Sign in with
Sign up | Sign in
Your question

What percentage of games will use 4 cores by 2009 -poll

Last response: in CPUs
Share

What percentage of games will use 4 cores by 2009?

Total: 103 votes

  • 5%-10%
  • 9 %
  • 10%-20%
  • 9 %
  • 20%-30%
  • 22 %
  • 30%-40%
  • 9 %
  • 40%-50%
  • 10 %
  • 50%-60%
  • 11 %
  • 60%-70%
  • 33 %
February 3, 2007 4:46:45 AM

Hi,

I am running a poll here about the percentage of games will use 4 cores by 2009, ur contributions are most welcome.
February 3, 2007 4:55:47 AM

Quote:
Hi,

I am running a poll here about the percentage of games will use 4 cores by 2009, ur contributions are most welcome.

That's only a couple of years away :idea:

I would assume most multiple core games will be of 2 cores.

There would have to be a lot of 4 core CPU's in service to make it feasible.
February 3, 2007 5:37:33 AM

I'd have voted but the vote maxxed out at 70%.
Its gonna be all of them.
Even in a year or so any game that hasn't taken advantage of multi-threaded programming will be left in the dust, because the difference between those that do, and those that don't, will be major. Major enough to mean 99% reviews over 66% reviews, and nobody buys 66% games.
Related resources
February 3, 2007 9:01:59 AM

NIZ:
What about casual games (think solitaire and company) that aren't very taxing on the system. They're games and people play those things :lol:  I get your point that most big titles will probably be multi-threaded and will benefit from multi-core, but games doesn't need fancy graphics or technologically advanced to be great games that get good reviews. Put the fun back in games, instead of just tech demos like D3 or FEAR. Great graphics is just but a small part in a good game.
February 3, 2007 9:35:30 AM

I am a bit sceptical. We have dualcores for more than 2 years and we have only few games optimised to benefit from a dualcore CPU. Jack said that soon there would be more engines optimised for QC. Until a quadcore at 2.4GHz can't beat a dualcore at 3.0GHz, I won't consider those engines as multicore aware.
February 3, 2007 9:39:52 AM

Vast majority of the gamer in the world will not have quad-core by 2009, so I don't think the programmers will take advantages of quad-core soon.
February 3, 2007 9:45:46 AM

Quote:
NIZ:
What about casual games (think solitaire and company) that aren't very taxing on the system. They're games and people play those things :lol:  I get your point that most big titles will probably be multi-threaded and will benefit from multi-core, but games doesn't need fancy graphics or technologically advanced to be great games that get good reviews. Put the fun back in games, instead of just tech demos like D3 or FEAR. Great graphics is just but a small part in a good game.


Ummm....You are aware that the cpu doesn't do graphics right? That's why we buy GRAPHICS CARDS. The cpu handles A.I., physics, sound processing(on onboard sound), it draws the wireframes, some particles, driver info, game logic, etc. The cpu controls how the game is played, the gpu just paints the picture. All the stuff cpu's do(A.I. and physics mostly) is what makes games fun. I can't wait until they start using the cpu more, it means more will be possible to do in games. Instead of the yearly refresh in gfx cards that just makes the same old game mechanic look prettier. I'm hoping they push quad and octo cores to the max by 2009-2010.
February 3, 2007 9:56:49 AM

In 2009 all major games developers will use multicore (quad+) engines, no question about it. I voted 60-70% since its highest grade, but IMO it will be 70+% :)  Less than 30% assigned to small game devs who will be working with single-dual cores and cannot afford to buy Tier1 engine and not enough inside recources to build decent quad+ engine on their own.
February 3, 2007 9:59:05 AM

You guys are forgetting two good points. 1 is that there hasn't been many multiple core systems on the market so the devs have had no reason to optimize for it. Now you have 360 and ps3 which are multi threaded systems. Devs can now cross platform threaded engines and make money as before it was a tiny minority. 2 is that there are three ways to thread an engine. Course grain, Fine grain, and a mix of the two. Course grain would equal setting certain tasks to a certain core(core 1 does physics, core 2 does A.I.). Fine grain would equal all the cores work together to handle all the given tasks(true threading). A mix of the two is exactly that(it dynamically switches back and forth between fine/course grain). The last option can decide not to use a core if it doesn't need it. If game engines use fine grain then it wouldn't matter if there was 96 cores it would still use all of them.
February 3, 2007 10:29:03 AM

