To be continued...

[Zuikaku]

Please, just polish the SRCW! there are just too much WW2 games already. And too few modern/cold war era global strategies. The market is flooded with WW2 era games....

[mrgenie]

So, for a Supreme Ruler - Fall of Soviet Union.. (1985 - 2000)
add all the units and specifics of that era...

There is a mod coming out that is close to the Fall of the USSR

awesome, keep me informed )

Also, may I suggest the creation (even if it's not realistic, but at least to have the possibility)
to have the EU, Arab league, OAS, GUS, SCO, ASEAN, etc as military blocks.. or at least to choose
upon starting the game if you want to enable/disable these blocks..

http://en.wikipedia.org/wiki/List_of_in ... anizations

[Tremere]

We have said previously that there is intention of a WWII era Supreme Ruler but that anything earlier is unlikely to be an appropriate fit for this game engine. You'll have to wait for some announcements.

Good news! New SR will be best WWII game! Good luck!
I hope it will occupy not much time.

Could you tell about any changes in engine?

[number47]

IF BG decides to make a new game, than that game definitively needs new engine as this one has reached it's limits. And hex size needs to be much smaller than current 16km...

[Tremere]

IF BG decides to make a new game, than that game definitively needs new engine as this one has reached it's limits. And hex size needs to be much smaller than current 16km...

You are right!
New engine necessary to develop.
During WWII more massive armies took part in war. Millions soldiers from both sides.

[mrgenie]

IF BG decides to make a new game, than that game definitively needs new engine as this one has reached it's limits. And hex size needs to be much smaller than current 16km...

You are right!
New engine necessary to develop.
During WWII more massive armies took part in war. Millions soldiers from both sides.

what do you mean? In one of my games I played germany and conquered everything up to the urals when Russia gave up and the country was split between
germany, japan, china, korea...

I've send an army force of roughly 2 Million troops into the soviet union. Not sure what you mean, but Adolf didn't send much more then I did..

About hexes smaller then 16km.. dmm, you do you know if you make them hexes like 8 km, thus 4 times more hexes, you're probably going to blow any current CPU-GPU system,
noting that you already gave a hint that on your system the game can't handle millions of troops?

I really doubt the game engine can be much faster, compared to other games on the market this is already the best performance wise. It's not worth spending zillion hours of work
to increase a 10% performance win.

Although they might consider using GPU as CPU power. With a Quad-GPU system the game will speed up very much, IF a lot of the coding can be vectorized. If not, you won't gain much boost
from this either. And that IS more work then a small studio as BG can handle I think. But yes, there might be a huge advantage for future development.

[number47]

I believe this part refered to my post.

About hexes smaller then 16km.. dmm, you do you know if you make them hexes like 8 km, thus 4 times more hexes, you're probably going to blow any current CPU-GPU system,
noting that you already gave a hint that on your system the game can't handle millions of troops?

Why connecting map size (hex size) with number of units? I fail to see how the current hex size is limiting unit numbers. Can you explain what you mean?

If we misunderstood each other, I'll explain what I meant with smaler hexes. Current hex size of 16km limits any tactical movement of troops near the frontlines. As it is now, you need to keep arty at least 2 hexes (16-32km, depending how you count the distance, are the units in question at the closest end of hex to each other or in the middle of the hex) from enemy infantry unit which is extremly wrong to say the least. Inf shouldn't be able to pose any threat to arty unit if it is farther than 1-2km from it...
The arty-inf example is used just to clarify my point but it is applicable to most other unit combinations since unit attack ranges go from 1-8km for land units...

Or in other words, current hex size ingame allows the following situation:
-one unit in the center of a surface of 3x3 hexes, can cover (instantly interact with any unit entering neighbouring hexes) an area of 2304 square km (16+16+16=48km; 48x48=2304 square km).
Kind of seriously wrong, wouldn't you agree?

EDIT: and when I said "new engine" I didn't mean "graphical engine" I meant "game engine"

[Tremere]

IF BG decides to make a new game, than that game definitively needs new engine as this one has reached it's limits. And hex size needs to be much smaller than current 16km...

You are right!
New engine necessary to develop.
During WWII more massive armies took part in war. Millions soldiers from both sides.

what do you mean? In one of my games I played germany and conquered everything up to the urals when Russia gave up and the country was split between
germany, japan, china, korea...

I've send an army force of roughly 2 Million troops into the soviet union. Not sure what you mean, but Adolf didn't send much more then I did..

About hexes smaller then 16km.. dmm, you do you know if you make them hexes like 8 km, thus 4 times more hexes, you're probably going to blow any current CPU-GPU system,
noting that you already gave a hint that on your system the game can't handle millions of troops?

I really doubt the game engine can be much faster, compared to other games on the market this is already the best performance wise. It's not worth spending zillion hours of work
to increase a 10% performance win.

