How Does 3D Mac Gaming Keep Up?

[Wiglaf]

So they're porting Return to Wolf, they've already done UT and Deus Ex...my question is, how do they run on macs at all?

Most mac video cards are ATI Rage 128s, and (correct me if I'm wrong) Geforce2s and Radeons aren't listed in any requirements from any mac porting company (and aren't in wide use either). Most PC games out there, especially FPS games, NEED a Geforce or Radeon card to run well. How is it that something like Quake III Mac will run fine, at full detail, on something like a Rage Pro (8mb SDRAM) with a 400 MHz G3 processor? A PC couldn't dream of doing that, I would think (after looking at those minimum specs for RTCW especially).

I don't get it.

[monolith94]

Inside every macintosh there resides a small green rabbit foot on a chain, attatched to the motherboard.

Been like that ever since they got rid of the clones; they were able to standardize, make 'em all uniform.

[Wiglaf]

Do a find for "Apple System Profiler" and tell me the VRAM total, would you? Odds are it's not above 16.

Isn't that a little interesting? Only the new iMacs have started to incorporate the 32mb MX card, as opposed to the Rage, but even that's a bit behind the PC market of Geforce3s and 8500s. How do games like RTCW sell or run well at all on our side of the court with those limited resources?

[HugoHillbilly]

I am a PC user so I wouldnt know. But one guess is that maybe the Mac OS is more efficient than Windows at running stuff.

And dont ask how I got here, I dont know myself. I clicked a wrong link I think.

[Wiglaf]

I am a PC user so I wouldnt know. But one guess is that maybe the Mac OS is more efficient than Windows at running stuff.

I'd think it's more of a hardware issue than a software issue, actually. What operating system you have probably wouldn't make the biggest impact (more demanding ones like XP probably slow things down though). I'd guess it's just more efficient coding in games and lower resolution/detail, but somehow that just doesn't click either.

And dont ask how I got here, I dont know myself. I clicked a wrong link I think.

By all means, stick around. It's great to see new faces around here.

[wgabrie]

If he doesn't know how he go here he may not know how to come back, but I hope he does.

[Asher]

Originally posted by Wiglaf
Isn't that a little interesting? Only the new iMacs have started to incorporate the 32mb MX card, as opposed to the Rage, but even that's a bit behind the PC market of Geforce3s and 8500s. How do games like RTCW sell or run well at all on our side of the court with those limited resources?

MacOS 9 was actually pretty bad for 3D gaming. MacOS X is supposed to be great, if the developers start taking advantage of some of the tools it has.

I guess your whole perception of "runs well" is in the eye of the beholder. John Carmack (the guy who made the engine RTCW runs on) says the x86 hardware is faster for 3D gaming, mostly because his engine is very memory-bandwidth hungry. Something current Macs don't really have.

And the 32MB GF2MX is more than a "bit" behind the GeForce 3s and 8500s. It's actually about 1/4 to at most 1/3 of the speed of the GF3s/8500s. But you can get those for Macs too.

I think the thing here, Wiglaf, is you're making something out of nothing. I bet you an old 500MHz G3 + Rage 128 performs just as well as a P3 800 + Rage 128. And I guarantee you a 2.2GHz P4 with a GeForce 3 Ti500 would waste even a dual 866 G4+ with a GeForce 3 Ti500.

I'm patiently waiting for those G5s though. If the benchmarks I've seen are real, that thing will be a real screamer.

[Wiglaf]

Oh Christ, look who's here.

MacOS 9 was actually pretty bad for 3D gaming. MacOS X is supposed to be great, if the developers start taking advantage of some of the tools it has.

How was 9 "pretty bad" for 3D gaming? Actually it posted very high fps in QuakeIII with only 8mb VRAM to deal with, in my experiance...(powerbook)

I think the thing here, Wiglaf, is you're making something out of nothing. I bet you an old 500MHz G3 + Rage 128 performs just as well as a P3 800 + Rage 128. And I guarantee you a 2.2GHz P4 with a GeForce 3 Ti500 would waste even a dual 866 G4+ with a GeForce 3 Ti500.

**yaawwwwnn**

I've said it before and I'll say it again. Intel's method of measuring clockspeed is ridiculous now, if you honestly believe they made the 2GHz-3GHz jump overnight with the Northwood. And it took a card with 32mb VRAM and a 700 or so processor on the PC side to post similar results to the powerbook I tested, actually (assuming similar resolution and detail settings). That's what I'm so screwed up about atm...how a so called "inferior OS" with a mobility 128 can equal a PIII700 and a Geforce2 during normal gameplay.