Quote:
You guys are forgetting two good points. 1 is that there hasn't been many multiple core systems on the market so the devs have had no reason to optimize for it. Now you have 360 and ps3 which are multi threaded systems. Devs can now cross platform threaded engines and make money as before it was a tiny minority. 2 is that there are three ways to thread an engine. Course grain, Fine grain, and a mix of the two. Course grain would equal setting certain tasks to a certain core(core 1 does physics, core 2 does A.I.). Fine grain would equal all the cores work together to handle all the given tasks(true threading). A mix of the two is exactly that(it dynamically switches back and forth between fine/course grain). The last option can decide not to use a core if it doesn't need it. If game engines use fine grain then it wouldn't matter if there was 96 cores it would still use all of them.


a game using 96 cores, easy man. Haven't u read about AMD 's APUs approach and WHY they have taken this approach. For the foreseable future, we will be only using 4 cores( with Hyper-threading is that is to ever happen).
February 3, 2007 10:33:13 AM

All new A+ games that will ship in 2009 will have multicore support.
February 3, 2007 11:34:47 AM

Alan Wake can utilize all 4 cores
February 3, 2007 12:01:50 PM

Quote:
Hi,

I am running a poll here about the percentage of games will use 4 cores by 2009, ur contributions are most welcome.

Almost all games optimized to use 2 cores will be able to use 4 of them just as well; when you code a multithreaded app you usually make it use all the threads it finds, no matter if they are 2, 4, 8 or 16
February 3, 2007 12:21:13 PM

Quote:
Almost all games optimized to use 2 cores will be able to use 4 of them just as well; when you code a multithreaded app you usually make it use all the threads it finds, no matter if they are 2, 4, 8 or 16

A 2 threaded game will not use 4 threads; hence it will not use 4 cores. :wink: It is true that a 4 threaded game may use 4 cores, but it may also use 1 core.

I voted 30-40%.
February 3, 2007 1:20:50 PM

Threading tends to be dynamic today; it's just a few lines of code to see how many threads a system has and that's all. To give you an example, in the 3D modelling/rendering probram I use (Blender3D) you have the 'threads' button option; when you press it, the engine grabs all the available threads it finds to render as fast as it can,... The newest version, the 2.43, gives even the possibility to chose the threads number.
It's just stupid to code a program to use ONLY or at most 2 threads, because since the first dual cores were out, there were rumors of future quads or octo-cores,.. however, this is another problem; version policy. They release a version for 2 cores, then the "NEW" version for 4 cores etc,... blah.
February 3, 2007 1:27:04 PM

Quote:

a game using 96 cores, easy man. Haven't u read about AMD 's APUs approach and WHY they have taken this approach. For the foreseable future, we will be only using 4 cores( with Hyper-threading is that is to ever happen).


You are right.... Multicores may keep going for awhile for servers, but the deskop/notebook market with go the APU route.... Everything in the computer world that goes parallel always comes back to serial.... Whether is is pata -> sata, or number of cpu cores.... Multicore is a temperary hack, because they couldn't make the 10Ghz processors like they wanted to....
February 3, 2007 2:00:05 PM

Quote:
It's just stupid to code a program to use ONLY or at most 2 threads, because since the first dual cores were out, there were rumors of future quads or octo-cores,.. however, this is another problem; version policy. They release a version for 2 cores, then the "NEW" version for 4 cores etc,... blah.


Absolutely. Any good developers would NOT do that. The idea would be to dynamically assign work to threads. Just coding for 2 would be a waste of time imo.
February 3, 2007 2:14:32 PM

2009 is too far away. Of course we will have multi core games by that time.

I would say by H2 2008 everything should be running dual-core as mainstream.
February 3, 2007 2:35:39 PM

The current way things are going most new games will be Dual Core aware by the end of '07. So having 60-70% of games 4 core aware in 2009 is definetly feasible.
February 3, 2007 2:43:38 PM

Quote:
I am a bit sceptical. We have dualcores for more than 2 years and we have only few games optimised to benefit from a dualcore CPU. Jack said that soon there would be more engines optimised for QC. Until a quadcore at 2.4GHz can't beat a dualcore at 3.0GHz, I won't consider those engines as multicore aware.


