Guild Wars 2 & Linux/Mac & SMP

Even tho its far away today its something i would like to discuss
Guild Wars 2, Support for multiple cores/cpus?? Support for Linux/Mac

As most PC's today are Dual or Quad core it would only seem natural for Anet to make it SMP capable, a feature i think was folishly left out of GW.
I know i would sure like to see it in GW2.

And more and more ppl today are using an alternative OS.
Running Linux, or using a Mac.
Myself im a 50/50 user, half Linux half Windows
Most of us Linux users know we CAN play GW and other games without any bigger problems.
But if im not mistaken the EULA says that u are not allowed to.
Imo taking the step and making a client for Linux/Mac users would be greatly appriciated and loved by alot of fans.
And if not making a client for it, atleast make it allowed but maybe not supported.

Just my 2 cents, ur oppinion?

Id like to see some mac compatibility

Well, it would have to be worth their while. Much of what you're proposing is expensive to implement. The number of people out there wanting a Linux client is so small it just doesn't seem cost-effective for them to bother with it. Trust me. I'm a software developer, and the amount of QA, testing, and additional support that it would require would not come cheap. As far as not being allowed in the EULA...I'm sure that has something to do with the possibility of a whole new realm of hacking. An open source OS could potentially allow for a whole new world of trickery...

As far as a Mac client...well...they'd have to decide if there was enough of a demand to make it worth the cost as well. Most Mac users I know aren't gamers.

As far as SMP being "foolishly" left out of GW...meh. It's not really needed. Most single-core CPUs from the past 3 years have plenty of power to run GW just fine. Why spend time threading your existing code (which can be a daunting task depending on the app architecture) when you'll get minimal benefit from it? Since I've not seen all of the features that will be in GW2, I can't comment on the benefit of SMP in it. If there will be a significant amount of physics and complex AI then it would be a nice feature.

I can see SMP possibly with GW2.
Mac/Linux.. not sure. Right now it seems very DX heavy. They would have to make it compatibility with OpenGL or make there own engine / code.

Agree with Dex about SMP and GW1 it easier to count the number of games that support it vs the ones that don't.

The *next* version of windows is going to force / or push SMP/64bit support on the software world. But in truth I can see 64 bit supported before SMP... in mass in the industry. Not only that GW started development 5-6 years ago and SMP when rare then.

Most mac users I know play WoW, because that's really all they have. I'd like to at least see a mac client for GW2, so that I may have more ammunition when trying to convert them.

I know it takes a bit of manpower and so on if they are gonna make a linux client to, BUT imo not that much that it wouldnt be worth it.
I would go as far as to say that most of the ppl that do play games on linux play some kind of rpg.
NCsoft is a very big company and its not like GW has been a bad deal for them, so i cant see why it would make it a "worth while" question.
Sure it may not be the biggest company out there, but looking at the market today and looking at older games, i cant agree with it being a daunting task for either SMP or Linux client.

A good example of games that are far older then GW and that have clients for linux/mac/smp supplied by the developer is Unreal Tournament 2004 & Quake 3.
And just saying the game doesnt need more then 1 core isnt justified as a reason imo, and i do belive that GW2 will have alot more features supported, just like DX10
The whole SMP part applys to all game developers today imo, and all games that are rather high end , and GW is imo high end, even tho it doesnt require a very powerful system, if u play the game maxed out at realy high resolution it does require alot of performance and it does look realy good

And it will always be a fact that some ppl cant afford to upgrade their system to the absolute latest, so ppl using systems like P4HT whenever GW2 arrives will most likely gain some benefit from SMP support

I do understand ur PoV Dex, as i code alot of PHP and small apps
/endrant

NCSoft is huge... but Anet is the company / developer is the one that programs the game - from the photo of the entire company from the fansite gift baskets... seem to be < 100 people. All NCSoft does is act as the publisher, arranges manufacture / distribution, etc.

Besides if they can implement DX9, DS3D and full blown EAX after it's release I think they can add that given time.

In the big picture that's still a microscopic demographic.

Converting any complex app from a single-threaded app to multi-threaded is a daunting task. I've overseen a team that's had to do it. Trust me, it's nothing at all like programming a bit of PHP. Making such a conversion with something like a complex game app would be outrageously expensive. It's not just the programming. The QA and testing process is VERY expensive, not to mention all the new support issues and bugs that would certainly arise from such a large re-work of the code. No offense, but I don't think you understand how much money we're talking about here. ArenaNet is simply not that large or well-established a company to be throwing money around like that if it's not necessary.

These games were designed from the start to use SMP, and with Quake 3 only the server components of them truly get any benefit from it. The client aspects of Quake 3 are largely single-threaded.

