Ok, I will try to summarize what I (believe to) know
AGAIN.
(I am not pissed. It is just a bit frustrating for me to discuss the same issues again and again.)
As I said, we - several experienced Civ4Col modders and a few experienced Civ4 modders - had been exchanging on this topic several times over the years:
(Mostly outside of normal public forums, since most players would not have been interested in such technical discussions.)
1. Using FPKs
significantly improves loading times.
That is easy to verify.
2. Using FPKs is a bit of a hassle, when still working a lot on graphics.
But that is an aspect of modding and does not affect the players.
3. Loading times extremely depend on your hardware and OS.
Things are quite different for 64Bit environments with SSDs and very strong graphic cards for example.
(Loading RaR with cleaned cache takes about
12s on my machine.)
4. It is possible to mix FPKs with normal folder structure.
Data in normal folders will override data in FPKs.
5. When using FPKs there are some "general advices" gathered over the years:
(Maybe those are not correct anymore today though because hardware and OS have changed so much.)
- It is better to have 5 FPKs of 100MB than 1 FPK of 500MB
(Try to do your packaging wisely.)
- Some graphics (e.g. some animated screens or some movies) do
not work when packed in FPK
- Some glitches with normal graphics (units / buildings / improvments) could be solved by moving the graphics out of FPK to normal folders again.
(It is not know exactly though if those issues were really related to FPKs. Similar things were reported without FPKs as well.
Could have been some temporary problem of the graphic card.)
6. The way graphical caching is done by Civ4-Engine is generally unknown to us modders.
There are many speculations about that, but nobody I ever talked to really knew for sure.
7. The concept of FPKs is quite old (10years).
However, I don't know any other game than Civ that uses them.
-------
Here it already ends with "facts".
The rest is experience, opinions and basic tests several of us had done.
(We did not use advanced methods of testing to verify.)
Those tests had been done quite some time ago though.
(Last discussions about FPKs I remember were more than a year ago.)
Most of us agreed that FPKs caused minimally worse performance ingame.
We are talking here about 5% to 10% longer Autoplays for example.
I ran 4 Autoplays with FPKs and 4 Autoplays with normal folder structure and then compared the times.
(The
average time for Autoplays was better with normal folder structure. Don't know the exact numbers anymore though.)
Nobody ever said that FPKs improved performance ingame though.
(I also could not think of any reason, why it should.)
But all of this is more "feeling" than really "scientific testing".
Since so many experienced modders generally agreed on this, I do trust in what I heard.
-------
As I said, everybody shall do whatever works best for him.
(Using normal folder structure without FPKs works best for me though.)