Poll 32 vs 64 bit

[Runite]

Most people don't understand anything about computer architecture, so they have a misguided sense of what is important for performance. I got eviscerated for saying that the last thread ~_~

 

[Gwayne]

It is interesting to see that the majority of the voters has a 64 bits OS. And I still wonder why people are so furious that this game should stay 32 bits

My laptop is running 64 bits OS since 2004 (linux) and 2008 (windows)

The advantage of direct access to more than 4 gig memory would be huge.

I have seen the changes of my 25 years of IT

 

[SchwarzKatze]

It is interesting to see that the majority of the voters has a 64 bits OS. And I still wonder why people are so furious that this game should stay 32 bits

My laptop is running 64 bits OS since 2004 (linux) and 2008 (windows)

The advantage of direct access to more than 4 gig memory would be huge.

I have seen the changes of my 25 years of IT

It's easy from the business perspective:
Customers whose machine can run 64 bit games: 80+%
Customers whose machine can run 32 bit games: 100%

Benefits from just compile the binary in 64 bit: more than 4gb of address space.
Does the game normally use more than 4gb of RAM?: From past experience, only after heavy modding or memory leak.

So without rewriting a lot of stuff to take advantage of 64 bit architecture, which is impossible for a game at this stage, the benefit is too small to justify losing customers.

 

[MajorHeartfire]

I am sure Paradox doesn't have anything against 64-bit - and let's be honest, they know more about programming than most of the people in this forum. That being said, the above poster is the most correct. Simply put, Paradox games don't use enough RAM to make 64-bit a necessity - and multi-thread is a different beast altogether.

 

[Oakfan]

A lot of people in this forum hate admitting when they are ignorant and like pretending they know what they are talking about no matter how painfully obvious it is that they do not.

A perfect example of this was a pair of posters on another thread who insisted that there was 'no such thing' as an 'historical fact'.

LOL really, wish I saw that thread. Need a good laugh now and then

 

[Atlantians]

LOL really, wish I saw that thread. Need a good laugh now and then

Here you go:

https://forum.paradoxplaza.com/foru...os-is-allying-everyone-randomly.895232/page-3

Scroll down halfway and find my comment that got 4 'disagrees'.

 

[Gwayne]

I do not know (nor pretend to know) actually how much impact 64 bits would have on the single player game. But I am sure that it will have a large impact on multi player games. Just because the network stack can be much bigger. And that is a big issue on 32bits operating systems.

64 bits has more impact than only on the 3.7 gig memory boundary

Also, if 80% of the customer base has 64 bits then it is ok to switch I think. Because the 20% left out will sooner or later upgrade. So that is also easy defendable from a business perspective if it creates a better game.

 

[KiwiNoob]

And if 20% = $2,000,000 in lost revenue is it still ok?

 

[SchwarzKatze]

I do not know (nor pretend to know) actually how much impact 64 bits would have on the single player game. But I am sure that it will have a large impact on multi player games. Just because the network stack can be much bigger. And that is a big issue on 32bits operating systems.

64 bits has more impact than only on the 3.7 gig memory boundary

Also, if 80% of the customer base has 64 bits then it is ok to switch I think. Because the 20% left out will sooner or later upgrade. So that is also easy defendable from a business perspective if it creates a better game.

Um, I just Googled what a "network stack" is, and it seems to be the layers of protocols encapsulating a packet, which I'm sure have nothing to do with CPU architecture.

And no it's not ok, it's losing more than one tenth sales for a nigh noticeable benefit.

 

[semaphore]

I do not know (nor pretend to know) actually how much impact 64 bits would have on the single player game. But I am sure that it will have a large impact on multi player games. Just because the network stack can be much bigger.

That doesn't really make any sense to me. Do you perhaps mean having to communicate over the internet needs a lot of RAM?

Something is going insanely wrong if handling network communications somehow requires much more RAM. Logically, there is a fairly low limit on what upload speeds a multiplayer game can reasonably expect to have. Even a few hundred kilobytes per second would be extremely demanding for most people's internet. Therefore, the temporary storage (i.e. RAM) needed for that data cannot possibly be of any significant size.

 

[Swinds]

Is anyone a software developer in this tread? Does anyone actually know the processes taken to make a game?

I ask because Podcat has actually answered on this and PDX decided on a 32bit program so that all 100% of people who want to buy and play HoI4 can. This means more sales and therefore more profitability.

I doubt PDX will make a 64 bit version of the HoI4 engine because it is too limiting on sales. It also means two versions of the same game to support, and I am a software support engineer and that is a really good way to double your bugs and development costs for no gain, no gain to PDX but extra cost mean less profitability and that leads to software companies closing.

At the end of the day the game looks great and this whole thread seems like a wast of time, but that is a forum and we all love arguing about silly things.

 

[wingo]

KiwiNoob, clearly you only see what you want to see. If you read my comment explaining what was running when, you can clearly see when running in the foreground like we normally play, EU4 alone (and nothing else) uses 2,5 cores - 1 whole core and the rest is split evenly among the 3 cores - which is expected, because there are a lot dependencies and synchronizing Clausewitz engine needs to do, so the other threads have to wait until everyone finishes their tasks - and only then they can start another round. You just cannot start any work on 31.12.1800 if all the work on 30.12.1800 did not finish yet.