It's perfectly justified. Why would you put a race car engine in a car that you're only going to be driving around the city? Threading can actually decrease the performance of software running on a single-CPU machine. As I said, if GW2 has things that make sense to offload onto another thread (like physics and AI), then SMP would be a welcome addition. DX10 doesn't have anything to do with it.

It's not that simple. SMP doesn't always net a worthwhile performance increase. Saying, "this game is pretty...it should be multi-threaded" doesn't make any sense. For GW multi-threading is overkill, plain and simple. There's no benefit at all. Seriously, you'd see no performance increase. Why would ANet bother?

Hyperthreading is a joke. There are so few situation in which it adds performance that Intel doesn't even talk about it anymore. In many cases it actually degrades performance. Turn it off.

Yeah, but that was a matter of adding some hooks into a few APIs. Threading a large existing single-threaded application would require you to disassemble your source, plan a threading strategy, do unit testing, do performance tuning, put it back together, and re-test your entire app to start tracking down all of the things you broke in the process. It's a big deal.

i think u missunderstood alot of it
seems like u thought i wanted them to make gw1 smp capable now
that is not it, i wanted to see the support in gw2
and just like u said, q4/ut2k4 was designed that way from the start
do the same for gw2
biggest part of the market today is dualcore/quadcore
there is no real single cpu/core systems realy being sold anymore
and before u say " there is single cpu/core systems" sure they are being sold
but looking at the bigger part of the market, atleast in gaming
its dual/quad core that dominates
i do realize in alot of systems the gfx is the bottleneck
but looking at new games and gw2 in the future the development goes fast
dx10 and what not, i seriously doubt any new high performance games will be single cpu/core , if u have a system with lets say 4 possible cores, why would u not make the game capable of using them all

Ok, fair enough. It did sound like you were talking about SMP with GW (1). When you said that SMP was "foolishly" left out of GW it made me think that. GW (1) would not benefit at all from SMP, and it was not "foolish" to leave GW single-threaded. It wouldn't have made sense to have made the client multi-threaded because any dual-core CPU already has enough power in a single core to run GW at max settings. Threading would do nothing for it.

It only makes sense to thread a game when it can actually benefit from using more than one CPU. If the game can get all it needs from a single CPU it's far more efficient to leave it single-threaded and let the OS balance load between the CPUs among the game and other running applications (e.x., the game runs on core 0 while background apps run on core 1, etc.) Just because most newer PCs have multiple core does not mean that every application needs to be multi-threaded. Again, if the game can run at full speed using a single CPU core, attempting to split threads between multiple cores is going to result in a NET LOSS of efficiency.

Again, I can't comment on whether or not I think GW2 should be threaded because I haven't seen any real specs on it. If a single core on...say...an A64 X2 3800+ can't handle the game at near-max settings then I'm all for it. Otherwise it just doesn't make sense.

I would love Mac support. Its easier than ever with Intel chips, and with directX 10 compatable cards going into Macs im sure they could do it. Please!

Hardware doesnt account as much as the OS its being made for, they would have to make it specifically for Mac OS X regardless if they have Intel or G based processors,

but compatibility with linux (aside from WINE) would be nice I really dont want to use windoz anymore....

Exactly. What they actually have to program for is Win32 and DirectX. It has very little to do with the actual hardware. Most of the base code that the game itself runs on would have to be re-written as well as the layer that interfaces with the graphics and sound APIs. No small task for such a small financial return.

There's plenty of cross-platform APIs like SDL (windowing, input, sound, basic graphics, networking), OpenGL (3d graphics), OpenAL (3d sound) etc. A game built on these could run on Windows 2000, XP, Vista, Linux, Mac OS X and possibly other operating systems with minimal effort.

If the userbases of non-Windows operating systems were larger and more interested in gaming it might actually be worth bothering with cross-platform development, for now I think the devs will just stick with DirectX they're already familiar with.

Well with the move by EA that all their games that will be released and BF2142, C&C 3, and more that is going to the Mac i wouldnt be surprised to see many more developers moving to the Mac side too...

I dual boot linux and windows, but the only reason windows is still on my computer is for gaming. I don't plan on including windows in the next computer I get so I'd love it if GW2 came out for linux as well.

I have a feeling that ArenaNet would probably not use any sort of wrapper like SDL. I'd guess they already have an abstraction layer in their engine. Case in point: the graphics can switch between dx8 and dx9, and the sound engine can use DirectSound or FMOD. It's not such a huge stretch to imagine them also creating an OpenGL rendering library similarly - which would go a long ways to porting to other platforms.

I'd guess they'd want to write their own native audio, graphics, networking, file, and system components to interface with native code. Talking about a cross-platform library is sort of a non-starter, because in order to reap the benefits of it, you have to start with that library.

