Home Trending Latest threads New posts Developer diaries

Technical Issue: I want multicore!

Jump to latest Follow Reply
  • We have updated our Community Code of Conduct. Please read through the new rules for the forum that are an integral part of Paradox Interactive’s User Agreement.
Crusader Kings II 69942 Bug Reports 19717 User Modifications 10604 Suggestions 5969 Tech Support 6373 After Action Reports
unmerged(144119)

unmerged(144119)

Private
1 Badges
Jun 17, 2009
20
0
  • Hearts of Iron III
I was pretty much disappointed when HoI3 wasn't dualcore ready in times where almost every serious game supported it. My hopes were high with Vic2.
Pretty much self-explanatory that I was disappointed too when Vic2 came out.
CK2 will be published in 2011/12, four to five years after EUIII.
I want an Engine update. I want P'dox to fully utilize my machine.
I know its hard and might cause trouble. But at the very least to expect is that a modern game runs better on a modern rig. Not as it is right now, where it doesn't matter what kind of CPU you use as long as it is a post-386 (exagerated, but you get my point.).

Grüße
 
That's not as easy as it sounds like. Taking a single-threaded imperative OOP program and turning it into a multi-threaded one is, in my experience, harder and more work than rewriting it from scratch as a multithreaded program, unless there are some really obvious ways to partition all the data.

Paradox wrote the Clausewitz engine to be the basis of all their new games, and as far as I can tell from the information I can deduce from it's modding capabilities, it can not be easily multithreaded beyond the obvious split of UI/data model.

In other words, we'd all like a pony.
 
As hard as it would be, I'd like to think PI would consider it. Having Multicore support would really make playing a game like CK easier.
 
Consider, yes. But never should it be able to stand in the way of a stable game, in playability that is. This is something paradox can be proud of when it comes to their games, they should strive to continue this.
 
It really isnt going to happen as has been stated it is too difficult to do and would require a complete rewrite of the engine. I have to say I have seen an improvement of the engine in the latest HOI3 patch and Vicky patch. And as CK2 is going to "probably" have fewer provinces than either of the previous games and fewer units, ie 1 unit per province, it should be relatively good performance.

But I share your pain I have a quad core machine with 8gb ram and run virtually every other game at full detailed settings and everything goes super fast but then I try late game HOI3 or Vicky, or even late game HOI2 and suddenly I feel I am playing on my speccy from 15 or so years ago.

Then again Paradox games are not the only ones who have the issue if anyone has played Harpoon classic trying certain scenarios in that for instance when controlling 4 CVBG in some of the later scenarios does cause issues at times.

I think you just need to grin and bear it, but as I said I think performance has really improved recently.
 
Tuna-Fish said:
That's not as easy as it sounds like. Taking a single-threaded imperative OOP program and turning it into a multi-threaded one is, in my experience, harder and more work than rewriting it from scratch as a multithreaded program, unless there are some really obvious ways to partition all the data.

Paradox wrote the Clausewitz engine to be the basis of all their new games, and as far as I can tell from the information I can deduce from it's modding capabilities, it can not be easily multithreaded beyond the obvious split of UI/data model..
Click to expand...

Great post. Just to expand on this: Johan and others at Paradox have stated that it would take too much of the developmental cycle to turn the Clausewitz engine into having multithread capabilities. In addition they feel that there still is a large number of systems of potential customers that don't have multithread capability, and that they see 3 to 4 more years as this being the case.


They said the next generation engine should be multithread, but this one will not
 
Tuna-Fish said:
In other words, we'd all like a pony.
Click to expand...

It's not about wanting a pony.:rofl:

As probably many of you I've spend thousands of hours playing P'Dox games. I got a lot bang for the buck so to say.:D
But i don't see any need on wasting time on a sluggish game engine that isn't kept up to date for mere convenience such as playability.

I do understand the point though that multithreading Clausewitz isn't possible at the moment. My point is: Then move on!:cool:

Because i won't buy CK2 if the retail box doesn't state its multithreaded.
Its a pain in the astonishingly well-formed parts of my body right now, playing Vicky2 or HoI3.:(
Sure it's fun. But it takes too long in late-game (and esp. MP) and spoils the game experience for me to that point, that I've stopped playing Clausewitz games at all. HoI2 is huge step backwards once you've seen HoI3, but at least it's running comparably smooth.

SAS said:
Great post. Just to expand on this: Johan and others at Paradox have stated that it would take too much of the developmental cycle to turn the Clausewitz engine into having multithread capabilities. In addition they feel that there still is a large number of systems of potential customers that don't have multithread capability, and that they see 3 to 4 more years as this being the case.
Click to expand...

Even my gf has a dualcore, my granny has a dualcore. My pals have duals or quads. The only one I know who doesn't have dual is my mom who hasn't got a computer at all. Dualcores are now fairly widespread. And in 3-4 years where are we then? Hell AMD released a server duodeca a while ago; Intel an octacore!
So this would be the point, where I stepped back from playing your future releases. 3-4 years from now on is a long time and I have better things to do to wait for the game to sort things out while reading my news.;)
I know I won't change your dev cycle but... actually I just wanted to rant...

EDIT: Just to put it straight: I do apreciate all the efforts Paradox takes to improve their games and what they have reached so far. In fact, I respect them more than any other game-developer. It's just that I want my pony, probably. :rofl:
 
Dual core didn't penetrate to budget builds until 2008 (and there's still far too many Pentium 4 refurbs kicking around).