I agree. Until all 4 cores are utilized specificaly for different tasks. Usage trading is what we have going on for the most part with software.

WAG; 95% of games utilizing dual-core by 09' and 60%+ using quad-core.

Quote:
You are right.... Multicores may keep going for awhile for servers, but the deskop/notebook market with go the APU route.... Everything in the computer world that goes parallel always comes back to serial.... Whether is is pata -> sata, or number of cpu cores.... Multicore is a temperary hack, because they couldn't make the 10Ghz processors like they wanted to....


Intels road map calls for almost 16 and eventualy 32 cores into the next decade. Stability and efficiency; sleeping cores - performance on demand. Burning a core out, it had been said that the CPU will continue functioning on the remaining. Multi could be around for a very long time. Although, progression and physics meet head-on at some point, the micron architecture will change inevitabily. How long though? Intels K structure is the biggest in transistors and chemicals in 30 years. 8O
February 3, 2007 2:45:22 PM

Open task manager. You will see that your computer may only use around 40 processes but over 400 threads. In theory each thread could be sent to a different processor, but in a timesharing enviroment one processor can do the work of many with little impact on the user.

Games can be multi-threaded fairly easly. The question is will they realy benifit from the extra cores. My guess is that the highend game engiens will be writen to take advantage of multiple cores and the smart ones will scale to numbers beyond what ever the current market is. Then any game using those engines will also benifit.
a b à CPUs
February 3, 2007 2:48:11 PM

When you say "by 2009" does that mean the beginning or the end of the year? Whether you are talking 2 or 3 years away makes a huge difference.
The move to dual core has been slow, but I think the move from dual to multi core will be a little faster - once developers make the first change the second becomes much easier.
I think most major games will become dual core in 2007 and 2008. Few will be multi core at the start of 2009, but a fair number (30% - 60%?) will become multi-core during 2009.
February 3, 2007 2:51:43 PM

Where is 0% of games?

its very difficult to create a game that uses 4 cores natively. Id say about year 2018...
The problem is with the humans and not the computers this time ;) .

I dont consider "support" the same thing as using.
February 3, 2007 2:53:48 PM

you are so negative! :lol: 
February 3, 2007 2:55:54 PM

Quote:
you are so negative! :lol: 


Nothing negative or nothing to break your hopes... its just extremly hard to create a game that uses 4 cores. Programmers have not reached that level yet but once we do we will see amazing things.
February 3, 2007 3:26:46 PM

well companies are already doing it, what makes you think they wont have their applications ready by 2009? I mean you say 0% but I would say there would be a big % by that time of applications that use 4 cores
February 3, 2007 3:39:36 PM

Quote:
well companies are already doing it, what makes you think they wont have their applications ready by 2009? I mean you say 0% but I would say there would be a big % by that time of applications that use 4 cores


There is no program to date that is native multicore. It. is not possible for now but when it does happen it will bring in a new age.
Being optimized for multi core isnt the same thing. Humans cant program at that level yet.
February 3, 2007 4:09:23 PM

Who gives a rats ass. Go get laid dude.
February 3, 2007 4:14:42 PM

Quote:
Where is 0% of games?

its very difficult to create a game that uses 4 cores natively. Id say about year 2018...
The problem is with the humans and not the computers this time ;) .

I dont consider "support" the same thing as using.


Well, I partially agree. It is definitely more difficult to write and especially test a program that starts new threads and/or processes. On the other hand, it's not really that bad. The most difficult parts are already covered by APIs like Microsoft.Net for example, so the game developer doesn't need to deal with all the complexity himself and reinvent the wheel. I'm sure that in the next year or two developers will have much better support for creating multithreaded apps. Besides, even if it's difficult, so what? Get one or two really smart developers on the team, pay them a lot, let them deal with this part. Not all the developers on a big team need to be good at this area.
February 3, 2007 4:48:53 PM

Quote:
Where is 0% of games?

its very difficult to create a game that uses 4 cores natively. Id say about year 2018...
The problem is with the humans and not the computers this time ;) .

