posted by Brad Oliver


That's interesting, because Civ3 isn't multi-threaded. In fact, about the only thing it could be related to is the sound callback code.

Oh well, I gave up the Civ3 source on Oct 15 2003 when I left Westlake, so it's someone else's problem now.

That's interesting, as every crash log I've looked at includes multiple thread reports.

Running a slow Mac it's clear to me that there are background processes running to interfere with screen updating, as I see occasional delays of a second or two. These are not due to other Mac apps, as they never interfere with anything except Civ3. I've just never seen anything like a minute.

 

[dojoboy]

Originally posted by Danapoppa

I just purchased Civ3 GotY for Mac, and I've been having the problem described in the thread titled "1.29 beta2." Every now and then the game pauses while a unit is being moved. The unit will stop in mid-stride and stay there for a good minute or so, then finally finish the move and let the game go on as normal.

I've not experienced this unit-pause issue, ever. I'm on an eMac, G4 (1GHz), 10.3.3, with 384 RAM. The only anomally I've experienced has been at the end-turn phase, when the end-turn button is blinking. When things are normal, it blinks at its normal rate (about every second). But, when the anomally hits, the end-turn button appears to hang, blinking maybe once every 5 - 7 seconds. In addtion, all map scrolling is delayed and unit activation (say I wanted to activate a fortified unit or open a city window) is also delayed during this anomally.

It no way effects game-play and does not repeat itself on the very next turn, but it will usually reappear later for one turn, then back to normal again. While typing this, an idea popped into my head. Tonight, I'll turn off the "pause-at-end-of-turn" option in preferences and see if this anomally has an opportunity to appear. But, I hate playing without the "pause-at-end-of-turn" option, no time for a bathroom break.

 

[Queue]

Queue, if you don't mind me asking, how do you output audio? I'm using USB output to a pair of Apple Pro Speakers via a Griffin PowerWave.

I am just useing the regular sound output jack.

 

[Brad Oliver]

Originally posted by AlanH
That's interesting, as every crash log I've looked at includes multiple thread reports.

The OS inserts threads into just about every app for OS-level tasks. Even a mostly single-threaded app like Civ3 is bound to have 6 or 7 of these, but they are outside the direct control of the program.

Civ3 itself runs in a very linear fashion, and as such the code attempts to break out from AI calculations as often as practical to do housekeeping like unit animations and sound updates. If the AI calculations are particularly long, then you will see stalls and hitches on slower Macs as those housekeeping routines don't get serviced.

 

[tao]

I also have this "freezes" up to 2 minutes on my 2x1GHz MDD. It happens in the mid of a unit move: I see the animation stop. I have background music switched off.

Also <cmd>P does not pause but only after the delay.

The machine (disk) is busy.

And when Civ III is paused, the load indicator shows both processors about equally loaded while running Civ III.

 

[ShnikeJSB]

Like I said in previous posts, I have "vanilla" civ3 running on OS 10.3.3 and a Titanium PowerBook G4/667 with a gig of RAM, and I get pauses as long as OVER 5 (FIVE) MINUTES over the length of a SINGLE turn, and I can't figure out what it is! As for the skipping/pausing/restarting music, I use ProcessWizard to up the "nice" value of Civ3 and quit some background tasks, and the music no longer has problems, though the pauses are still there with the turns... Anyone? -JB

Moderator Action: Please can we limit this discussion to one thread. We will all get more out of it if the posts are in one place, not three

 

[n8mac]

It's been a while since I've been to this site. I just stopped back in to see if anything new was happening, plus i'm back to playing civ after beating EUII and some console games.

I have seen this twice, both times not back to back. My system is a...
iMac 400
Built in speakers
OS 10.1

From what I've read here and what I've seen in this thread this not a problem with the...
system
hardware
speakers

...but a civ problem, possibly 1.29. What I experienced was during my turn I had a large stack of units that I wanted to move. I hit J to move all my calvary to the next square without a problem. Next I hit J to move all my riflemen to that same square and the same problem the rest of you described happened, each rifleman took over a minute to move. Looking closer I found that it was animating the whole time, as if the frames per second had dropped drastically. After they were done moving I saved and restarted civ and the problem was gone.