On the other hand, bargain-basement builds can't play Clausewitz games anyway, Clausewitz requires a bit more out of the graphics system than those computers usually have. At the level of builds that would actually include a discrete GPU, I'd say since about late 2006/early 2007 they've been standard issue (Pentium D, Athlon X2 and Pentium Dual-Core). In other words, they were on the roadmap but not on my desk when they were first designing Clausewitz.

I do understand not multithreading Clausewitz if HOI3 really is as far as you'll push the engine (though it limits the headroom available for HOI3 mods). I think CK is the last major Paradox IP they haven't made a game for on Clausewitz (assuming CK was Europa). I'm less enthused with the idea of not getting a really threaded engine until 2014.

Given the amount of 'stuff' that happens each tick, they should be able to extract some reasonable parallelism out of the next engine. Combat in Stalingrad for example, won't affect the combat in the Sahara, nor will it affect what the weather will be in Chile. Nor should what Henry the VIIIth is thinking right now about that saucy noblewoman affect the Pope's decisions about trade policy (at least not until his spy network tells him next hour).

There's some things you just can't do at the same time though. Take for example moving ships and spotting. You can have all the ships move and then all the ships spot, or the other way around, but you cannot have some ships move, and then spot, and then the rest move and then spot (especially if which ships move first is nondeterministic). If you did that, sometimes WWI BBs would catch up with fast DDs in calm waters, since they got to do all their movement before the DDs. You can however probably thread the 'all the ships move' and 'all the ships spot' parts.

In the near term, both major CPU manufacturers have come up with ways for multicore CPUs to scale clockspeed up significantly when working on poorly-threaded tasks, which should help for those of us getting 4+ cores.
 
Last edited:
I doubt there will be multicore support, but at least some things are fairly easy to move to separate threads. For example whole event system.

However if the engine was not designed with multithreading support in advance, it's extremly error prone. And does not actually warrant an effort.

But I really hope EU4 will support multithreading.
 
Zolotaya said:
Great post. Just to expand on this: Johan and others at Paradox have stated that it would take too much of the developmental cycle to turn the Clausewitz engine into having multithread capabilities. In addition they feel that there still is a large number of systems of potential customers that don't have multithread capability, and that they see 3 to 4 more years as this being the case.


They said the next generation engine should be multithread, but this one will not
Click to expand...

Right, I think most of us understand that converting Clausewitz to multicore isn't technically feasible. The question we all went to know is when you're going to give Clausewitz a state funeral with the full honours it deserves and move on to building your next generation engine from scratch?

Clausewitz started seeming really dated as of HOI3, which admittedly is the game that pushed it the most. Even so, it runs like molasses on my machine. I have 8 cores and almost everyone has at least 2. Computer technology is just going to keep advancing even further over the next few years making its datedness more and more apparent.

I realize it most be frustrating to invest so much time and money into an engine only to see it become obsolete in a few years, but the reality is that these sorts of fast technology cycles and part and parcel of the industry you're in.

On the bright, side think about all the cool games you'll be able to make when you can make features work more complexly, with more calculations, without having to worry as much as you have to now about causing a big hit to performance.
 
Zolotaya said:
Great post. Just to expand on this: Johan and others at Paradox have stated that it would take too much of the developmental cycle to turn the Clausewitz engine into having multithread capabilities. In addition they feel that there still is a large number of systems of potential customers that don't have multithread capability, and that they see 3 to 4 more years as this being the case.


They said the next generation engine should be multithread, but this one will not
Click to expand...

This is true. I have a single core processor. However, I think its a mistake to have the developers over-extend passed the engine's capabilities. The engine obviously has its limits, so don't push them and unnecessarily create slowdown.
 
If the Clausewitz engine has 3-4 years left on it, and CK2 comes put in early 2012, the engine will have 1.5-2.5 years left. Assuming their policy of one new game a year continues, maybe we'll get a new time period for the Grand Strategy games before the new engine (and presumably EU4) is launched! I'd love to see a Cold War game. It would be like Vicky but with espionage, nukes, and a potential WWIII! So like a V2/HOI3 hybrid. :D
 
I think I read in the forums some time ago that HOI3 is multithreaded, but not where it counts: ie the Clausewitz-part. PI can't scale their games, and they can't afford to lock the singlecores out just yet.Instead I would prod PI to make the most out of 4+ GB ram and SSD disks, and some other small things they can do today.

With further advances in chip design we'll eventually have faster clockspeeds (smaler nodes) and a higher IPC. And of course we'll see 6+ physical cores in budget builds.
 
Meh, the release of eu3 was just bad timing that's all. Intels dual cores came out about the same time paradox were developing their new engine. They had no idea it would be such a success. If they could go back and do it again I'm sure multiple core support would be in.

As it is now, they'll just have to limp by best they can and try to squeeze what they can out of this engine. Tell it like it is. Trying to justify the lack of multiple core support by saying that it's not that common, is laughable. Multiple cores are more common today than pixelshader 2 ever was back in 2007.
 
I just wonder: Did Paradox say that CK2 will still use the Clausewitz Engine?

For HOI 3 and Vicky 2 it took around 9 months to a year to develop them. For CK 2, it should be 18 months - perhaps because they also develop a new engine?
 
You must log in or register to reply here.
Top Bottom
Back