I dont consider "support" the same thing as using.


no it isn't difficult, unless you don't know about multi-threaded programming. You don't write code explicitly for a certain number of cores, you just write multiuhtreaded code and let the scheduler worry about where it puts them.
February 3, 2007 5:02:12 PM

Quote:
well companies are already doing it, what makes you think they wont have their applications ready by 2009? I mean you say 0% but I would say there would be a big % by that time of applications that use 4 cores
it all depends on how you would classify a multi core capable or truly multi core game.

in essence we've had multitsking games with software that splits off into multiple io devices already, the soundcard and graphics card along with the processor for instance..... and now Ageia with their stupid overpriced physics card.... (ok my rants done in regard to Igeia)

Valve's recent statements along with John Carmacks made mention of this practice along with 2 other ways......

I suspect the authors question in this thread was in regard to a truly simultanious multi processor enabled application and in this case I would also vote for 0% by 2009..... hasn't anyone noticed it takes 4-6 years to develp a game... Unreal 2007 which is a relatively simple game in that it only requires a graphics and physics engine is delayed and has been 4 years in development.... Crysis is delayed yet again and likely will be again possibly making it out by Xmas at the earliest.... so do I forsee any game making it to multi core in the next 2 years... no.

multi-threaded software development didn't start recently it's been an ongoing development for over 30 years and no one has successfully been able to take advantage of it ever save for in repetitave apps not related to video games....

Sony took a gamble this year with PS3 and hopes that fully multi core games will be around in the next 5-10 years... the plant being games will continually get better because the software will be able to take better advantage of the hardware........that is their gamble and they are being blasted by every software developer in the world bar none..... ignoring that modern games cost as much as 20 million to develop right now..... which is outright killing the industry, ignoring the statements by every software developer in the world that has already said optimising for dual core.... not quad core but Dual core is going to be an outright nightmare and that we would be lucky to see any substantive benefit in the next 5 years.... ignoring that according to every software developer learning to multi thread will not be a secret yet unlocked but that it will be a terrible pain in regard to every program developed each time.... as in you'll have to start over every time you develop a new program.... ignoring the statement from Nvidia's own engineers that stated clearly that no school in the world currently has an adequate curriculum to teach programmers to multi thread, that no software tools are available to develop multi threading in the fashion desired by the author of this thead..... the timeframe seems quite short.

I would likely say that multi core supporting software will be ad-hock for the next 5 years showing little to no benefit that couldn't be attained by a faster CPU with add in peripherals cards and that software that truly starts to take advantage IE: possibly a 50% improvement in performance will likely not arrive for another 10 years if ever we attain that level of improvement.

p.s. I have seen a few companies looking for cash to develop a hardware development solution that would analyse software while it runs and automatically optimise the software to better tailor it to multi core..... this seems feasible but how much to develop and how long it will take is another matter entirely.
February 3, 2007 5:27:05 PM

Quote:
Who gives a rats ass. Go get laid dude.


done!

I stil think it will be close to 2018 before we see somethign good.
February 3, 2007 5:41:02 PM

Quote:
NIZ:
Put the fun back in games, instead of just tech demos like D3 or FEAR. Great graphics is just but a small part in a good game.


Just a slight aside, but I cant believe you consider FEAR a "tech demo". It was a phenomenal FPS, which used movie style special effects such as ghosts/vanishing people to really set the mood of a horror movie that just happened to be a game you could play through. As well as the squads who communicated and (with varying degrees of success) tried to work as a team to pin/flank/kill you. If a multicore engine means they can merge and improve upon the mood/setting/AI of FEAR, the polish and physics of HL2, and the interactivity of something like Dark Messiah of Might and Magic (where your position in a fight is vital compared to your enemies so you must think) then I'll buy a quad core CPU tomorrow just so I can enjoy the new depths of entertainment.

And if you thought Doom 3 was just a tech demo, oh my god. I totally agree with you. After HL2/FEAR I was so disapointed.
February 3, 2007 6:54:30 PM

Quote:
Where is 0% of games?

its very difficult to create a game that uses 4 cores natively. Id say about year 2018...
The problem is with the humans and not the computers this time ;) .

I dont consider "support" the same thing as using.


Are you saying the source engine update that's coming doesn't support quad natively? http://techreport.com/etc/2006q4/source-multicore/index... read up.

