View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0003036 | Spring engine | AI | public | 2012-03-25 23:48 | 2017-05-09 19:45 |
| Reporter | askuhn | Assigned To | Kloot | ||
| Priority | normal | Severity | minor | Reproducibility | always |
| Status | closed | Resolution | won't fix | ||
| Product Version | 88.0 | ||||
| Summary | 0003036: RAI causes hang detection triggered on large maps with many metal spots | ||||
| Description | When I load a single player game with an RAI bot, spring hangs for about 1 minute after the loading screen before eventually starting the game and leaving some stuff in the infolog. If I load with AAI, E323, or any of the other AIs, it starts normally, no delay after the loading screen. Only tested in BA 7.63 & 7.68 & Chicken Defense, Spring MT, Ubuntu 11.10 64-bit, 88.0 from PPA Thanks! :) | ||||
| Tags | No tags attached. | ||||
| Attached Files | |||||
| Checked infolog.txt for Errors | |||||
|
|
Oops, didn't know that ~ was a special character here on mantis. I meant that spring hangs for "about a minute" before starting the game. |
|
|
Does it happen with a regular (non MT-Sim OMP) spring? |
|
|
Just tested, & yes it does. Attached the infolog for that also. |
|
|
Probably OMP related then. Would be nice if you could try it without OMP to confirm that. |
|
|
can you please attach script.txt? |
|
|
I do have a build from git, but how do I recompile without OMP? Do I need to recompile? I'll be sure to attach script.txt from this point forward. Thanks. |
|
|
This is not duplicate of mine reported "hangs with MT bug" since singlethreaded spring for me never hangs. Unfortunately on same game demo(from zk site) it doesnt hang only when joining online game and you can reproduce as long as game is running, with singlethreaded it always connects. |
|
|
I'm not familiar with linux builds but there should be an option "OPENMP" that you can disable. wolas: maybe so, but I think it is related based on the weird stack traces, which both start with address 0x3. |
|
|
Compiled 88.0 master branch from git using -xopenmp=none as a flag (found with some internet searches). Looks like openmp was disabled, but It still hangs after load. :( I want to clarify whats happening to me, just to be sure there isn't any confusion... There is only a hang when there is an RAI bot in the game, and the game will proceed after hanging for a minute or two. I might try to build the latest develop soon to see if there are any changes. If you guys have anything else for me to try, I'll do my best to reply. |
|
|
askuhn: did you try on a smaller map? i guess thats just an RAI issue, it takes very long to initialize on large maps... |
|
|
That's interesting, abma. I used the same setup on my smallest map (8x8), and there was no hang. The reason I had not tried this is because RAI never used to hang on any map (even my larger 24x24 for example) in previous engine/game versions (but I'm not sure exactly when the issue appeared). |
|
|
so... "normal" spring (non-mt) does hang, too? |
|
|
Yes he said it does... |
|
|
No change on latest develop, 88.0.1-52-g24bd079. Even tried deleting RAI's cache to no avail. I'm pretty much out of ideas on this one as a regular user... |
|
|
I can reproduce it, seems to be a "bug" in RAI. From glancing at the code it calls cBuilderPlacement::FindResourceDistance to find the shortest path from each metal spot to every other metal spot (and repeats that entire process for every unique MoveDef) but there are 80 spots on eg. Comet Catcher (3160 pairs) and at least 20 MoveDefs in BA --> your CPU will choke. |
|
|
Glad you could get to the root of this one. I wish I could dig into the AI code and help maintian it, but I don't have the know-how... Thanks for all the responses :) |
|
|
since no volunteers have stepped up in 5 years and I have no plans to dive into RAI myself, it's safe to call this a wontfix. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2012-03-25 23:48 | askuhn | New Issue | |
| 2012-03-25 23:48 | askuhn | File Added: infolog - RAI Hang - 88.0.txt | |
| 2012-03-25 23:48 | askuhn | File Added: infolog - AAI No Hang - 88.0.txt | |
| 2012-03-25 23:51 | askuhn | Note Added: 0008501 | |
| 2012-03-26 03:18 | zerver | Note Added: 0008502 | |
| 2012-03-26 03:23 | zerver | Relationship added | has duplicate 0003035 |
| 2012-03-26 03:31 | askuhn | Note Added: 0008504 | |
| 2012-03-26 03:32 | askuhn | File Added: infolog - RAI Hang - 88.0 - non MT-Sim OMP.txt | |
| 2012-03-26 19:38 | zerver | Note Added: 0008507 | |
| 2012-03-26 21:48 | abma | Note Added: 0008508 | |
| 2012-03-26 22:07 | askuhn | Note Added: 0008510 | |
| 2012-03-26 22:13 | askuhn | File Added: infolog - RAI Hang -with script1 .txt | |
| 2012-03-26 22:13 | askuhn | File Added: script1.txt | |
| 2012-03-26 22:48 | wolas | Note Added: 0008511 | |
| 2012-03-27 01:30 | zerver | Note Added: 0008514 | |
| 2012-03-27 03:18 | askuhn | Note Added: 0008516 | |
| 2012-03-27 03:20 | askuhn | File Added: infolog - RAI Hang - -xopenmp=none.txt | |
| 2012-03-27 10:21 | abma | Note Added: 0008518 | |
| 2012-03-27 16:21 | askuhn | Note Added: 0008524 | |
| 2012-03-27 16:32 | abma | Relationship deleted | has duplicate 0003035 |
| 2012-03-27 16:33 | abma | Note Added: 0008525 | |
| 2012-03-27 19:49 | zerver | Note Added: 0008528 | |
| 2012-03-27 23:25 | askuhn | Note Added: 0008529 | |
| 2012-03-28 00:24 | Kloot | Note Added: 0008530 | |
| 2012-03-28 00:26 | Kloot | Note Edited: 0008530 | |
| 2012-03-28 16:14 | abma | Summary | Hang detection triggered for Spring 88.0 (MT-Sim OMP) only when loading a game with RAI => RAI causes hang detection triggered on large maps with many metal spots |
| 2012-03-29 02:21 | askuhn | Note Added: 0008541 | |
| 2012-03-29 09:10 | abma | Relationship added | has duplicate 0003042 |
| 2012-04-12 19:03 | hoijui | Description Updated | |
| 2016-02-25 02:41 | abma | Relationship added | has duplicate 0005109 |
| 2017-05-09 19:45 | Kloot | Assigned To | => Kloot |
| 2017-05-09 19:45 | Kloot | Status | new => closed |
| 2017-05-09 19:45 | Kloot | Resolution | open => won't fix |
| 2017-05-09 19:45 | Kloot | Note Added: 0017581 |
