32bit has nothing to do with AI processing. The bit size is like the size of a plate. If you eat a lot, it helps so you don't need two plates, but if your portion is small enough that 32 bit plate is enough, it makes no difference. The reason AI takes a long time is because it has to do everything sequentially. How do you contemplate a good place to settle if you don't know the surroundings? That alone is already sequantial. Then the combat, then building order, and so on and so forth
In comparison to civ 4, 5 had slow turn times even early game. Civ 4 has almost instantaneous turn times on standard maps early to mid game while civ 5 can take a couple of seconds. Late game becomes even worse for civ 5. Sent from my HTC Desire 820s dual sim using Tapatalk
Civ 5 was poorly optimized (performance wise). It is that simple. Hopefully civ 6 will change this with new 64 bit engine & better utilisation of modern processors. I am cautiously optimistic though, Xcom 2 has pretty crappy optimization which is another recent Firaxis title. Let's hope they get it right this time. Sent from my HTC Desire 820s dual sim using Tapatalk
His point holds even if you back-adjust for era appropriate hardware running the game. 4 wasn't particularly good, but 5 is optimized that badly.
Something running badly doesn't mean that it was poorly-optimised. "optimisation" is thrown around as a catch-all phrase for magically improving game performance. Sometimes the base systems simply don't scale well, leading to bad performance under certain conditions (larger maps, longer games, more units, etc, et. al). In these situations "optimisation" won't help, because the code can't be "optimised" any more than it already has been. You can call that poor design, or poor implementation, or whatever you want. But people need to stop misusing "optimisation" like the developers never attempt it or something.
I cannot recommend this enough. If not viable with civs, this should at least be done for city states. There are 16 city states on a standard sized map, which themselves cause a large amount of lag by shuffling their units every turn without any purpose to it
This would be viable if sensible unit stacking was allowed. But as long as 1UPT is here to stay, the AI needs to move its units to prevent stupid traffic jams we can't solve.
I'll go with both poor implementation and poor optimization on the topic of turn times. If optimization were given priority by Firaxis, they would not be trivially outperformed by mods. Turn times in 5 were a tremendous hidden cost of playing it. I'm glad people are noticing, even if this discussion fell on too many deaf ears years ago when I pointed out that players were waiting hours extra per game compared to the previous title, leaving both on recommended hardware. It is the single greatest factor in me not playing it and not insignificant for others. What I wonder is how many people it drove away without them clearly identifying it as the reason, instead citing "immersion" or "boring" or "not as much to do" etc.
If a game is running sub-optimally on default settings on (large map, 12 civs, 24 city states) and you still have horrendous turn times, it is poor optimization. Obviously if you increase the number of players and add mods then you can't argue that, but the game was still badly optimized even on GOOD machines where I still saw theg rey tiled and the switching of players in Hotseat caused a 1-3 second break to activate the other player only for the map to be completely re-rendered (which affected me enjoying the mode). That is poor design and optimization.
I realise that, which is why I said for city states only. There is no point to them moving all their units every turn, except when at war (which, lets admit it, is quite rare in vanilla). It doesnt even solve traffic jams. That being said, I'm glad firaxis is toning 1UPT down for 6.
I've had city states responsible for traffic jams on more than one occasion. Yes, I would like them to move their units out of the way, please.