*DAMN R6 Forum

*DAMN R6 Community => Gaming (All your Gaming needs are here!) => Topic started by: [:] Narauko on March 14, 2005, 06:18:58 am



Title: What makes every game diff. in performance?
Post by: [:] Narauko on March 14, 2005, 06:18:58 am
What is it that makes the diff. in the game. I don't think I have ever owned a computer that exceded the minimum requirements of a game, except for Duke 3d and UT maybe.

I ran Return to castle wolfenstein (30 minute load screens), Ghost Recon (Ran fine) and Medal of Honor (10 minute load screens) on a G3 350mhz 8mb vid 256 RAM. Minimum for Ghost Recon alone was something like G3 700 16mb. RtCW had requirements that my current G4 Laptop just meets.

On the G4 667* 256RAM 16mb VRAM Laptop I run Halo (fine), Ghost Recon (excellent).

My real question is what makes every game different in engine performance and why can't developers make games to manage all the data better. Games just seem to get more complicated and more hassle to run. But I shouldn't expect to have to buy a new computer every time a game comes out...should I?

nark?

p.s. Although I'm crazy, I'm not even going to attempt trying to run RvS.


Title: Re: What makes every game diff. in performance?
Post by: BTs_GhostSniper on March 14, 2005, 04:24:22 pm
Actually, the minimum system requirements for Ghost Recon is a 450MHz G3.  It played great when I had my 600MHz G3 iMac...and even better on my G5 (I never owned a G4).


Title: Re: What makes every game diff. in performance?
Post by: BTs_Mysterio on March 14, 2005, 05:16:48 pm
The fact on macs is we have a bad version of Open GL. Until we fix that the system requirements will be retardedly growing. On the real issue, You can't expect 16 MB of VRAM to cut it really. You may have been able to run halo on very low settings, but turn them up 1 notch and it will be trouble. The RAM you have and VRAM combined is just minimum you could expect to run RTCW or MOH.

If you actually did a model comparison (See original Unreal Engine 3.0 video released last year), you would say that the detail level is between a stick figure and some picture. You should be expected to buy a new computer if you are this upset about how games run. I've had this G5 for almost 2 years and it runs all games fine.

You're 667 powerbook is over 4 years old. That's 3 engines old. You need to realize these things need to require more and more. No matter what platform a 4 year old computer won't do the cutting edge things. Developers can only do so much for the management of data. Especially considering with 256 RAM you hardly have enough to keep the system and 1 more application running.

Currently I've made the following upgrades to my G5:
Combo Drive ----> 16X (DVD burn) Superdrive from Lacie
512MB RAM ----> 1.5 GB RAM
160 GB Storage ---> 320 GB Storage

Soon:
ATI 9600  ----> NEXT GENERATION TOP OF THE LINE
Thus why I am so against gaming on laptops. They can't over do the upgrades. I intend this computer to last over 4 years because with upgradability it can. Or maybe I'll go buy a quad G5, however, you answer your question: games are different because they get better.


Title: Re: What makes every game diff. in performance?
Post by: crypt on March 14, 2005, 09:25:54 pm
p.s. Although I'm crazy, I'm not even going to attempt trying to run RvS.

Good, you'll only be dissapointed that you can't run a great game. That's what happened to me, but I'm still patiently waiting for R6: Lockdown coming in June for all systems and PC.


Title: Re: What makes every game diff. in performance?
Post by: Cobra on March 14, 2005, 11:38:38 pm
Nark:
Computer applications in general perform differently because no two applications are performing exactly the same instructions and/or operating on exactly the same data.  Games, in particular, can have a huge range in the size of the data set it needs to keep track of (think TicTacToe vs. Doom3), what it needs to do with the data (update all geometry, test for and handle collisions, render visible geometry, apply post-rendering effects, etc...), and how it's going to work on chunks of data (organize the game world into quad-trees? cache static objects in VRAM?).  In addition to this, game performance is dependent on the hardware that is running it.

I guess the only answer is that no two games perform the same because they are not the same.  Why don't car engines perform the same?
 
Quote
and why can't developers make games to manage all the data better
They can, and good developers care very much about maximizing the performance of their game on their target hardware.  This requires smart decision-making about how to organize game data and how to make efficient use of the hardware they are targeting.  Sloppy developers (and developers who have run out of development time) might stop working on the program after they see that it can run on the hardware.

Especially considering with 256 RAM you hardly have enough to keep the system and 1 more application running.
Yeah.  Nark, I really wouldn't suggest running games on OS X with anything less than 512MB of RAM.  With 256MB, a lot of games are going to push you over the limit of your RAM, and OS X will resort to paging data to/from the hard disk, which will likely make your game completely unplayable.