And you keep bringing up background tasks when you can clearly see that when I paused EU4, CPU usage was 4% so what background tasks are you talking about?

Not to mention you keep ignoring all users and devs saying there is multithreaded support. So you keep repeating your lies despite all the evidence to the contrary:

oh my... this again...

Will paradox games be 64bit? Yes, probably when they build their next engine version. Before that, probably no.
Why? Because 64bits offer no advantage unless your program is designed to use more than 4GB of memory. Their engine is not designed for that. Usually paradox games use less than 1GB.

About multithreading:
They added multithreading with some of the HOI3 updates. It improved performance dramatically. Especially if your CPU had less single core power (HOI3 in my old laptop went from unplayable to quite fast in one update). Now the engine theoretically uses as many cores as you have but since it is extremely hard to divide the program flow to many threads, the advantage given by multiple processing cores gets smaller and smaller when you add cores. I have found that EU4 runs better with 4 cores compared to 2 cores but i noticed no difference in 8 core and 4 core systems.

For those who are not programmers:
you cannot just divide a program for several cores. You have to specify which things are done where and synchronize the process and mostly the cores end up waiting others to finish their tasks to be able to continue. Multithreading is extremely powerful tool with easily scalable tasks such as image or sound processing (or in any application where you can have a large series of unrelated operations) but much more difficult with games that have to follow certain synchronized logic.

My hardware is Quad-Core Core i7 920 with hyper-threading running at 2.9ghz. Chrome was running in the background with a few idle tabs, and Task Manager of course. Baseline CPU usage was 1-2% with no game running.

EU4, in my game at 1535.

- 1920x1080 windowed fullscreen/borderless mode
- I launched a coalition war with about 10 nations on either side, then set the game at speed 5.
- I left the UI fairly zoomed in, close enough for ships and armies to be moving about:

EU3, loaded the 1701 bookmark.
- 1920x1060 windowed (I dropped 20 pixels off the resolution so it didn't get clipped.)
- Played as France, putting me at war with 4 or 5 nations with 3 or 4 allies. Game speed 5.
- UI zoomed in again.

So, yeah. EU3-DW is not really "multi-core", as they say I think we can assume the maxed out CPU core is one thread for AI/game calculations, and then there's another main thread for the UI which shows as the 20ish % usage CPU. It might be that there's also a second active thread for the UI, showing as the 10%-ish core usage (polling for user input maybe?) Or that could just be Task Manager (seems a bit high for that, though.)

Note that coincidentally we see 25% total usage in both shots. Actually the EU3 usage was averaging more like 15%, I just happened to capture it at a higher spike. EU4 was very consistent around 25%.

It is still note-worthy that EU4 pretty much maxed out one CPU while others were no more than 50% (even taking into account the fact that there are only 4 real cores between these 8 graphs.) I suspect there's still not the ideal level of parallelism in its AI operations and general housekeeping operations. But clearly it's vastly better. And I expect EU4 is doing a lot more work than EU3 had to do overall.

And of course the key question is not "how does the game balance across cores", but simply "does the game run smoothly." And yes it definitely does on my system. So while there might be technically and optimally more balancing/parallelising to be done, it's not actually impacting the game; at least on my hardware. So any further work would quite likely be premature optimisation, and I think they've optimised it very well for the requirements.

Maybe EU5 will spawn one thread per AI nation Or a thread pool of 10-15 so that AI nations can be swapped in and out when they're at war or otherwise doing more work, giving them a dedicated thread when they need it and a shared thread when not. Hopefully by then the AI will be so sophisticated, it will warrant that!

There you go!

All our modern games try to utilize full parallelization of complex problems, unfortunate we cannot all ways take advantage of multiple cores/cpu:s

AI is multithreaded in most of our games

also this:

 

[KiwiNoob]

Actually read the post from TheBloke. Look at the pictures and realize they fly in the face of your theory that the CW engine is well multi-threaded.

 

[Teodosio7]

I run Ubuntu, what shall I reply to the poll?
Anyway, 64bit here. I haven't run a 32bit os in... ten years? Possibly more.

 

[Gwayne]

Um, I just Googled what a "network stack" is, and it seems to be the layers of protocols encapsulating a packet, which I'm sure have nothing to do with CPU architecture.

And no it's not ok, it's losing more than one tenth sales for a nigh noticeable benefit.

Its just that network on 32 bits architecture is limited to 1 Gb, in memory. So yes it has impact

And it's hard to just say that it's a barely noticeable effect if you do not develop the software. Everything we say he is just an assumption. If the game would be much better because of 64 bits it could attract more customers. So if you lose 10% because of no 32bits but gain 40% because of a great game it would still be a win for paradox. And that is how they will calculate. At the end of the day, they need to make money.

My assumption is based on my knowledge of high-end databases. And I know that one of our problems with 32 bits was network because the more connections you have the more memory it uses. And if you are hosting a game you will get a lot of connections and a lot more data. That is why I think that for muli-player games, especially with many users 64 bits will almost be mandatory.

