View Issue Details
| ID | Project | Category | Date Submitted | Last Update | |
|---|---|---|---|---|---|
| 0001772 | AI War 1 / Classic | Bug - Other | Dec 4, 2010 12:02 am | Dec 22, 2010 7:17 pm | |
| Reporter | Draco18s | Assigned To | Chris_McElligottPark | ||
| Status | resolved | Resolution | fixed | ||
| Product Version | 4.044 | ||||
| Fixed in Version | 4.054 | ||||
| Summary | 0001772: Multi-Stacked FFs: When one dies, everything is left unprotected for a moment | ||||
| Description | Screenshot series to follow, will post in order. | ||||
| Tags | No tags attached. | ||||
| Internal Weight | |||||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
In order: 1) Things unprotected, no remains. 2) 2 Remains (tractor and grav turret), things protected. 3) 2 FFs died, things unprotected 4) 3 remains (3 tractors--grav turret rebuilt), things protected. |
|
|
How long of a moment are we talking? The game only processes so many TryProtections calls per second to avoid major cpu death on planets with tons of ff's. There's a punch-through so that command stations _always_ get to retry if they just lost their ff protection (and even then shots landing in the same frame as the ff dying can get through, it's just rare). |
|
|
Actually, having ANY ship that just lost it's FF protection retry probably wouldn't be a bad idea. That would drain fairly little CPU power in most cases, as it's not common to have a ton of ships all losing FF protection at once. I suppose the two exceptions would be: if the ship doesn't have supply, or the player is human and is in a brownout at the time. Or if the ship has been paralyzed (EMPed). Those could get expensive. |
|
|
A couple seconds, actually. It was long enough that it wasn't difficult to screenshot. Notably we were on Time -2, but that helped alleviate the lag of 3000 Neinzul ships warping in (allow more time for processing -> consistent slowdown). But I still think it was for a full second of game-time, as the Neinzul ships were able to take down a shield from full radius to dead as it would never receive protection from the other FFs, despite being smaller, and then everything else was still unprotected long enough to get shot to rubble. About every 3 waves we'd have to rebuild the FF stack (about 15 to 20 Mk1s and 5 Mk2s). As for the every-ship-retry, you could add them to a stack and only process so many per game tick, so that if a ton DO lose protection (say lost supply or a brownout) you'd spread the processing out, and in most cases there wouldn't be a new FF anyway, so it wouldn't matter if they had to wait. |
|
|
Chris, if you want to have that special flag default to true, that would probably be sufficient to try what you're suggesting. I make no warranty on what happens to cpus ;) |
|
|
Ok: * A change has been put in place to make ships immediately re-check for new forcefield protection as soon as they lose protection from a prior force field. This was previously the case for command stations only. This will have some performance impact, but will greatly help correctness in the case of multi-stacked force fields. And in most cases, the performance impact will hopefully be negligble, but we look forward to your feedback on that. |
|
|
Sweet. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| Dec 4, 2010 12:02 am | Draco18s | New Issue | |
| Dec 4, 2010 12:02 am | Draco18s | File Added: Screenshot_2010_12_04_00_00_32.png | |
| Dec 4, 2010 12:02 am | Draco18s | File Added: Screenshot_2010_12_04_00_01_08.png | |
| Dec 4, 2010 12:03 am | Draco18s | File Added: Screenshot_2010_12_04_00_01_14.png | |
| Dec 4, 2010 12:03 am | Draco18s | File Added: Screenshot_2010_12_04_00_01_33.png | |
| Dec 4, 2010 12:05 am | Draco18s | Note Added: 0005124 | |
| Dec 4, 2010 12:05 am | Draco18s | Note Edited: 0005124 | |
| Dec 4, 2010 11:24 am | keith.lamothe | Note Added: 0005197 | |
| Dec 4, 2010 11:24 am | keith.lamothe | Assigned To | => keith.lamothe |
| Dec 4, 2010 11:24 am | keith.lamothe | Status | new => feedback |
| Dec 4, 2010 11:29 am | Chris_McElligottPark | Note Added: 0005200 | |
| Dec 4, 2010 11:33 am | Draco18s | Note Added: 0005203 | |
| Dec 4, 2010 11:33 am | Draco18s | Status | feedback => assigned |
| Dec 4, 2010 11:35 am | Draco18s | Note Edited: 0005203 | |
| Dec 4, 2010 11:58 am | keith.lamothe | Note Added: 0005211 | |
| Dec 22, 2010 7:04 pm | Chris_McElligottPark | Note Added: 0006767 | |
| Dec 22, 2010 7:04 pm | Chris_McElligottPark | Status | assigned => resolved |
| Dec 22, 2010 7:04 pm | Chris_McElligottPark | Fixed in Version | => 4.054 |
| Dec 22, 2010 7:04 pm | Chris_McElligottPark | Resolution | open => fixed |
| Dec 22, 2010 7:04 pm | Chris_McElligottPark | Assigned To | keith.lamothe => Chris_McElligottPark |
| Dec 22, 2010 7:17 pm | Draco18s | Note Added: 0006772 |
