Barefeats posts Doom3 results, Aspyr Comments on Performance

[MacBytes]

Category: Games
Link: Barefeats posts Doom3 results, Aspyr Comments on Performance
Posted on MacBytes.com

Approved by arn

[Flickta]

Finally no bs! Aspyr comments make everything understandable and... well, while not being optimistic, they leave some place for hope.
I was so tired of all that "lack of optimization" "crappy hardware" "windows conspiracy" stuff...

[adamfilip]

yikes well.. those numbers arent so great in comparison

[2GMario]

2. Compiler differences. gcc, the compiler used on the Mac, currently can't do some of the more complex optimizations that Visual Studio can on the PC. Especially when inlining small functions, the PC has an advantage. Add to this that the PowerPC has a higher overhead for functional calls, and not having as much inlining drops frame rates another few percentage points.

from a developer stand point (which i am for web) allways kinda wondered what they used on PC's for games

Visual Studio... just interesting. not a bad thing, not a good thing (in my opinion) just nice to see someone admit to the language they use

not to say that MS follows the C standard all that close. thats another discussion in itself.

-Mario

[Fender2112]

I liked the details on the difffernces between Mac and Windows and why things are faster on Windows. Not that I understand all of it. But I now have a good idea of why apps can not just be recompiled and expected to work.

[Little Endian]

http://forums.macrumors.com/showthre...5&page=3&pp=25

Credit where it's due.... Macbytes seems to be late these days in posting info. Lately I have been finding out things in normal forum discussions before they even pop up on Macbytes.

[unoriginl]

I don't play games much so I don't really care about Doom 3, but I thought this article was very interesting in describing the different design philosophies of the two architectures. For example, the OpenGL abstraction layer in OS X may cause a performance hit for some applications, but it keeps things nicely encapsulated... which makes software more maintainable and robust. These kinds of design decisions separate the Mac from the PC and are, I think, what make OS X and the PPC architecture far more beautiful than the x86 architecture.

[gekko513]

Originally Posted by Flickta

Finally no bs! Aspyr comments make everything understandable and... well, while not being optimistic, they leave some place for hope.
I was so tired of all that "lack of optimization" "crappy hardware" "windows conspiracy" stuff...

The lack of optimization part was actually confirmed in the comments made by Aspyr, but you have a point on the two others.

[daveL]

Originally Posted by Little Endian

http://forums.macrumors.com/showthre...5&page=3&pp=25

Credit where it's due.... Macbytes seems to be late these days in posting info. Lately I have been finding out things in normal forum discussions before they even pop up on Macbytes.

Really. I saw this and said "this isn't news". Anyway ...

[Neuro]

Perhaps a better PC vs Mac benchmark would be running Doom3 with your 20 'usual applications' open in the background...

The point being you can't realistically keep 20 apps open on Windows without something going wrong pretty soon, especially while playing demanding 3D games. I bet the performance of Doom3 wouldn't degrade as much on OSX as Windows in this scenario.

I think for the Mac to be a good games machine, it would need a 'greedy mode' that pretty much freezes all background proceses like Windows does.

But more importantly:

1) Doom3 is a crap game with no long-term appeal. It's just a paid-for demo of ID's new engine...
2) Apple need to put better default graphics hardware in their systems. The iMac would be sooo much better with a half-decent gfx chip.
3) Some of the OSX native shareware games on the Mac are getting pretty impressive graphically and I'm sure Tiger will take this much further. Remember that Windows developers have had the benefit of DirectX for gaming for many years.

[nagromme]

I found that very informative--and encouraging!

On the top Mac, at 1600x1200, the Mac was averaging 57 fps: close to the game's built-in limit of 60 that supposedly exists. My next Mac will be top-end, and will meet my expectations.

Speed isn't as good as the PC version, though. But Glenda Adams' reasons why were encouraging too.

CPU and compiler issues are the least significant factors:

* Lack of PPC "integer conversion" optimization - which would be a massive undertaking on an app first built for Windows. Not worth it.