Secondary, I play the game on a laptop with 32 Gb of memory what is probably not the standard. I did not run a 32bit OS in more than 10 years. I wanted long enough I think.

 

[semaphore]

My assumption is based on my knowledge of high-end databases. And I know that one of our problems with 32 bits was network because the more connections you have the more memory it uses. And if you are hosting a game you will get a lot of connections and a lot more data. That is why I think that for muli-player games, especially with many users 64 bits will almost be mandatory.

The requirements of a database is inherently inapplicable to a computer game game. Your 32-bit performance issues are because databases (Oracle?) wants to cache the data in memory to avoid unnecessary I/O when serving requests. It has little to do with the number of connections per se, and everything to do with the fact that a database wants to be able to perform queries on many gigabytes of data as fast as possible.

This game does not do any of that. It mainly just communicate and synchronise game states. It handles much less data than a typical business database to begin with, and it certainly doesn't need to service however many requests for data operations that a database is expected to perform.

Databases and games are simply incomparable.

 

[Gwayne]

The requirements of a database is inherently inapplicable to a computer game game. Your 32-bit performance issues are because databases (Oracle?) wants to cache the data in memory to avoid unnecessary I/O when serving requests. It has little to do with the number of connections per se, and everything to do with the fact that a database wants to be able to perform queries on many gigabytes of data as fast as possible.

This game does not do any of that. It mainly just communicate and synchronise game states. It handles much less data than a typical business database to begin with, and it certainly doesn't need to service however many requests for data operations that a database is expected to perform.

Databases and games are simply incomparable.

I do agree that the scale is different, but a game as this is also using a database architecture. And updating and syncing game states can become pretty large over time. Especially when you have many players. But as said before I do not know how paradox does this. A game-state can be huge if you put a lot of data in it.

Point is that 32 bits operating systems do not scale very well, and we are living in 2015. Every new machine has a 64bits OS and a minimum of 8GB.
So sooner or later paradox will have to change to 64 bits, just to take advantage of bigger capacity.
Bigger capacity is more players, more details, bigger AI, bigger map.

This discussion btw will always end up as a repeat because one group basically defends that there is no need for 64 bits (Those with 32 bits operating systems and no money or will for a upgrade) and the other group who will point out the advantages of 64bits (Those who already upgraded) (My perception)

Stardock with Galciv 3 already made that change because they wanted to be able to scale their game to 50+ players.

But a lot of games are now using 64 bits, like the Witcher 3, fallout 4 etc etc. In 2016 the majority will be 64 bits. Paradox will be aware of that.

So my assumption is that they are building now a new 64 bits engine that will be released with the next generation games. But not yet this one.

They have not much choice in this.

Edit: typo's and added last part

 

[semaphore]

I do agree that the scale is different, but a game as this is also using a database architecture. And updating and syncing game states can become pretty large over time.

It is so different that comparisons are frankly invalid. It's thousands of requests per second to examine maybe tens of gigabytes of data each, versus 10 people playing a game that is perhaps 2 gigabytes in total including graphics. The concerns are simply not the same.

Which is why, as I've explained, the fact that databases need more RAM to cache their data cannot be extrapolated into multiplayer games need more RAM.

 

[Swinds]

Its just that network on 32 bits architecture is limited to 1 Gb, in memory. So yes it has impact

And it's hard to just say that it's a barely noticeable effect if you do not develop the software. Everything we say he is just an assumption. If the game would be much better because of 64 bits it could attract more customers. So if you lose 10% because of no 32bits but gain 40% because of a great game it would still be a win for paradox. And that is how they will calculate. At the end of the day, they need to make money.

My assumption is based on my knowledge of high-end databases. And I know that one of our problems with 32 bits was network because the more connections you have the more memory it uses. And if you are hosting a game you will get a lot of connections and a lot more data. That is why I think that for muli-player games, especially with many users 64 bits will almost be mandatory.

Secondary, I play the game on a laptop with 32 Gb of memory what is probably not the standard. I did not run a 32bit OS in more than 10 years. I wanted long enough I think.

OK were do I start.

Crysis as an example. This game was developed to have the best possible graphics without too many compromises. How did it sell? It required better hardware than what was available to run at full speed. Low sales. Limited the people that could play it.

MOH:MW2 as an example. The grapics were good/OK but well within nearly all PC gamers rigs specs. How did it sell. By the bucket load. Virtually anyone could play it because it was a console games port.

Now compare the two companies fortunes....

Simple fact if you want to be successful do not limit the amount of customers that can buy your product.

The Game looks great and by all I have seen IMHO much better than HoI3. Why are people complaining about something that makes effectively no difference???? Do non of you play games on your phones. Because they are poor compared to PC games so therefore not fun or great??

How many packets of Data does the Game send via the NIC in a multiplayer. I suspect much much lower than your database. I would expect (I am not a programmer on the game) this is low because they have to take into account not everyone has super fast broadband. Battlefield sends relatively low amounts of data for that very reason. Databases are not Games and are not comparable.

 

[KiwiNoob]

Databases and games are simply incomparable.

Unless it's an MMO