Although they might consider using GPU as CPU power. With a Quad-GPU system the game will speed up very much, IF a lot of the coding can be vectorized. If not, you won't gain much boost
from this either. And that IS more work then a small studio as BG can handle I think. But yes, there might be a huge advantage for future development.

I wrote about massively armies that cannot be displayed corerectly in 16 kilometres hex system. I don't think that increasing of number hexes will slow down game so seriously.
7 units for one 16 kilometres hex is very small. I can mistaken, but 7 infantry units is about 5000 soldiers, less than two regiments. On the front line concentration of forces was very high sometimes.

[Ragu]

I've send an army force of roughly 2 Million troops into the soviet union. Not sure what you mean, but Adolf didn't send much more then I did..

Barbarossa was around the over 3.9 million axis personnel. That's a few more than 2 million

Edit: Actually the numbers are a bit more than that.

Divisions: 166
Personnel: 4,306,800
Guns and Mortars: 42,601
Tanks: 4,171
Aircraft: 4,389

[Chauncey]

Which really suggests that unit size should go up if anything. As I've said before, battalion-scale (w/o any option of combining into larger units) gets very awkward and inefficient when dealing with large numbers of units. The mere thought of trying to organize and execute a Barbarossa-scale offensive, battalion by battalion, gives me a migraine.

Because I know some joker is going to accuse me of it, I might as well state now that I'm not advocating turning SR into HoI. The latest iteration of HoI is division-level if I remember correctly. There ARE options between division and battalion. Not to mention SR is different in enough other respects that it would make for a totally different style of game.

[Balthagor]

We are getting way ahead of ourselves. We haven't even closed SR-CW or announced if WWII will be our next project...

[number47]

We are getting way ahead of ourselves. We haven't even closed SR-CW or announced if WWII will be our next project...

We got bored of talking about bagels

[Chesehead]

I'd vote for some sort of 4D space game.. I think there is a lot of potential there compared to WW2 only because it's been done so much and I think the SR engine is better suited for long term matches compared to the ~6 years of WW2.

[Chauncey]

We are getting way ahead of ourselves. We haven't even closed SR-CW or announced if WWII will be our next project...

We got bored of talking about bagels

True.

I also agree the space game idea could be great.

[mrgenie]

a space game? I think you're talking about a whole new development and not about enhancing supreme ruler But yeah, a space game with massive armies would be nice ))

Anyway, what do hexes have to do with CPU performance seems to be the question here.

Since i have no insight in SR SDK, I don't really know the effect on this game. But generally the most CPU eating task of any strategy game is the unit path method. Searching for where to go to and find the best possible solution to get there.

In Civilization, where I worked a lot on the SDK i can guarantee you that on small maps this makes up maybe 20% of the total CPU power and lets say it takes 1.000.000 operations to do the math (I didn't actually count it, but let's just assume)
if you take a map the size twice, the path-finding now takes up 60-70% of the total CPU demand, with about 4-6 times the amount CPU power (depends on a lot of factors like land units not able to cross water or hills, etc) that's why some mods can make the game really slow if they start to mess around with the path finding method or allow units to enter water areas or worse, building tubes into oceans.
if you take the size once more twice, you're not 4-6 times the amount again, but around 8-10 times more CPU, on top of factor 4-6 we already had.. so even while your number of units remains the same, your total amount of CPU demand for handling those units especially the path finding simply explodes.

As I watched SR, they took a different approach, not path-finding the whole map, but regions depending on events triggered by other events. You might think this thus wouldn't be effected by the amount of hexes, but that's wrong. The game still must be close to reality, so if an event triggered by some other event 160km away has 10 hexes to do the path-finding, the moment you decrease it to 8, you obviously have to do the path-finding over 20 hexes.. and that's of course square, so 10x10 = 100 while 20x20=400.
So in an ideal situation with perfect coding and everything going perfect you'd have to handle 4 times the amount of codes. Unfortunately the perfect path-finding theory doesn't exist in numerical mathematics, it's about approximation. So you'll have to assume on small numbers the doubling leads to 4-6 times more cpu demand while on larger distances you'll get to 8-10 times or worse..

But like I said, I don't know the SDK, maybe BG took a good approach to get closer to theorerical values, but you're still, if you want to keep it realistic, triggering events depending on other events for some realistic distances in km/miles and thus doubling the amount of hexes between 2 points for path-finding, and thus in an ideal situation eating 4 times more CPU performance.

And yes, to get from A to B under circumstances can even require more then Distance * Distance to get path-finding running well, especially if you have obstacles in the landscape..

I can understand BG won't release this part of the SDK as in my humble opinion they did the best on this part out of all the games I know off, maybe all the games on the planet even. But even the best engine can get crippled if you're going to explode the number of nodes to do the path-finding on, even if it's the best programmed method on the planet.


In short: I think the map is large enough to get close enough to reality. Diversity of the game and getting rid of bugs is more important then enlarging a map that's already one of the largest in this branch of the gaming industry.