* Compiler differences. These could improve over time possibly.

Both of those make just a few % difference each. OS design is more significant:

* "More robust and modern OpenGL implementation on OS X." That's right, the Mac's OpenGL actually has ADVANTAGES: it's more usable by many applications at once. (Things like window minimizing, Exposé, and Screen Zoom use it, for instance, leveraging the GPU instead of wasting CPU power.) I would NOT want to give that up for a more primitive Windows implementation. The price for that functionality is that it will need a bit faster hardware to get the same game performance. I can accept that price since I get something for it!

* OpenGL in OS X is separated from the video drivers. Certain Windows-style hacks and workarounds aren't possible. Again--this is good OS design, and I accept the price.

The MOST important factor:

* "Amount of time Apple/ATI/nVidia have had to optimize specifically for Doom 3." On the PC, they've spent multiple programmer-years optimizing for JUST this one game. Apple has only had a few months.

The good news: "Apple/ATI/NVIDIA/Aspyr are all going to be continuing to work on increasing performance over the coming months/years." As often happens with the porting process, the initial release of Doom 3, and the OS/drivers that support it, are not the final word. Performance will only improve with Tiger, future patches, new video drivers, etc.

Also, it sounds like the Unreal 3 Engine will blow Doom 3 away. I know UE3--complete with editing tools!--is already being worked on for Mac. I wonder if its initial release will be stronger on Mac than the Doom 3 engine's was?

(PS, anyone who complains that this "old news" should have been submitted to MacBytes earlier... would do better to have submitted it earlier THEMSELVES than to fill the thread with posts about lateness I'd never seen this link myself, so I'm very glad it was posted and discussed! Anyone who HAD seen it someplace can simply not waste their time with it.)

[Some_Big_Spoon]

it's a smack in the face, but one all (rational) mac folks could expect. Games basically "take over" the PC, on the Mac, it's not the same. Glenda states this quite well. And, best point of all, Doom 3 will be a benchmark utilitiy. Apples takes a look at the performance and goes back and tweaks the drivers, same for ATI and Nvidia.. What's nice is that even for folks like me who could give a toss about gaming, the performance bump bleeds out into faster graphics for video/graphics/garageband.. etc..

I'll let you gaming monkeys do the job of bitching to Apple to go faster :-)

BTW, GCC 4.0 on it's way!! Woo Hoo!!

[daveL]

Originally Posted by nagromme

(PS, anyone who complains that this "old news" should have been submitted to MacBytes earlier... would do better to have submitted it earlier THEMSELVES than to fill the thread with posts about lateness I'd never seen this link myself, so I'm very glad it was posted and discussed! Anyone who HAD seen it someplace can simply not waste their time with it.)

Then I guess you missed the point: There's already a thread on this, and it's several days old. It contains the exact same quotes from Glenda Adams.

[Santaduck]

There's an interesting commentary at arstechnica here , regarding Glenda's comments at Barefeats.

BTW Macologist.org will have an update followup piece on Doom3 (we're working with Barefeats on some measurements) in short order, with benchmarks as well as real-play-feel, and more. You might be surprised that the game is quite playable on a lot of machines, the benchmarks alone don't tell the entire story.

[nagromme]

BTW, more discussion here:
http://www.insidemacgames.com/forum/...howtopic=19287
(Glenda Adams has popped in there as well.)

It consistently sounds like Glenda Adams knows what she's talking about, and the armchair game porters don't.

Originally Posted by daveL

Then I guess you missed the point: There's already a thread on this, and it's several days old. It contains the exact same quotes from Glenda Adams.

And someone submitted it on MacBytes?

I don't know why MacBytes needs to be limited to things that have never been discussed on MacRumors' forums. That seems like needless searching during MacBytes approval--and more importantly, it seems like it would needlessly remove links I and others are glad to have MacBytes post for me. To what benefit? Hiding stories from MacBytes visitors for the sake of avoiding a tiny irritation to people who closely follow MacRumors' forums? I don't follow the forums, and I'm sure many other people who check MacBytes don't either.

