Removing the Fog of War to speed up turn times?

Wasn't it more the case that the engine tries to check for obstacles as it goes along each hex? There is no 'these hexes are fine, just go without checking each time' omniscient part of the turn processing engine.

---Sig---
Playing Touhou games (Go to this Wiki if interested. No, nobody else is that good/insane as that replay). No rush for SE

The hex checking process was extremely slow, but it was optimized a long time ago so that playing with no fog of war wont really do much good. Using Simultaneus Movement mode gives you the fastest turn processing possible in the game at this stage. That said, I'm afraid the only way to get turn processing to go really fast is to simply get a very fast CPU. In fact, if you're playing with alot of AI players you probably wont be able to get a 'too fast CPU' for many years to come, unless Aaron does some major changes to the game's processing system.

If you do choose to use Simultaneus mode then HDD speed will also matter some, since it reloads the saved game twice for every turn.

Oh, and you can turn off fog of war yourself by simply selecting 'All players can see all systems' in the sensor options when creating a new game.

There are somethings that can be whittled down in the scripts from what I've seen, that's going to be my next major undertaking.

From what I've seen though there is a fair bit of room where things can be optimized, just how much can be done without rewritting everything from scratch will be the real question.

Unfortunately the script execution isn't what's causing most of the turn processing time. It's the ship movement phase that needs to be optimized( and the combat phase needs more optimizing ), and that can only be done by Aaron.

One possible solution might be to reduce the system movement speed of ships. For an extreme example, having all ships move at only 1 hex per turn would probably cut down the turn processing times immensely.

Hmm, that is an interesting solution... it would make ships horribly slow, of course, but if you simply reduced the number of hexes in a system and made each hex bigger on the screen (this is possible in Settings.txt! ) you would be able to do that... you'd have the additional advantage of being able to show actual ships on the screen in system view instead of just tiny specks

