A New Dawn Beta Builds

I have played on large maps with 1,5Gb RAM well into the modern and future eras with ROM 2.92 and AND 1.7x series without getting that many MAFs (50-100 Turns, more when Alt-Tabbing or playing windowed). Point is moot anyways since turn time by then was 5+ mins. :lol:

PS: 3GB switch was active and I have since then gotten a new pc.
 
i know is prefered to have more -

but how can you explain that im in civ4 for 6 years - and never hit a maf before?

specially with my mod with and and without it....

something in the new code must have something to do with my new mafs, there is no other explanation, i havnt installed any software, program or made any changes to my cpu or the compiling sequence im using on the dll.

are there some more logs i can upload that will help?

*************
Stormwind, so are you also getting mafs now??
can anyone else confirm this?
 
The more I think on this, the more I like this idea. Basically, it steals one of the good ideas from Civ5, declaring support for or against Civ's, and adds on it (secretly allying against a particular civ/group of civs).

I'm going to release Beta 2 soonish tonight, but I will work on adding this for beta3 for next weekend. ;)

I also like the idea but I think war declarations and pacts should not only look at military power but also take into account if each civ has an embassy and trade agreements, and most of all possible power.

For example I'm playing 10 civs 6 continents large map. Only friend is catherine on another continent. 4 trade agreements with her but a i have a small army running pacifism. Alexander attacks me along with catherine which is a nice sneak attack on a main city, but i have 21 cities in 1500 ad and about even in tech so the attack did nothing for either of them but ruin my relationship with catherine...she should of known it was doomed and not attacked. Take into account distance, number of cities defender city has and make a projection of possible power.
 
@DudeSDCA & Dancing Hoskuld

It's due to Koshling's canConstruct caching. I'll see what I can do it fix it.

Also, It seems I am going to miss the Beta 2 update. I have everything ready, but I found a critical CTD in the DLL that appears within 300 turns. Koshling and I are working on a fix, and should have a release before Wend.

Did you manage to fix this? If not don't spend your time on my bugs - that's my responsibility. Assuming you didn't get this fixed I'll work on it shortly. The problems are (from memory and making assumptions based on the described problems):

1) Queueing a building doesn't invalidate the cache (building one does, but I forgot about queueing)
2) Acquiring techs via goodie huts similarly won't fluch all the city caches (should)

Thanks for the bug reports folks :) This is why we beta test!
 
Hey Guys good work again.

But there is one question left you wrote of speed improvements in 1.76b.
For me it would be important why in the late games of huge maps with 12 civs the roundtime takes about 2 minutes ?

e.g yesterday I started a LAN Game with a mate on a huge map 12 civs including my mate and me.
In the year 2004 the game started getting slower and slower in the year 2020 it took 2 minutes at least for each round.

I know you guys workin a lot on that.

PC:

Phenom X4
4GB Ram

Yeh, it's these later game slow downs we'er especially trying to wok on. They are due to multiple factors and we're slowly chipping away at them. Expect each release to improve somewhat in this area, but it's never going to be the case that some slowdowns will no occur as the game develops I'm afraid. You should find that 1.76 is significantly better than 1.75 in this regard, and as beta goes on the situation should improve further, but obviously there comes a point where you hit diminishing retruns.
 
i know is prefered to have more -

but how can you explain that im in civ4 for 6 years - and never hit a maf before?

specially with my mod with and and without it....

something in the new code must have something to do with my new mafs, there is no other explanation, i havnt installed any software, program or made any changes to my cpu or the compiling sequence im using on the dll.

are there some more logs i can upload that will help?

*************
Stormwind, so are you also getting mafs now??
can anyone else confirm this?

Think we've found the cause of this one. Afforess emailed me some details of a case he saw, and I have pinned down what I believe to be the cause in the code. Hoping to fix it later today, so if you're pulling source from SourceForge and building your own DLL you can pull the fix this evening with a bit of luck. Not sure when Aff plans to circulate a binary for the majority that don't build their own DLL.
 
hi koshling,

im very glad to here youve located the problem. and more over - that it isnt due to my pc specs...

i indeed updating from the svn.

im checking it every day :)

thank for all the work your putting, and helping the mod.

if you need some logs, saves anything from me to help, just name it.
 
