Author Topic: Prerelease 1.009L (AI Voice Taunts -- Last 1.009 Prerelease).  (Read 13409 times)

Offline x4000

  • Chris McElligott Park, Arcen Founder and Lead Dev
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 31,651
UPDATE:  The full version of 1.009 is now out, so please use that instead: http://www.arcengames.com/share/AIWar1009.zip

That version is an upgrade from version 1.008, so you have to already have 1.008 (or greater) installed. Just unzip it into your game folder (usually C:\Program Files\Arcen Games\AI War\ unless you specified something else). Please make sure that your unzip process keeps the folder structure from the zip file, rather than just unpacking all of the files into the base target directory.

What's new since 1.009H (I was a minor release):
(Cumulative release notes since 1.008 are attached at the bottom)

-------------------

-The AI will now taunt the players (with audio as well as text) when the human players lose a big battle, or when the AI destroys the command station of a human player.

-All of the remaining zoom/rotation/scale related uses of floating point math have been converted to use fixed-point math instead.  This should prevent any potential related multiplayer desyncs based on different machine architectures.

-Previously, when the client in a TBS-friendly pause game hit pause, it would sometimes just hang on Pausing... and never actually go to full pause.  Fixed.

-A shift has been made to the way the planetary summary items are aggregated, that should prevent the issue of duplicate entries (one white, one red/yellow/blue).

-There was previously a desync that could be introduced when loading a saved game from an older version with the 1.009 versions for the first time.  Simply saving the desynced game again and loading it would fix it, but now the root issue is also fixed.
« Last Edit: July 09, 2009, 02:58:51 pm by x4000 »
Have ideas or bug reports for one of our games?  Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!

Offline Admiral

  • Hero Member
  • *****
  • Posts: 547
Desync problem on latest 1.009 (!)
« Reply #1 on: July 07, 2009, 08:12:43 pm »
For my play session, we tried this straight off.

Almost immediately, we got a desync error. I have attached the save game that we load (on the desktop) which almost immediately gets one or two desync messages on the host computer.

I also attach the save games from the host and the laptop from immediately after a desync.

This happened twice on two reloads.

Cheers!

Offline Admiral

  • Hero Member
  • *****
  • Posts: 547
Desync happens on 1.009H as well as 1.009J
« Reply #2 on: July 07, 2009, 08:26:33 pm »
Will test G next.

Offline Admiral

  • Hero Member
  • *****
  • Posts: 547
Desync also happens on 1.009G... Will keep going back...
« Reply #3 on: July 07, 2009, 08:29:28 pm »
(no text)

Offline Admiral

  • Hero Member
  • *****
  • Posts: 547
Desync also happens on 1.009F
« Reply #4 on: July 07, 2009, 08:33:19 pm »
Next I will try E, but I need to eat dinner first... Maybe Colonel X will have a fix before then. (Ready, set, go!)

Offline Admiral

  • Hero Member
  • *****
  • Posts: 547
Desync happens on 1.009E
« Reply #5 on: July 07, 2009, 09:01:09 pm »
It happens, and seems to be even worse - many more repeated desync messages instead of just one or two.

We're going to start a new game on the latest release.

Offline x4000

  • Chris McElligott Park, Arcen Founder and Lead Dev
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 31,651
Re: Prerelease 1.009J (AI Voice Taunts -- Last 1.009 Prerelease).
« Reply #6 on: July 07, 2009, 09:20:54 pm »
Any idea what one of you might have been doing directly before the desync?  From what I am seeing here, it seems like a lot of your rotations are ever so slightly off, which is interesting.  What kind of machines are these?  Are they both the same OS?  Are the both AMD or Intel?  Are the both 32 or 64 bit?  Neither of those should matter, but they will help me narrow this down.  I've loaded up your savegame on a pair of machines, and am not getting a desync with it.  So I don't know if there is some difference between your two machines that is causing this, or if it's related to some activity that is causing the desync...
Have ideas or bug reports for one of our games?  Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!

Offline Admiral

  • Hero Member
  • *****
  • Posts: 547
New Bug
« Reply #7 on: July 07, 2009, 09:43:37 pm »
In the ship list on the left side, I have two entries for Fighter. One is gray, the other is blue (my team color). They keep swapping every second or two. One shows nothing, the other shows the proper amount of data.

As to the other problems:

The save was from 4 days ago, so I don't recall what was happening with it, sorry.

The machines are different (and posted in other places):

Laptop:
Core 2 Duo
Vista Home x32
3G RAM
ATI video

Desktop:
8G RAM
Core 2 Quad
Vista Ultimate x64
nVidia Video

Both .NET 3.5 SP1 with the same hotfixes.

Hope this helps.

Offline x4000

  • Chris McElligott Park, Arcen Founder and Lead Dev
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 31,651
Re: New Bug
« Reply #8 on: July 07, 2009, 09:49:14 pm »
In the ship list on the left side, I have two entries for Fighter. One is gray, the other is blue (my team color). They keep swapping every second or two. One shows nothing, the other shows the proper amount of data.

I still can't duplicate this, but I'll take a look again for it.

The save was from 4 days ago, so I don't recall what was happening with it, sorry.