The fact on macs is we have a bad version of Open GL.
What?  Microsoft officially dropped native OpenGL support for Windows years ago (in a push to make DirectX the standard in graphics rendering on that platform), I think at version 1.1.  OS X natively supports the most up-to-date version of OpenGL (1.5), since the OS is built on top of it and relies on it for many OS-related drawing tasks.  What do you mean by "we have a bad version of OpenGL?"

Quote
Until we fix that the system requirements will be retardedly growing.
Are the system requirements for current Mac games unacceptable?  Doom 3 requirements are virtually identical on OS X and Windows...


Title: Re: What makes every game diff. in performance?
Post by: *DAMN Mauti on March 15, 2005, 12:16:36 am
I recently discussed the topic optimizing games with Elandrion and beside the fact that it is very expensive to optimize a game is that you have hundreds if not thousands of different PC configurations. If you opitimize a game for a certain graphic card you exclude other user at the same time.

On consoles you have the contrary - developers can focus on 1 used hardware and get the max out of it by using all special graphic libraries they offer. Just look at the PS2 and what it can still offer graphic wise although it is as old as my iMac 400 G3(over 5 years).

My hope is that they finally start to use MP(multi-processor) systems like many Macs offer, but I guess even the parallel developed Close Combat won't have such a feature. AI is stressing the CPU very much - just compare UT2004 fly through and bot matches.

Good night,

Mauti


Title: Re: What makes every game diff. in performance?
Post by: BTs_Mysterio on March 15, 2005, 01:43:35 am
What?  Microsoft officially dropped native OpenGL support for Windows years ago (in a push to make DirectX the standard in graphics rendering on that platform), I think at version 1.1.  OS X natively supports the most up-to-date version of OpenGL (1.5), since the OS is built on top of it and relies on it for many OS-related drawing tasks.  What do you mean by "we have a bad version of OpenGL?"

The OS isn't built on the same type of Open GL. We have a shit load of laziness by graphics card driver makers. There is no evidence to the contrary. Perhaps you should do a comparison between Direct X on the PC relative to Open GL on mac. This is not Apples fault, although they should put more effort into helping with it, it is the fault of those who write the drivers for our cards to interface with the core of the computer.

Are the system requirements for current Mac games unacceptable?  Doom 3 requirements are virtually identical on OS X and Windows.

Not identical. You must recall the technological relativity of Mac vs PC. It has only been in recent rev that all the power series computers are available over 1.5 GHz which leaves a lot to be desired of the other computers. The fact that they cannot get the specs any lower says something about the myth, is it megahertz anymore?

My hope is that they finally start to use MP(multi-processor) systems like many Macs offer, but I guess even the parallel developed Close Combat won't have such a feature. AI is stressing the CPU very much - just compare UT2004 fly through and bot matches.

Multicore (probably MP too) support will be in Unreal Engine 3.


Title: Re: What makes every game diff. in performance?
Post by: Cobra on March 15, 2005, 03:22:20 am
Mauti:
My hope is that they finally start to use MP(multi-processor) systems like many Macs offer, but I guess even the parallel developed Close Combat won't have such a feature. AI is stressing the CPU very much - just compare UT2004 fly through and bot matches.
Yeah, at this point, the dual-processor setup that many Mac users have has gone largely ignored by game developers (which isn't too surprising, since most commercial games that are released for OS X are ports).  As the PC world begins to lean more and more in that direction, I expect that many more games will be natively written to take advantage of multiple processors before too long.

Myst:
The OS isn't built on the same type of Open GL.
It's the same OpenGL — Quartz links to the same OGL library and calls the same functions as a game would.

We have a shit load of laziness by graphics card driver makers. There is no evidence to the contrary. Perhaps you should do a comparison between Direct X on the PC relative to Open GL on mac. This is not Apples fault, although they should put more effort into helping with it, it is the fault of those who write the drivers for our cards to interface with the core of the computer.
Yeah, the current state of video card drivers on OS X is pretty dismal.  The topic of PC vs. Mac OpenGL performance comparison is an interesting one.  At low triangle counts, a Power Mac with a 6800 Ultra (up to the low 1000's range) is almost twice as fast as an equally equipped PC.  As the the triangle count increases, the PC catches up and eventually (~10,000 triangles) performs 2x as well as the Mac.  At high poly counts, the 6800 performance on the Mac is strangely low.  I guess this isn't completely unreasonable, since it's still a new-ish card, but it's really time now for Apple to solve the problem of video performance for its top-of-the-line GPUs.  Preliminary Doom3 benchmarks are seriously underwhelming, though it's hard to say how much of that is the fault of Aspyr/id/Apple (Aspyr says it's Apple's fault, but I'm not sure).  I've heard that Apple's been putting a lot of effort into improving their video drivers (Apple writes the drivers for the NVIDIA cards, I think ATI writes their own) for Tiger.  I hope they've had some success...


Title: Re: What makes every game diff. in performance?
Post by: BTs_Mysterio on March 15, 2005, 04:14:56 am
We should get a open source project for improving video drivers on mac.