Tried out the mapscript R2R_Erebus_1_09R, bu tit put the Civs WAAAY to close to each other, in the attached i had 5 civs just in that square area, on a Gigantic map??
Any way to put a tile limit space apart, in the python someplace??
It is pretty random. Sometimes they are close, sometimes they are spread out all over the map. Look at a few more and you should see what I mean.
On the other hand I only see 3 civs in your screen shot - two towards the bottom (purple and blue, based on the circle colors) and one up by the upper left of the WB toolbar (yellow). The bottom two are quite close for such a large map, but the yellow one is not as close as it seems due to the mountain ranges blocking direct travel for the most part, limiting the connection to one gap in the peaks. It is even possible that some civs that appear close to each other are, due to peaks and ocean, blocked of completely from each other until ships or mountaineering or until they have units that can travel through mountains via the mountain leader promotion, or things like that.
There is a mechanism for setting some preference values for things like the importance of distance in the placement - this isn't a hard limit, just a weight that influences the value of a region when placing them on the map (this was in the original Erebus script and is how it influences the placement of the various FfH2 civs with their widely varied characteristics). I even added an additional system for terrain and feature preferences. However, none of this is actually in use at this time. All the civs currently use the default settings.
If you wanted them to prefer to be farther apart, you could change the default weight for distance. Changing it from 2.0 to 3.0 would give them a significantly greater preference for being spread out.
These are the defaults as set in the __init__ function of the CivPreferences class:
self.idealAltitude = 0.35
self.idealMoisture = 0.7
self.needCoastalStart = False
self.allowForestStart = False
self.altitudeWeight = 1.0
self.distanceWeight = 2.0 #distance is most important for generic civs
self.regionType = "None" # QoW
So they all want the same things. Note that they do try to spread out due to the distanceWeight being set to 2, but since all the other preferences are the same they do still value the regions the same which tends to resist that preference. In general, they are going to want a grassland (with the occasional lush and a little muddy and even the occasional marsh plot) area due to that 0.7 moisture level, and all around the same altitude (which is semi-random in arrangement, but the water plots are all set to 0 and it tends to go up as you move away). This leads them to "clump up" in the greener areas of the map and mostly, but not all, in the regions next to the water areas. If all the "good" regions are taken, then the later ones will get spread out a bit more - it depends on the number of areas matching the preferences vs. the number of civilizations. Frequently there is enough that they can all get those green areas, but not always.
I may go through all the R2R civs and add preferences at some point, but currently it is just using the default values for all of them. There are some region types defined but since the civs don't have the basic preferences defined none of them have preferred region types either. They don't all need to be defined so I might just start with a few that should be on the coast and a few other historical placement types, like Mongols on plains, Egypt in the desert, and Incas amongst peaks. If just one or two of the civs on a map had preferences that were different, the rest would tend to have the room to be more spread out across the areas matching the default preferences.
If anybody feels like defining some preferences for civilizations it would be done in the GetCivPreferences function. There is a commented out example in there. The new region type data has some explanation of what the values mean in the __init__ function of the new CivRegionType class and there are already some types defined in the GetRegionTypes function.
The next release version of C2C is coming out soon (1st Oct I think). You should test it, 3 GB *should* be enough for a smooth gameplay (even without viewports).
I don't think it's a maf problem. It happens in the early part of the game, on a standard map with 18 civ. Memory usage is really good til now. 