What about Alan Wake? http://www.anandtech.com/tradeshows/showdoc.aspx?i=2841... they basically said this game will not run on single core

What about crysis?
http://www.gamespot.com/pc/action/crysis/news.html?sid=... (towards the bottom of the interview he states that it will support single/dual/quad/multi(which I assume means more than 4))

Those are all due in 2007. Your 0% is incorrect. There are sure to be surprises throughout the year also.

They are not just supporting multi core they are utilizing and improving performance.
February 4, 2007 5:59:21 AM

Quote:
Just a slight aside, but I cant believe you consider FEAR a "tech demo". It was a phenomenal FPS, which used movie style special effects such as ghosts/vanishing people to really set the mood of a horror movie that just happened to be a game you could play through. As well as the squads who communicated and (with varying degrees of success) tried to work as a team to pin/flank/kill you.
~snip~

I did like the looks and the AI, but rest just didn't quite do it for me. Perhaps I was put off by the cliche story, but I didn't really enjoy the game as most other people did.

Quote:
Ummm....You are aware that the cpu doesn't do graphics right? That's why we buy GRAPHICS CARDS. The cpu handles A.I., physics, sound processing(on onboard sound), it draws the wireframes, some particles, driver info, game logic, etc. The cpu controls how the game is played, the gpu just paints the picture. All the stuff cpu's do(A.I. and physics mostly) is what makes games fun. I can't wait until they start using the cpu more, it means more will be possible to do in games. Instead of the yearly refresh in gfx cards that just makes the same old game mechanic look prettier. I'm hoping they push quad and octo cores to the max by 2009-2010.

My point was games are putting too much emphasis on technological goodness and spit shine graphics instead of putting it where it counts, like story and solid game mechanics. Physics and AI are all great things to help you immerse in the game, but what if the other elements sucked? Then no amount of AI and physics will make it fun. I guess how I worded it could be mistaken for I thought CPU did the graphics. Apologies for the confusion.
February 4, 2007 1:53:50 PM

See Gothic 3. Yuck.
February 4, 2007 2:25:25 PM

If you count all games, I'd say less than 5%. Remember, you have to account for what actually consists a large % of PC gaming... card games/etc.



Now, if you count only 3d games, I'd say closer to the 5-10% mark. Maybe as high as 20%.

Thats of course counting the games that will use the 4 cores efficiently. I'm sure the games that will get SOME benefit from 4 cores will be closer to 50%.
February 4, 2007 4:57:33 PM

Bringing card games into the mix is a little like saying that the number of athletes that use steroids is very small when you consider most of the people playing sports are little kids.

Anytime a program calls a new function it can easily be made into a new thread. Then all threads compete for an open core. The only real problem is that many aspects of a game need to happen in an order and for thread based programing to gain any benift things have to be able to be calculated in an asynchronous maner. There are aspects of games that can benifit from this. When you only have 1 core the added overhead of multi-threading can be hard to justify.

On the most basic level by 2009 every game will use every avalible core, just as every other thread will. The number of games that can tax four or more cores to the point you can't run a scan in the background will be very small if any.
February 5, 2007 3:11:23 PM

Quote:
Bringing card games into the mix is a little like saying that the number of athletes that use steroids is very small when you consider most of the people playing sports are little kids.

Anytime a program calls a new function it can easily be made into a new thread. Then all threads compete for an open core. The only real problem is that many aspects of a game need to happen in an order and for thread based programing to gain any benift things have to be able to be calculated in an asynchronous maner. There are aspects of games that can benifit from this. When you only have 1 core the added overhead of multi-threading can be hard to justify.

On the most basic level by 2009 every game will use every avalible core, just as every other thread will. The number of games that can tax four or more cores to the point you can't run a scan in the background will be very small if any.


Your analogy is correct in that the number of athletes that use steroids is extremely small because you asked the wrong qustion. The correct question would have been what percentage of PROFESSIONAL athletes use steriods.

The question is what percentage of games will use 4 cores by 2009. That includes massive mutiplayer online, First person shooters, real time strategy, turn based strategy, card games, sudoku, chess, pinball, kid games such math blaster, carmen sandiego, etc.. If you believe the question should be what percentage of first person shooters or MMO's will use 4 cores by 2009 then start that thread. The OP topic as written includes everything.
February 5, 2007 9:02:14 PM

Mainstream games are clearly trending to multicore. 2009 is far enough out that we should see the majority of mainstream games able to use 4 cores or more by then.
!