Of course, this would slow down combat, seeing as combat speed is proportional to system speed - but hey, there's also a "combat speed per move point" in the VehicleSizes.txt

~~~
The power of the ancients is now yours to command. You are truly the Master... of the obvious

Yeah, definitely something to keep in mind for modders. Slower ship movement isn't a problem as long as the mod has been designed and balanced with that in mind. Combat speed isn't a problem either since that can be tweaked with the individual ship settings, like you mentioned.

Just to get a notion of what could be gained, I tried this with BM 1.11 in a 6 player game. I set the engines' movement points down to 25( originally 100 ), so the ships moved only 1/4 the normal distance. At turn 50 a no-combats turn processing went from 10 seconds with the normal engine speed to 6 seconds with the modified engine speed( tried several times, results were consistent ). This could bode really well for large games with turn processing times much longer than 10 seconds.

I'm not sure about you guys, but I'd be much happier with bug fixes and feature additions then turn time optimization at this point.

Then again, I'm patient and run my Core 2 Duo 2 3.2ghz, so guess I'm biased.

I'm with noxxle on this one. Waiting for many seconds, sometimes minutes, and being forced to play games with few AI players and small galaxies, is not my idea of fun. 10 seconds or so I could be willing to wait between turns in a late game, but much more than that and I quickly lose interest. Spending more time watching 'Processing Turn' than actually playing the game... *shrug*

Turn time processing should be the main focus now all the major bugs are gone.It is still too bloody slow.

Please tell me you all are kidding, right?? You complain about 30+ sec wait times (and call it a game breaker)?
Wow, how impatient are we? Well, what should I expect from a society that is skipping grammar for speed.
Seriously lets worry about actual game-breakers.

And my guess is that the other games are much more simplistic. Try playing Dominions 3. Turns times on that game can get long, and I have yet to hear anybody complain about it on the forums there...
If a game is worth playing, which SE5 is (Dom3 as well), then its worth waiting IMO. The saying 'good things come to those who wait.' is used for S'n'Gs.
And I realize I'm probably ranting... so I'll stop

Uh, turn processing times in Dominions 3 are not even a fifth as long as SEV's.

-

I'm pretty sure I'd enjoy the game about ten times more if the turn processing times were instantaneous. Granted, that's asking way too much of Aaron, but the point is that the lower they are they more fun the game is, for me and probably everyone who spends most their time playing singleplayer.

That said, I'm not really interested in having the turn processing times decreased from 1 minute to 50 seconds or so; that'd be a too insignificant improvement.

This conversation again I see. From my perspective, once the turn processing was patched to work in the background while you Alt-Tabbed to something else solved the issue for me. Doesn't matter if it takes 30 seconds or 20 minutes - there is always something else to multitask it with. Well it's not really "multitasking" in the strict sense - more like "massive contextual jumps". The easiest thing to 'multitask' the waiting with is to actually write a diary/report/story out of what's happened thus far.

But of course, not everyone is comfortable with such 'multitasking'. Hence this conversation arising time and again.

---Sig---
Playing Touhou games (Go to this Wiki if interested. No, nobody else is that good/insane as that replay). No rush for SE

BTW, SE4 was much more simple compared to SE5, esp. from an AI perspective (scripting vs. setting thresholds is a HUGE difference). You will never have turn times like SE4.
And yes, Dom3 can have very long turn times... try playing with large maps with lots of AIs and you will run into 1+ minute turn times (which mainly happen in late/later game)
Also I didn't say SE5 is worth waiting for because its complex... I said its a good game, thus worth waiting for which is a BIG difference.
Actually I wish I could ALT-TAB out of the game... but it crashes every time, so thats not an option for me, but then again I'm not the one complaining.

If by 'changes' do you mean the scripting and general expandability... then yes, I do think that they are worth it. Simply for the fact that not only does it make the game better, but it opens the game up to be used for other people that want a 4x game, but maybe not the one that came in the box. Now, am I biased? Oh yeah. I have always found that any game that allows the user to change it, especially fundamentally, is worth a LOT of extra points. However, any time that you add that kind of flexibility something is going to falter. In this case, whats lacking is the base game is lack-luster (by that I mean that the AI isn't developed and just a lack of polish to the game all-around). Am I willing to put up with these short falls? Because of all the other offerings, yes.
Hopefully that all made sense.

Short answer: yes.

SpaceEmpires.net | Space Empires Wiki

Well, I guess I stand corrected... somewhat. I still believe that the changes are for the better.
I suppose the change to hex based is whole cause of increased movement processing... since the computer now has to process 50% more options for movement.
So is it the real-time combat that has caused the main increase in turn process time?

SEIV wasn't really that much faster, from what I remember. Once you got into late game you'd quickly get very long turn processing times in that game as well. Combat was mainly where SEIV was a fair bit faster, because of the turn-based solution it used.

Anyway, as I've said before, I might have been able to accept long processing times if they were necessary; if it was the only way to have the AI play well, or offer this degree of moddability in the game. The problem is, at the current time the long turn processing appear to be mostly a problem of unoptimized game code.

The Witcher is a game which suffered similar problems( from a player POV ). Loading times, going from inside a house to the outside, or from an outside area to another, would take forever, discouraging alot of players from playing the game. Now, one patch later, the loading times are suddenly less than a third of the original ones and the game has become much more playable. Yet none of the gameplay was lost in the process. This goes a long way in showing how big the difference between a optmized and a unoptimized application can be.

So now that the worst bugs have been ironed out, I'd love for Aaron to start optimizing all he can, and then go ahead and use the same game engine for the next Space Empires so that he wont have to go through it all again.

And he should really hire someone to assist him. Another designer/programmer/tester, even part-time, would probably do wonders for the game. One man doing everything( except art/music/sounds ) in a project this complex is way too ambitious, and it shows.

Ahaha, Aaron has made the comment "I find I can't add much to the game as I'm busy constantly re-writing it to fit with the latest programming tools/languages available every several years" or somewhere along those lines. Yeah, expanding beyond a 1-man project would be the solution but then you have management problems to learn about. All down to him in the end.

---Sig---
Playing Touhou games (Go to this Wiki if interested. No, nobody else is that good/insane as that replay). No rush for SE

There won't ever be other programmers for SE5; it was apparently way too late in the process to bring another programmer up to speed a year ago. SE6 should be getting a small team funded by Strategy First, though.

Raapys:
There is only so much you can do to speed up the resolution of a real time combat engine.. Aaron doesn't want to remove all the model rendering stuff that goes on in strategic combat, since they are necessary to determine the effect of bolts and torpedos (as well as everything else based on actual model collisions). Doing that would also result in weird visuals in the replay. For example, there has to be somewhere to place each beam as you fire it, so the replay doesn't have the beam clip through the firing ship model. It would be possible to forgo accurate visual replays and such, and resolve combat without all the real time clock rendering data at the speed of math, but Aaron does not want to do it.

Regarding movement routines, Aaron has made optimizations to them many times over the last 2 years. Maybe there is more he can do, maybe not.

SpaceEmpires.net | Space Empires Wiki

Hmm, if such is the case, then perhaps implementing/finishing multi-core support might be the best option? Given that CPU cores probably aren't going to become much faster in the next few years, it'd help greatly to at least be able to take advantage of 2+ cores.

Then again, I can see how adding multi-threading to a finished product would be rather time consuming.

Really? Sweet! I wonder where you heard that...

Then again, that might make SE6 suffer the "MOO3 Syndrome", where handing the series off to new developers will ruin the whole thing... hopefully Aaron will be able to keep them focused; Steve Barcia was sadly not included in MOO3, but if Aaron stays in charge of SE6...

~~~
The power of the ancients is now yours to command. You are truly the Master... of the obvious