So, what you're saying is that it was desyncing literally immediately after you started.  Neither of you had really done anything with it tonight?  That is very curious indeed -- I've been running it for 30 minutes now with no desync.  I think this might have something to do with one last little vestige where I am using floating point numbers instead of fixed-point numbers.  I wouldn't expect the rotations to matter, but that is what is getting incorrect.  I'm working with that now, and I'll post a new version for you to try when that is ready.

Are you having any trouble with the new savegame?

The machines are different (and posted in other places)

Thanks for re-posting -- I just wanted to make sure I had accurate data.  The floating point precision is different between 64 and 32 bit machines, but I can't for the life of me think why that would matter for the exceedingly limited use that I have for it in the game at this point.  I routinely play with a mix of 32 and 64 bit machines without issue, which is doubly curious.  We'll see what happens once I have that out -- at the very least, it will avoid clouding the desync logs with inconsequential rounding-error clutter...
Have ideas or bug reports for one of our games?  Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!

Offline Admiral

  • Hero Member
  • *****
  • Posts: 547
Another bug: Pausing
« Reply #9 on: July 07, 2009, 10:09:21 pm »
Hey again,

Pausing (TBS-mode on both computers).

Pausing on the laptop (the client, not the host) causes the host to be immediately paused, but the laptop sits there "Pausing..." until the host saves the game, then it saves "Paused."

Cheers!

Offline Admiral

  • Hero Member
  • *****
  • Posts: 547
Flashing fighters
« Reply #10 on: July 07, 2009, 10:10:51 pm »
Here's a save with the flashing fighters.

Offline Admiral

  • Hero Member
  • *****
  • Posts: 547
Re: New Bug
« Reply #11 on: July 07, 2009, 10:15:33 pm »
So, what you're saying is that it was desyncing literally immediately after you started.  Neither of you had really done anything with it tonight?  That is very curious indeed -- I've been running it for 30 minutes now with no desync.

Yes, it desyncs within a minute with little being done to cause it.

Are you having any trouble with the new savegame?

The new game is running fine, with the aforementioned problems (pausing, flashing pair of fighters on the side bar).

Thanks for re-posting -- I just wanted to make sure I had accurate data.  The floating point precision is different between 64 and 32 bit machines, but I can't for the life of me think why that would matter for the exceedingly limited use that I have for it in the game at this point.

Interesting that that would be a problem... IEEE math should return the same on two machines, unless they are using different precisions. I thought the Intel chips all did 80-bit internal before rounding to 64-bit or 32-bit external, using one of the rounding modes.

Are these desyncs ones we can safely ignore? Is there a way to cause a resync after a desync? :)

Thanks!

Offline Admiral

  • Hero Member
  • *****
  • Posts: 547
Flashing fighter sidebar bug
« Reply #12 on: July 07, 2009, 10:17:12 pm »
I made it go away by hitting F1 twice.

It might have been caused by hitting F1 earlier; I had never hit it before in any other game prior to this one.

Offline x4000

  • Chris McElligott Park, Arcen Founder and Lead Dev
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 31,651
Re: New Bug
« Reply #13 on: July 07, 2009, 10:20:50 pm »
Yes, it desyncs within a minute with little being done to cause it.

Okay... if you can figure out exactly what you are doing right before it desyncs, that might be helpful if the floating point thing doesn't fix it.

The new game is running fine, with the aforementioned problems (pausing, flashing pair of fighters on the side bar).

Well, that's good to hear (I guess).  I'm glad it's working for you, but by the same token that probably means there is something off with one of the ship types you are using in the other game.  Something that triggers this on your two machines, anyway.

Interesting that that would be a problem... IEEE math should return the same on two machines, unless they are using different precisions. I thought the Intel chips all did 80-bit internal before rounding to 64-bit or 32-bit external, using one of the rounding modes.

Actually, they can be highly inaccurate, even between intel machines.  In the past, before I switched to fixed-int math (early in alpha) for most calculations, I was getting small rounding differences between 64 and 32 bit machines in multiplayer.  These differences were extraordinarily tiny, but when compounded they would eventually be large enough to cause a desync.  I'm no expert on why that happens, but it definitely does happen.

Are these desyncs ones we can safely ignore? Is there a way to cause a resync after a desync? :)

The only way to resync after a desync is to have the host save and reload, and then the client reconnect.  Or the client can just reconnect, but I recommend saving first just in case.  There aren't any desyncs that can be safely ignored, because they will eventually cause a great divergence even if they are trivial at the start (plus, it fills the host's screen up with desync information continuously).  I'm working on the fixed-int thing, should have that soon...
Have ideas or bug reports for one of our games?  Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!

Offline x4000

  • Chris McElligott Park, Arcen Founder and Lead Dev
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 31,651
Re: Flashing fighter sidebar bug
« Reply #14 on: July 07, 2009, 10:21:55 pm »
I made it go away by hitting F1 twice.

It might have been caused by hitting F1 earlier; I had never hit it before in any other game prior to this one.

Could be... I'll check.  But hitting F1 will clear everything with the planetary summary, so that's not surprising that that would fix it once it was started.  Otherwise it should just go away on its own after 60 seconds.  Thanks!
Have ideas or bug reports for one of our games?  Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!