I remain hopeful that they'll eventually go cross-platform to mac and linux, but they haven't given any indication that they plan to do so yet.

Well,

if they would move to OpenGL instead of DirectX, that would make it a lot easier.
Secondly, the argument that there is no userbase for Linux/Mac client is not valid.
The reason is simple, if there is noone developing for these platforms, how can a user base be created ?
Further, as the userbase for Linux/Mac globally is increasing, a responsible company in my view should have clients for the platforms too.

Americas Army did it, well, sort of at least, they made a windows client and then had some highly skilled guy write the Linux and Mac client software for them, he basically translated it. And trust me, he did it extremely well. His name currently slips my mind.

So it's not a matter of not being able, it's a matter of willingness.
There is politics too it, I fully know.

But the argument about costs and all is just not a valid one. It can be done for low costs, if they systems engine and all around it is set and designed for all 3 platforms, being Microsoft, Linux and Mac.

So Anet, what is your idea so far, as there is plenty of time, why not listen to the users on this one as well.

/Santa

That may be true, but not everyone has a new dual/quad core computer. There is a huge installed base of single core computers out there. Not too many people are likely to buy a new dual/quad core computer just to play GW (or GW2). (Btw, many parents buy a Mac because their kids can't play so many games on it.)

I think many of you are missing one important point in all this. Many of you talk about writing a Linux/Mac client as if the said client was static and that's all you had to do. But you are forgetting that GW (and I assume, GW2) are written to be dynamically updated/patched - not a fixed client like say, a game cartridge (which are often developed cross-platform). As we all know, new file downloads are a common event when you log into GW. The extra work load of updating and patching clients for Mac and Linux wouldn't be worth it for the small user base. (And don't bother saying that Mac and Linux have fewer bugs than Windows - it's the game that would be patched, not the OS.)

Perhaps, if GW2 was written in some cross-platform dev system, these problems could be minimized, but that doesn't help GW(1), and you'd still have bugs and glitches that would only involve interactions with a particular OS.

Does anyone know the amount of WoW users that play on the Mac? Im sure its about 1 million. A lot of students use Macs, and I'm sure a lot of them play WoW.

I understand development of a LINUX client would not be cost effective & a MAC client is a maybe. How cost effective would it be to make sure some of the WINDOWS emulators available would be compatible with Guild Wars 2?

The advantage gained, especially for Linux users, would be a wider market. Linux does not need constant hardware upgrades to maintain performance. People with older computers would have less problems running a demanding program like a high end game. Those people would be more likely to buy more games software if if they did not have to buy hardware upgrades so often.

It also won't be necessary after a deal was struck between Novell and Microsoft last month enabling future distro's to use the Windows AutoUpdate and Direct X architecture and enabling nearly full support for windows games on the Linux distro's produced by Novell (and the Linspire guys, forgot the name)

and this deal cost 23billion $'s... and here was me thinking that Linux didn't cost anything?!?

heh, <3 my vista anyway, so i don't care

Yes, Linux implementations do tend to be more resource efficient than even a well-maintained Windows install, but not enough to make an enormous difference when it comes to gaming. Windows doesn't need "constant hardware upgrades to maintain performance" either, just a little system maintenance. On top of that, most systems are still GPU limited in most games, so better resource management won't make more than a small difference in most cases. This difference gets even smaller as games become more complex. A P3 with a FX5900 isn't going to run Oblivion at an acceptable level...I don't care what OS you're using. You're going to need more powerful hardware even if you're using Linux (assuming you can even get really good Linux drivers for your hardware...).

As far as a wider market...I still really don't think that's significant enough to devote dev resources to it. Windows Vista, which is considered a total failure as far as sales compared to past Windows releases, surpassed the MacOS's ENTIRE WORLDWIDE INSTALLED BASE in the first 5 weeks after its release. MacOS has 3x the installed base of Linux. Even if you could convince a handful of people to install/dual-boot Linux to squeeze more performance out of their old hardware it wouldn't make enough of a difference in sales for ANet to see a significant change to their bottom-line.

Don't get me wrong, I like Linux. I use it to host services all the time. I just wouldn't bother porting games to it if I ran a game dev shop.

Just my 2 cents...

linux and mac makes up a total of about 4% of the computers world-wide, in south korea im pretty sure that only windows computers are compatible with anything.....out of that 4%, the few people who would play gw2 (compared to the whole gw2 community) could learn how to dual boot...not worth it for a. net

even if 1million american users wanted to play guild wars on a mac...a.net will just say to dual boot because that takes man power out of things they could be doing to support the rest of the large gw community (consider europe, and asia where macs are a lot less supported and a lot less popular)