I guess your whole perception of "runs well" is in the eye of the beholder. John Carmack (the guy who made the engine RTCW runs on) says the x86 hardware is faster for 3D gaming, mostly because his engine is very memory-bandwidth hungry. Something current Macs don't really have

Ok, but we're still seeing RTCW now running very well, according to apple, on a Geforce2MX equipped iMac with 32mb VRAM and 256SDRAM. A PC, for optimal performance, would take up a bit more than a budget card and a "less than hardcore" RAM total.

And I bet you'll see people pulling it off with excellent fps using only 128RAM and a Rage Ultra (16mb). It happened with Elite Force and Deus Ex after all.

[Asher]

Originally posted by Wiglaf
Oh Christ, look who's here.

Haven't you always been here?

How was 9 "pretty bad" for 3D gaming? Actually it posted very high fps in QuakeIII with only 8mb VRAM to deal with, in my experiance...(powerbook)

Essentially it had inefficient code. A similar clocked P3 outpaced a G3 running Windows 2000 vs OS 9 using the same video card, which should not happen. G3s should definitely be faster clock for clock, the OS' inefficiency (or perhaps lack of good drivers?) slowed down performance.

I've said it before and I'll say it again. Intel's method of measuring clockspeed is ridiculous now, if you honestly believe they made the 2GHz-3GHz jump overnight with the Northwood.

I've said this before, and I'll continue to say it again until you seem to understand. The method of measuring clockspeed has not changed, last I checked it still worked by amount of cycles per second (Hertz or Hz). That said, I know what you're trying to say. And what you're saying is irrelevant, I don't really care what the clockspeed is on any processor. I care about (Instructions Per Clock) X (MHz rating). I don't see what your point is, at all. Every time we mention a processor you go on some tangent about "oh it's not fair you think the other processors are faster based on clockspeed!" Huh? No one said that. G4s do more per clock than P3s, which do more per clock than P4s. It's part of Microprocessor design: With each design, every processor seems to be getting longer pipelines, which means higher frequencies and doing less per clock. There are some very distinct advantages to doing this method, as well as some disadvantages. I won't bother explaining that to you, because you aren't even reading this. In fact, you're just going to reply saying "Intel is BS, they just make 2.2GHz processors because they suck" or something. Whatever. 2.2GHz processor owns the current G4s in every benchmark out there. The G5 may change that, though.

And it took a card with 32mb VRAM and a 700 or so processor on the PC side to post similar results to the powerbook I tested, actually (assuming similar resolution and detail settings). That's what I'm so screwed up about atm...how a so called "inferior OS" with a mobility 128 can equal a PIII700 and a Geforce2 during normal gameplay.

Well that's certainly interesting, because literally every benchmark I've seen is the complete opposite of what you're saying. If you want anyone to believe that, you're going to have to post me some Quake 3 timedemo scores and screenshots. You can email them to me as well if you want. Otherwise it's complete bull, and you know it.

Ok, but we're still seeing RTCW now running very well, according to apple, on a Geforce2MX equipped iMac with 32mb VRAM and 256SDRAM. A PC, for optimal performance, would take up a bit more than a budget card and a "less than hardcore" RAM total.

RTCW doesn't even require a GeForce 2 MX, it only needs a GeForce for good performance. You're pulling figures and benchmarks out of your ass, I'll believe it when I see concrete proof. You're just spouting rhetoric and hoping everyone here believes you.

I'm calling your bluff. I want proof. Else you're useless.

[Wiglaf]

Essentially it had inefficient code. A similar clocked P3 outpaced a G3 running Windows 2000 vs OS 9 using the same video card, which should not happen. G3s should definitely be faster clock for clock, the OS' inefficiency (or perhaps lack of good drivers?) slowed down performance.

I know you've got one bookmarked, so can I see a link for that? What version of OS9 are we talking about?

I've said this before, and I'll continue to say it again until you seem to understand. The method of measuring clockspeed has not changed, last I checked it still worked by amount of cycles per second (Hertz or Hz). That said, I know what you're trying to say. And what you're saying is irrelevant, I don't really care what the clockspeed is on any processor. I care about (Instructions Per Clock) X (MHz rating). I don't see what your point is, at all. Every time we mention a processor you go on some tangent about "oh it's not fair you think the other processors are faster based on clockspeed!" Huh? No one said that. G4s do more per clock than P3s, which do more per clock than P4s. It's part of Microprocessor design: With each design, every processor seems to be getting longer pipelines, which means higher frequencies and doing less per clock. There are some very distinct advantages to doing this method, as well as some disadvantages. I won't bother explaining that to you, because you aren't even reading this. In fact, you're just going to reply saying "Intel is BS, they just make 2.2GHz processors because they suck" or something. Whatever. 2.2GHz processor owns the current G4s in every benchmark out there. The G5 may change that, though.