The more I think on this, the more I like this idea. Basically, it steals one of the good ideas from Civ5, declaring support for or against Civ's, and adds on it (secretly allying against a particular civ/group of civs).

I'm going to release Beta 2 soonish tonight, but I will work on adding this for beta3 for next weekend. ;)

I am glad you like my idea.

Another thing I have noticed that the AI is pretty bad at times when placing cities. I have seen a few cities that, if only moved one tile left or one tile right, would gain access to 3+ resources. In vanilla BTS this isn't such a big deal, as cities weren't required to have the resource within the cities vicinity in order to build the buildings related to it.

That being said, I don't think that the AI weighs a resource properly for this mod. Might want to see if it is possible to edit the AI to see if you can weigh them differently.

If editing the AI in this way is either not possible or not practical, here I think is an amenable workaround which would solve the problem of the AI being too stupid to ensure that a resource is actually within a city vicinity:

Make each building which requires the presence of a resource be tied to the amount you have rather than vicinity.

Here is an example of how it would work, as I am typing hurriedly so I doubt it is totally clear...

It is the beginning of a game. I have two cities. I was lucky and my first city had stone in the vicinity, and I built a Stoneworker's Hut here before settling the second city. I am now ready for a third city, and get really lucky and place it next to a stone resource AND a marble resource. Now currently I would be able to get 2/3 cities with Stoneworker's Huts... The first and third cities. However, in reality I now have two stones and one marble. 2 + 1 = 3. Of course the way it would factor out is two cities would receive the stone bonus of the Stoneworker's Hut, and the third would only receive the marble bonus. The first round of resources allocated would not be that big of an issue, as the difference between the two types of bonuses are minute...

However, once your borders start expanding and you start gaining more stone/marble/obsidian than you have cities, you start to have a problem... What if I have 3 cities, one with stone, one with marble, and one with obsidian? What if I want all three resources to be diverted to my Capital, the city which would benefit most from the added production? With this addition I have mentioned, I could artificially ensure that the Capital would get all 3 by not building Stoneworker's Huts in the other two cities, thereby forcing the game to send the benefit to the Capital, the only city with the Stoneworker's Hut... But I think an ability to in some way control this would also help... However this to me would be a luxury, not a necessity.

That is how it would work, as the AI is too stupid to properly place cities with the appropriate resources within the city vicinity, something which is mandated by this mod.

As an aside, I don't think the RoM mapscripts give enough priority to the early-game resources. What the hell am I going to do on a continent with 10 oil, 10 coal, 10 bauxite ore, and 10 stone? Why can't it be 5 oil, 5 coal, 5 bauxite, and 25 stone? Truthfully I often edit my own maps after generating them, and apportion the appropriate resources to the appropriate locations, ensuring that all Civs actually have stone... However this is something the mapscript should do, I should not have to do this manually.

Oh and guys... Just go buy more RAM. I doubled my RAM from 4 to 8 just to play this mod. It's cheap as dirt and is by far the easiest piece of a PC to upgrade.
 
Did you manage to fix this? If not don't spend your time on my bugs - that's my responsibility. Assuming you didn't get this fixed I'll work on it shortly. The problems are (from memory and making assumptions based on the described problems):

1) Queueing a building doesn't invalidate the cache (building one does, but I forgot about queueing)
2) Acquiring techs via goodie huts similarly won't fluch all the city caches (should)

Thanks for the bug reports folks :) This is why we beta test!

1) Yes. Queuing a building doesn't invalidate the cache, so you can add multiple copies of the same buildings (and, consequently, buildings don't go gray when you add them to the queue).

2) This is not just for goodie huts (but the problem is similar). The queues appear to be updated on end turn, not in real-time. So acquiring a goodie hut doesn't give you the new buildings. Similarly, using a missionary to spread a religion doesn't give you access to the new religion's temples. I think I remember this even happening when I connected a resource to my city the first time (that butchering shop didn't show up until the next turn).

Hope this helps.
 
Make each building which requires the presence of a resource be tied to the amount you have rather than vicinity.

Here is an example of how it would work, as I am typing hurriedly so I doubt it is totally clear...