I'm not very knowledgable about this subject but I am guessing that the animation file for the session this occured did not load correctly. I think I can safely say that I have narrowed the problem down to civs animation.

Hope this helps!

 

[AlanH]

I think Brad Oliver indicated the likely cause of such problems in his previous post:

If the AI calculations are particularly long, then you will see stalls and hitches on slower Macs as those housekeeping routines don't get serviced.

I used to see this sort of thing regularly, but it seemed to go away. I can't think of a specific change that improved it, but the things I know have changed over the period are:

- OS X. My Mac's performance has improved with each new version. I'm particularly deeply suspicious of version 10.1, and I think we've had this conversation before. There are so many major differences between 10.1 and the later versions that many software packages simply don't run in 10.1. It was Beta software in all but name (there's an old adage that you should avoid running version 1 of anything), it's at least four years old and I guess you don't see too many updates for it. I really recommend you move to Panther, or Tiger when it arrives.

- More memory. This is critical. I went from 256MBytes to over a GByte and it makes a huge difference to performance. MacOS X uses virtual memory - stashes bits of code or data on disk when it needs to load something else into memory and has no spare space. If you don't have a lot of memory and it has to keep swapping code in and out of memory this could clobber the speed of execution. If there's a bug in the way it decides what to swap this *could* conceivably be responsible for a sudden slowdown.

- CPU speed. I've upgraded my CPU from 350MHz to 1GHz. Civ is so much nicer to play now. But I think the freezes disappeared before that.

You said "Hope this helps!" ...

I'm not sure what you expect to happen as a result of resurrecting this thread. Brad Oliver has left Westlake and as far as I can tell there's now no one who gives a tinker's cuss about investigating or fixing any open issues in the Mac version of Civ3. If you want Civ3 to perform better I can only suggest you change something about your system, with the above thoughts as suggestions. No one's going to change the Civ3 code.

 

[n8mac]

10.1. It was Beta software in all but name

No its not, apple never stated it was. There was a beta and 10.0 sold at full price before it. Does it bother you that I use it?

More memory. This is critical.

This would make sense for often or always slow downs but this is not the case. Why do you keep blaming the hardware and OS when this game works as old as OS 9. The original poster had the problem on a G5, certainly not a lack of hardware there.

I'm not sure what you expect to happen as a result of resurrecting this thread.

You apparently didn't take the time to read my post. I shed some new light on the topic because I experienced the same problem. By stating the fact that in the same session calvary did not move slowly yet the rifleman did (alternating different units) and that the slow down occured on fast computers that common sense would tell you that is was not system or hardware performance but a rare game error. Had this been much more frequent a fix would have been written a long time ago.

I simply shed some new light on the subject that was still dark. I haven't been here a while and saw this unclosed thread. All the time I see reviving of old threads, especially on topics that have been discussed many many times before, yet you put mine down, why? Bad day?

I think Brad Oliver indicated the likely cause of such problems in his previous post

There was two more posters after Brads comment, yet you failed to put them down! Brad apperently did not have this problem before and was guessing from the information he recieved. Just because Brad makes a statement regarding a problem does not automatically make him right and closes the discussion.

No one's going to change the Civ3 code.

I could care less if someone changed the code! If you haven't noticed this thread started out with a rare problem and everyone including you was curious as to what it was. Is it so wrong to help others find out a problem even if solving it due to hard code is impossible?!

 

[tig3001]

Just an FYI,

I agree with AlanH as far as more computing power equals more performance but I've run Civ3v quite well on my beige G3/234/384MB/OS 9.22. Slower but with no long pauses.

Tig

 

[AlanH]

I apologise if my response came over as critical of n8mac. It is really a rant against the fact that there will never be any repairs for Civ3 problems on Macs now, so we just have to live with them.

Yes, G5 duals seem to have a specific problem. I cannot comment on that as I do not have one.