I'm not saying a 2.2 GHz P4 isn't faster than an 867 G4 (although I haven't seen any direct comparisons, guess it's still a safe bet), but that's to be expected. The Northwood was released just this month after all, and the G5 is currently in the works to bring Apple processors right back on top. Of course, if you want to bring up the fact that the 867 is now more than enough to handle any mac game or app out there (by quite a bit actually), or that it increases Photoshop productivity like no one's business over the P4 (only natural), then we've got something to talk about. See the inside mac games hardware review.

I'm glad you aren't using clockspeed to judge the processor worth, but you seem to be convinced that a 2200 MHz setup inherently crushes one hovering below the 1GHz mark, which obviously isn't true. Although the strong majority of Intel/computer buyers just don't know any better.

Well that's certainly interesting, because literally every benchmark I've seen is the complete opposite of what you're saying. If you want anyone to believe that, you're going to have to post me some Quake 3 timedemo scores and screenshots. You can email them to me as well if you want. Otherwise it's complete bull, and you know it.

https://www.q3center.com/benchmarks/q3test/

Mac G3-400, 128 MB RAM, ATI Rage 128
Avg: 45
Max: 100 Use OpenGL 1.1.2, Q3Test
-
PII 400, 128MB RAM, RIVA 128
Avg: 25
800X600, decent image quality, Q3Test
-
P3-550, 256MB RAM, ATI Rage 128
Avg: 40.9
640x480, Q3DemoTEST v1.09, 32bit

Ouch.

RTCW doesn't even require a GeForce 2 MX, it only needs a GeForce for good performance. You're pulling figures and benchmarks out of your ass, I'll believe it when I see concrete proof. You're just spouting rhetoric and hoping everyone here believes you.

At such a high resolution, no way could a PC nail RTCW with an MX. No waay. Especially since XP is widely considered to be slower than 9x in gaming, it'd be a mess (you trade stability for speed, oh goodie). OSX, or maybe OS9, is now crunching the sucker at amazing framerates with much less juice than the average PC "hardcore" rig (1.2 PIII, Ti200). Again, just check out Aspyr's site and the system reqs there. The Sims on a 233 with 96 RAM? Yup, worked for lots of people apparantly (check the IMG forums and search). Harry Potter (UT engine) on a first rev iMac? Easily done.

Something, whether it be more efficient coding or hardware, makes macs much faster at handling an older generation video card than a PC.

[Asher]

Originally posted by Wiglaf
I'm glad you aren't using clockspeed to judge the processor worth, but you seem to be convinced that a 2200 MHz setup inherently crushes one hovering below the 1GHz mark, which obviously isn't true. Although the strong majority of Intel/computer buyers just don't know any better.

Again, you don't understand. Clockspeed means nothing, I'm talking total performance.
The fact that the G4 is under 1GHz means nothing to me, it could be 100GHz for all I care. What I do know is its performance is quite a bit less than PC processors now, but the G5 is supposed to change that.

www.q3center.com/benchmarks/q3test/

Mac G3-400, 128 MB RAM, ATI Rage 128
Avg: 45
Max: 100 Use OpenGL 1.1.2, Q3Test
-
PII 400, 128MB RAM, RIVA 128
Avg: 25
800X600, decent image quality, Q3Test
-
P3-550, 256MB RAM, ATI Rage 128
Avg: 40.9
640x480, Q3DemoTEST v1.09, 32bit

Ouch.

Did you read the full site? You left out some numbers like:
P3-450, 128MB, ATI RAGE128
Avg: 51

That site is extremely poor, can you find one with standardized settings and whatnot?

At such a high resolution, no way could a PC nail RTCW with an MX. No waay.

My P3 600 + GeForce 2 MX runs it fine in 1024x768. Think what you will.

Especially since XP is widely considered to be slower than 9x in gaming

Oh come on. XP is faster in some games than the latest Me, it's SLIGHTLY slower in others. Performance difference is negligable at best. You keep acting like XP is 25% slower or something. The numbers are like 2-3%.

Something, whether it be more efficient coding or hardware, makes macs much faster at handling an older generation video card than a PC.

Perhaps that's because no one on a PC uses older generation video cards, so those drivers don't get tweaked like the Macs? I mean, PC users can save the money saved on their PC and buy GeForce 3s.

You've still not proven a thing. I want real benchmarks with standardized settings. The ones you cited have P3 450s beating G3 400s as well as G3 400s beating P3 550s. Which is it?