It is the beginning of a game. I have two cities. I was lucky and my first city had stone in the vicinity, and I built a Stoneworker's Hut here before settling the second city. I am now ready for a third city, and get really lucky and place it next to a stone resource AND a marble resource. Now currently I would be able to get 2/3 cities with Stoneworker's Huts... The first and third cities. However, in reality I now have two stones and one marble. 2 + 1 = 3. Of course the way it would factor out is two cities would receive the stone bonus of the Stoneworker's Hut, and the third would only receive the marble bonus. The first round of resources allocated would not be that big of an issue, as the difference between the two types of bonuses are minute...

However, once your borders start expanding and you start gaining more stone/marble/obsidian than you have cities, you start to have a problem... What if I have 3 cities, one with stone, one with marble, and one with obsidian? What if I want all three resources to be diverted to my Capital, the city which would benefit most from the added production? With this addition I have mentioned, I could artificially ensure that the Capital would get all 3 by not building Stoneworker's Huts in the other two cities, thereby forcing the game to send the benefit to the Capital, the only city with the Stoneworker's Hut... But I think an ability to in some way control this would also help... However this to me would be a luxury, not a necessity.

This is a core Civ V concept (limited resources tied to production) - I believe Civ V focuses it more on tying resources to units, whereas above you're proposing to tie it to certain buildings.

Its an interesting concept - the only two concerns I think I'd have are: overpowered cities and balancing. If I could "designate" several of my cities to be power cities by building all the resource-required buildings (stoneworks, butcher, tannery, grocery, ironworks, etc.), I could really stack up the bonuses (maybe even up to +100% unit production, for instance). So I think this would required some significant re-balancing of building bonuses to compensate...

Ultimately, I'd focus on making the AI smarter than re-tooling the way the buildings work, but that's just my humble opinion!
 
I am glad you like my idea.

Another thing I have noticed that the AI is pretty bad at times when placing cities. I have seen a few cities that, if only moved one tile left or one tile right, would gain access to 3+ resources. In vanilla BTS this isn't such a big deal, as cities weren't required to have the resource within the cities vicinity in order to build the buildings related to it.

That being said, I don't think that the AI weighs a resource properly for this mod. Might want to see if it is possible to edit the AI to see if you can weigh them differently.

If editing the AI in this way is either not possible or not practical, here I think is an amenable workaround which would solve the problem of the AI being too stupid to ensure that a resource is actually within a city vicinity:

Make each building which requires the presence of a resource be tied to the amount you have rather than vicinity.

Here is an example of how it would work, as I am typing hurriedly so I doubt it is totally clear...

It is the beginning of a game. I have two cities. I was lucky and my first city had stone in the vicinity, and I built a Stoneworker's Hut here before settling the second city. I am now ready for a third city, and get really lucky and place it next to a stone resource AND a marble resource. Now currently I would be able to get 2/3 cities with Stoneworker's Huts... The first and third cities. However, in reality I now have two stones and one marble. 2 + 1 = 3. Of course the way it would factor out is two cities would receive the stone bonus of the Stoneworker's Hut, and the third would only receive the marble bonus. The first round of resources allocated would not be that big of an issue, as the difference between the two types of bonuses are minute...

However, once your borders start expanding and you start gaining more stone/marble/obsidian than you have cities, you start to have a problem... What if I have 3 cities, one with stone, one with marble, and one with obsidian? What if I want all three resources to be diverted to my Capital, the city which would benefit most from the added production? With this addition I have mentioned, I could artificially ensure that the Capital would get all 3 by not building Stoneworker's Huts in the other two cities, thereby forcing the game to send the benefit to the Capital, the only city with the Stoneworker's Hut... But I think an ability to in some way control this would also help... However this to me would be a luxury, not a necessity.

That is how it would work, as the AI is too stupid to properly place cities with the appropriate resources within the city vicinity, something which is mandated by this mod.

As an aside, I don't think the RoM mapscripts give enough priority to the early-game resources. What the hell am I going to do on a continent with 10 oil, 10 coal, 10 bauxite ore, and 10 stone? Why can't it be 5 oil, 5 coal, 5 bauxite, and 25 stone? Truthfully I often edit my own maps after generating them, and apportion the appropriate resources to the appropriate locations, ensuring that all Civs actually have stone... However this is something the mapscript should do, I should not have to do this manually.