But even if MacBytes should work that way, it doesn't. The best you can do is ignore MacBytes topics that aren't news to you. It's quick and easy

[Santaduck]

Nagromme: Agreed.

Old News Naysayers: Macbytes readers & MR forum readers are two different (although overlapping communities). Just looking at the the thread view counts for non-news forum items vs. click through hits for news items, it's apparent that the latter is a much larger number. Shame on you for not posting it to MB earlier, if you are the type of person that would complain a few scant days later.

We'll pat you on the back for being special that "you saw it first", but for goodness sake, let the word get out to others. And even if you want to play "me first", you'll lose because MR forums wasn't the first place to hear about it.

MR + MB is great, but they should not be mutually exclusive with scoops for the greater majority who simply scan through the MB headlines.

[daveL]

OK, I see your point. I came to MR before MacBytes began. I always see the MacBytes links in MR. I don't normally go to MacBytes directly, but I understand what you're saying. I'll shut up now.

[nagromme]

That's OK Just defending what might be perceived as an attack on whoever submitted the story. (And whoever approved it. And those who discussed it. And the whole MacBytes Way of Life As We Know It! And orphans!!!)

I'll calm down now.

[Santaduck]

i didn't mean to come across quite so stridently either we're all on the same side after all ...

As a relative latecomer, I've always somehow put MR & MB in the same box, so I never think of MB as 'stealing' from MR, or vice versa. I also fluctuate between browsing one or the other, but when I'm busy MB is definitely easier than scanning two pages of forums-of-interest if it's been more than a month.

[nagromme]

The new Unreal 3 engine IS being built multi-threaded

http://www.insidemacgames.com/news/story.php?ID=11005

[Mav451]

Originally Posted by nagromme

The MOST important factor:

* "Amount of time Apple/ATI/nVidia have had to optimize specifically for Doom 3." On the PC, they've spent multiple programmer-years optimizing for JUST this one game. Apple has only had a few months.

What the hell is "multiple programmer-years"? Is that man-power literally? Or is it something like man-hours (a la Project Management where you just multiply workers * hours = man-hours)?

I don't like how she spins it in there. Not cool.

[daveL]

Originally Posted by Mav451

What the hell is "multiple programmer-years"? Is that man-power literally? Or is it something like man-hours (a la Project Management where you just multiply workers * hours = man-hours)?

I don't like how she spins it in there. Not cool.

Well, if you understand what a man-hour is, I don't understand your confusion. A programmer-year is obviously one man-year's effort from programmers, get it? A man-year = 2080 man-hours. She's a technical manager, so she thinks in resource staffing requirements for any given project. How does that constitute "spin"? Not cool.

[nagromme]

Originally Posted by Mav451

is it something like man-hours (a la Project Management where you just multiply workers * hours = man-hours)?

Yes. Programmer-hours is both more specific and gender-neutral, but the math is unaffected

They really HAVE had FAR less time to optimize things outside the game on the Mac side. For example, because the PC version obviously came out first. That's not falsehood or spin, that's relevant data affecting both the current state of the game and the hope for future improvements.

[Mav451]

Originally Posted by daveL

Well, if you understand what a man-hour is, I don't understand your confusion. A programmer-year is obviously one man-year's effort from programmers, get it? A man-year = 2080 man-hours. She's a technical manager, so she thinks in resource staffing requirements for any given project. How does that constitute "spin"? Not cool.

I called it a spin, b/c she is blaming time constraints; which IS reasonable if she went on to explain it further. I am question what she is implying with that:

1) It will improve (with more time - we are working on it)?

If that's the case, how long should we wait?

2) I am saying "not enough time", b/c we don't FEEL like optimizing it further? (effort issue - we are stalling, but we won't put additional man-power into it)

Which is it? If she came out and said it straight, it wouldn't have been spin. Its fortunate I took PM320 last semester to understand that, but the implications are blurry.