[FlameFlash]

I don't know if this has been beaten to death or not yet, but one of the main reasons one of the Quakes (I forget which one) was put on a Mac first (yes, I said that a game actually came out for a Mac first) was because of the fact that with a Mac you don't depend a whole lot on this DirectX garabage... you just put in the CD and go... with anything. Clockspeed? Who cares? You're right on that one. My cousin with an AMD and a 1.2GHz enjoys pointing and laughing at my G4 500... then I point and laugh at the fact that he's lost I don't know how many hard drive partitions and my Mac is still running fine... no reformatting or anything.

But I digress.

The point is, unlike the Dark Side... we don't depend on DirectX and Microsoft for our graphical drivers, so our graphics don't have to be emulated by some other program before our video cards see them, our video cards just run them, with the lesser VRAM.

[Asher]

Originally posted by FlameFlash
then I point and laugh at the fact that he's lost I don't know how many hard drive partitions and my Mac is still running fine... no reformatting or anything.

Heh. Yeah, I'd point and laugh at him as well. I've reformatted my computer once, and that was because the Linux partition program froze up during a resize. (Linux...ugh)

The point is, unlike the Dark Side... we don't depend on DirectX and Microsoft for our graphical drivers, so our graphics don't have to be emulated by some other program before our video cards see them, our video cards just run them, with the lesser VRAM.

Er...just what are you talking about? I think you have some serious misunderstandings on what DirectX is and how Drivers work.

The video card manufacturer makes the drivers (ATI, Nvidia, etc.). Not Microsoft.

We also don't DEPEND on DirectX for our games at all. You can use OpenGL for graphics, there's other open audio methods too, etc. DirectX is popular BECAUSE it is a Good Thing.

DirectX doesn't emulate a single thing at all unless your video hardware doesn't support a certain feature, at which point it emulates it in software. You CANNOT do this on a Mac. OpenGL does not support it. It's one of the DirectX's main features.

You seem to think that somehow PCs are slower for doing 3D gaming because of "emulation" DirectX does. Heh. Whatever, that's just so totally off for how it works.

You write directly to the hardware through an API (Direct3D, OpenGL, etc). On the Mac, you're restricted to OpenGL. PC has DirectX and OpenGL. DirectX is popular because it makes game development easier overall. OpenGL is popular for portability, because Macs and Unix don't run DirectX.

Edit: Which Quake came out first for the Mac? This is certainly news to me! Also, NO Quake games use DirectX. So what you're saying is BS, whether you meant it to be or not.

You have some deep misunderstandings on how PC 3D games work.

[- Groucho -]

Asher, you are the unchallenged king of Mac vs. PC trolls on this board. I bow before your ability to make Mac champions scramble for esoteric benchmarking data and trivialize those areas where Macs reign supreme (i.e. photoshop). The only question I have is why? Did Steve Jobs steal your lunch money as a child? Did some guy with an Apple IIe mock your Commodore 64 so much that you swore a lifetime of revenge?

Seriously, what gives? Do you even own/use a Mac? If not, why do you bother?

[Wiglaf]

Did you read the full site? You left out some numbers like:
P3-450, 128MB, ATI RAGE128
Avg: 51

How a computer with a 100 MHz slower processor and 128mb less RAM (not to mention a sharper resolution) can post a higher score like that is beyond me with details maxed. That's a pretty big blip on my bull**** detector.

My P3 600 + GeForce 2 MX runs it fine in 1024x768. Think what you will.

You own a P3 600 and a Geforce 2 MX? Since when? MX200 or 400?

Oh come on. XP is faster in some games than the latest Me, it's SLIGHTLY slower in others. Performance difference is negligable at best. You keep acting like XP is 25% slower or something. The numbers are like 2-3%.

I don't consider Me to be a part of the 9x series, it's total garbage. XP really has some small issues that make the performance issue worse, that's all I was trying to say. 2-3% is cutting it a little low though.

Perhaps that's because no one on a PC uses older generation video cards, so those drivers don't get tweaked like the Macs? I mean, PC users can save the money saved on their PC and buy GeForce 3s.

No one on a PC uses an older video card? You sure about that?

You've still not proven a thing. I want real benchmarks with standardized settings. The ones you cited have P3 450s beating G3 400s as well as G3 400s beating P3 550s. Which is it?

Not until I get a link to back your "OS 9 has inefficient coding" remark. For whatever reason you totally ignored that point immediatly after bringing it up...

Seriously, what gives? Do you even own/use a Mac? If not, why do you bother?

He's not really trolling in the traditional sense. Actually some of his points are valid, so you probably shouldn't cut him off so quickly, even if he is a little brat at times about "kiddie tables" and whatnot.

You have some deep misunderstandings on how PC 3D games work.

So do you, if you think a PIII450 with a Rage and 128 of RAM can post 50 fps in any real benchmark at 800x600. I've only seen that done with G3s or systems with much more RAM and processing power. In reality the number would be around 35-41 or so. Not an average of 50.