Oh and guys... Just go buy more RAM. I doubled my RAM from 4 to 8 just to play this mod. It's cheap as dirt and is by far the easiest piece of a PC to upgrade.

Changing the vicinity requirement would be bad IMO. We need to attack this in the AI. Shouldn't be too terribly hard. However, are you sure that the resources the AI is 'just missing' were ones that that AI player had revealed at the time of city founding? It doesn't cheat so it has no clue that such resources exist until it acquires the necessary tech.
 
Changing the vicinity requirement would be bad IMO. We need to attack this in the AI. Shouldn't be too terribly hard. However, are you sure that the resources the AI is 'just missing' were ones that that AI player had revealed at the time of city founding? It doesn't cheat so it has no clue that such resources exist until it acquires the necessary tech.

Any reasons why the vicinity requirement would be bad? Even though I wrote that idea on a whim while contemplating the dumb AI city placement issue, the more I think about it the more I like it. If you have some examples why this would be bad let me know so I can think about them too.

As for the revealing issue, I see your point. When I was noticing the city placement that was not on my mind... I'll double check that as I have plenty of saves at various eras to glance through... However... You might have just revealed another solution... Perhaps the AI should cheat, but only in this respect since truly appropriate city placement is vastly more important in RoM:AND than it is in vanilla BTS.
 
1) Yes. Queuing a building doesn't invalidate the cache, so you can add multiple copies of the same buildings (and, consequently, buildings don't go gray when you add them to the queue).

2) This is not just for goodie huts (but the problem is similar). The queues appear to be updated on end turn, not in real-time. So acquiring a goodie hut doesn't give you the new buildings. Similarly, using a missionary to spread a religion doesn't give you access to the new religion's temples. I think I remember this even happening when I connected a resource to my city the first time (that butchering shop didn't show up until the next turn).

Hope this helps.

Actually there IS code to cope with the queuing (it just invalidates the cache entry for THAT building type), and it seems to be working fine for me (I can't seem to multiply queue the same building). There are problems with tech acquisition and religion founding however, which I have fixed (not pushed to SVN yet though).

I'd appreciate details on the multiple queuing issue since I can't reproduce it - is it specific cities/specific buildings or are you seeing it as a general issue?
 
Any reasons why the vicinity requirement would be bad? Even though I wrote that idea on a whim while contemplating the dumb AI city placement issue, the more I think about it the more I like it. If you have some examples why this would be bad let me know so I can think about them too.

As for the revealing issue, I see your point. When I was noticing the city placement that was not on my mind... I'll double check that as I have plenty of saves at various eras to glance through... However... You might have just revealed another solution... Perhaps the AI should cheat, but only in this respect since truly appropriate city placement is vastly more important in RoM:AND than it is in vanilla BTS.

Well, I am first and foremost a C2C modder, but we are converging the DLL aspects which is why I am merging my changes back into AND (and visa versa), so for me it's the C2C perspective that drives this opinion. The entire goods system in C2C is predicated on the in-the-vicinity resource code, and changing something that fundamental would have huge ripple effects there, which is why my gut reaction was negative. It would really need Dancing Hoskuld or Hydromancerx to comment on how it might play with the C2C goods system.

I also don't like (as a general principal) allowing the AI to cheat. Slippery slope. I DO agree with you though that we may need to improve the placement AI if it's not just a reveal issue.
 
hi koshling,

im very glad to here youve located the problem. and more over - that it isnt due to my pc specs...

i indeed updating from the svn.

im checking it every day :)

thank for all the work your putting, and helping the mod.

if you need some logs, saves anything from me to help, just name it.

Three changes just pushed to SVN:

1) Fixed the memory leak that was causing MAFs
2) Fixed some issues with the canConstruct cache (though multiple queuing I couldn't reprodice as per above post so may be still outstanding)
3) Fixed incorrect ability to work mountain peaks in the city cross until you discover mountaineering
 
I wish I could do anything for you guys and for the mod.
This is the greatest mod for a game I ever saw.
 
I wish I could do anything for you guys and for the mod.
This is the greatest mod for a game I ever saw.

Play. Test. Report issues. That's the help that's most needed ;-)

Also (/blatant advertising on) check out C2C which is a mod mod of AND
 
Top Bottom