Performance: b6497 (w/ 6503 content)

[yandonman]

Performance metrics:


Verbose logs are turn 200 to 201. Clicking turn button _feels_ like a minute.

Time counters of interest (time at least 3 digits)

Code: Select all

	Line 295: 2013-11-10 00:27:06,281 DEBUG Server : ProductionQueue::Update time: 106
	Line 692: 2013-11-10 00:28:08,287 DEBUG Server : ServerApp::PreCombatProcessTurns time: 1973
	Line 6300: 2013-11-10 00:28:10,130 DEBUG Server : Universe::GetEffectsAndTargets time: 1513
	Line 85775: 2013-11-10 00:28:15,999 DEBUG Server : Universe::ExecuteEffects time: 5868
	Line 85776: 2013-11-10 00:28:16,001 DEBUG Server : Universe::ApplyAllEffectsAndUpdateMeters time: 7384
	Line 93492: 2013-11-10 00:28:18,679 DEBUG Server : Universe::GetEffectsAndTargets time: 1487
	Line 173681: 2013-11-10 00:28:24,670 DEBUG Server : Universe::ExecuteEffects time: 5990
	Line 173682: 2013-11-10 00:28:24,672 DEBUG Server : Universe::ApplyMeterEffectsAndUpdateMeters on all objects time: 7480
	Line 177123: 2013-11-10 00:28:26,346 DEBUG Server : ServerApp::PostCombatProcessTurns time: 18027

[Geoff the Medio]

I notice a few instances of variations of:

Code: Select all

Universe::StoreTargetsAndCausesOfEffectsGroups done processing source 269 cause: SP_EXOBOT effects group 23 time: 70

Those likely add a disproportionate but not dominant amount, as it otherwise looks like a long list of individually short effect executions that add up to a long time collectively.

The target set determination is about a fifth of the effect execution time, so not dominant.