 

[Pinky]

Hey there,

It's been a while since I came here, but here is my input. This problem has been happening to me since 1.29b1. CivIII hangs, during MY turn, when moving a unit from 1 tile to a neighbour tile. See my replies here for detail (during the initial beta release):
http://forums.civfanatics.com/showthread.php?t=50695&page=3&pp=20
Whatever the cause, it is not a normal CivIII behaviour. It is not the AI. I have played many games in 1.21g without that problem. The computer never hanged for that long during the AI's turn, even late in the game. This is not even during the AI's turn...Going back to 1.21g solved that problem, so it is 1.29 specific.
Anyhow, never found a workaround...

Dual 866MHz 768Mb Ram, and 10.2.something at the time.
Good luck !

 

[dojoboy]

Hey there,

It's been a while since I came here, but here is my input. This problem has been happening to me since 1.29b1. CivIII hangs, during MY turn, when moving a unit from 1 tile to a neighbour tile. See my replies here for detail (during the initial beta release):
http://forums.civfanatics.com/showthread.php?t=50695&page=3&pp=20
Whatever the cause, it is not a normal CivIII behaviour. It is not the AI. I have played many games in 1.21g without that problem. The computer never hanged for that long during the AI's turn, even late in the game. This is not even during the AI's turn...Going back to 1.21g solved that problem, so it is 1.29 specific.
Anyhow, never found a workaround...

Dual 866MHz 768Mb Ram, and 10.2.something at the time.
Good luck !

I know of the issue, because several had reported it. I'm using 1.29b2 and have not experienced this at all. I really believe its machine specific. Why? I can't say.

 

[AlanH]

@Dojo - are you running a dual CPU? I suspect not. It seems to me that we are seeing a lot of duals with this problem, and Pinky's is another one. That doesn't explain m8mac's problem, of course.

 

[dojoboy]

@Dojo - are you running a dual CPU? I suspect not. It seems to me that we are seeing a lot of duals with this problem, and Pinky's is another one. That doesn't explain m8mac's problem, of course.

Nope, uni-CPU. Have these folks tried shutting one done to see what would happen?

 

[Queue]

Nope, uni-CPU. Have these folks tried shutting one done to see what would happen?

Yes I have tried that, and it seems to clear up the problem. However, the pause does not happen enough to make me turn off the second processer everytime I want to play Civ

 

[Pinky]

Well that's good to know. And now the real question :
How does one turn off the second processor ?
Heh..never thought that my little processors would battle one another for civIII...infernal machines !

 

[ShnikeJSB]

Hey all! Just here to throw in my 2 cents. I had previously posted about my problems with slowdowns also earlier. I found that not only did installing the 1.29b2 patch help, but also renicing helps A LOT too. I use Speed Freak to make the frontmost app -15 (+75%) and the background apps +5 (-25%). It really helps with not only speeding up turns but also the sound/music quirkiness I experience. Plus, just turning OFF the sounds/music all together helps too. I recommend at least trying to patch and the renicing to help things a bit. I run plain vanilla Civ3 on a TiBook G4/667 with 1 gig of RAM and OS 10.3.5. Good luck everyone, and here's to looking forward to a Civ4 and a new Powerbook (eventually...). LOL! -JB

 

[Duke of Ybrik]

I know this thread has been off for months, but I feel the need to throw in an observation on this issue (it's a real pain to me, also). I am running a dual board 450mhz G4 with 256MB, so go juice is not a problem.

My machine seems to be searching for a network connection during these long, unit mid-stride hangs. I solve it by doing a cold boot on my firewall router coupled with unplugging/repluggin the ethernet cable from the back of the tower. You hear the HD churn for a few seconds, then units leap back into motion.

If I don't do this, I can wait up to 2-3 minutes before it self-resolves. It is very curious that these hangs appear to be ethernet connected, since Civ III does not have an online game option (unlike Civ II).

 

[AlanH]

Once more it's a dual CPU config that shows the problem.

As you say, it's difficult to believe the Civ3 software is directly sensitive to the state of the ethernet connection, since network play wasn't even a part of Civ3 at teh time when that version was released.

You may simply be rearranging the priorities in the CPU. Unplug the ethernet connection or kill the router, and the network software goes into overdrive to reconfigure itself to try to recover. Praps that breaks some semi-locked up thread conflict between the activities of the two CPUs.