[bigh]

Does anyone else find this "Hall of Fame" list to be embarrassing to the Mac platform? As an old-time Mac gamer (remember Colony?), I think that this list is more of a post-mortem on Mac gaming than anything else. Mac gaming started dying ten years ago, and was finally pronounced dead by Apple when they wisely switched to Intel chips. This list shows that it was euthenasia.
If there's any hope for Mac gaming at all, it's in increased market share. And ironically, that looks to be one result of the move to Intel... the very move that killed Mac gaming. I think any chance of a resurrection depends on whether Mac game developers can redefine what they do for a living. As I see it, the only hope is for a return to dual-platform games... buy the game and play it on either platform. If the Mac market grows to sufficient size, or even looks like it might, this could be a possibility. And if I were a Mac game programmer, I'd be trying to sell my expertise to the big game houses as a consultant to perhaps pull this off. It may take a few more years, though.
Boot Camp and the rest have finally made gaming a non-issue on the Mac, and for most of us cheering for the platform that's fantastic news. I feel sorry for the heroic Mac game houses who have hung on so long, but in the same way I feel sorry for typsetters or traditional photo retouch artists. The world has changed, time to move on to greener pastures and redirect your talents. There are many of us who have had to do so.

[Steve_S]

Quote:

---

Cinebench is a highly optimized test -- and yes, it performs the same on both.

---

It's not just some synthetic benchmark either. It's part of a commercial application. Just the same, does your argument still hold for the Quake 3 benchmarks?

Quote:

---

In the benchmark, the CPU doesn't have to multiprocess: there's no audio, there's no network, there aren't any other players or robots to worry about, etc.

---

Agreed, but these are not OpenGL implementation factors are they?

Quote:

---

Whether we wish to blame that on porting, drivers, or OS X, I don't know -- but it's a very easily observed behavior.

---

Again, I'm not disagreeing with your observed benchmarks. I'm merely suggesting you consider where the blame is best placed. The DirectX APIs perform somewhat better than OpenGL APIs on Windows. Most games are written for Direct X on Windows. So, naturally, when ported to the Mac, developers have an uphill challenge to get performance parity. Games on the PC that are written for OpenGL though, should have equivalent performance on the Mac. If these OpenGL based games perform 20% less on the Mac, then I look to the developer to answer for this.
On the positive side, early indications are that Khronos is doing a good job with regards to getting the OpenGL API back on track. I've read about high performance in newer versions coming soon, but don't have the links handy.

Quote:

---

As for the quality of the ports ... don't you think Aspyr would do everything in their power to insure that the port is running as quickly as possible on the Mac? The better the game performs, the more units they'll sell. So I tend to blame OS X and/or the video drivers more than I blame the company doing the port.

---

I don't agree with this at all. Of course I think the folks at Aspyr make a reasonable effort at optimization. But, that doesn't mean I agree it's optimized as well as the PC versions. I also don't agree that a marginally slower port will have any impact on sales. If Mac gamers want a game, they will buy the game as long as it's a native OS X game. I've yet to hear a story of someone comparing benchmarks of a Mac port to a PC original and decide they will skip the Mac port or possibly buy the PC version instead. That's nonsense and you know it. Mac game developers need to make a stable port that performs reasonably well. Squeezing out an extra 10 - 20% performance improvement could well take an extra 50% development time. So, no, the economics don't support your suggestion that Mac porting companies do EVERYTHING possible performance wise. I'm not suggesting they're slackers, rather I'm suggesting to you that there are market realities which make your assumptions unrealistic.
The point to all of this is that performance issues are to be shared. Sure, some gaming performance issues can be direct to Apple, ATI, nVidia. However, if any benchmarks exist (and they do) which show Apple's OpenGL APIs to be on par with Microsoft's OpenGL APIs, then you surely have to hold the gaming developers responsible - if anything, even more so.

Quote:

---

The fact that Apple has taken proactive steps to increase OpenGL's speed on certain Macs (with multi-threaded OpenGL) also implies to me that they're aware they have a performance issue to address.

---

Agreed, but keep in mind two things.
1. Other programs have been demonstrated to meet or exceed Microsoft's OpenGL even without the multi-threaded OpenGL implementation that comes with Mac Pros. So, while we all agree there is room for improvement on Apple's part, etc. that doesn't excuse the Quake 4 performance delta.
2. Developers such as Aspyr need to modify their code in order to take advantage of this. This particular optimization is not inherent with an OS update. The games must be updated to take advantage of it. I have a Mac Pro. Has Aspyr released any patches that will take advantage of this new capability? In short, have the developers done their part? I'd guess not. To be fair, Apple also needs to make this available to all multiprocessor Macs. I don't know if this was done with 10.4.8 but would assume not. (again, share the blame)
Also, I'd like to mention that developers have difficult choices to make. For example, they pretty much need to make universal binaries for current releases. To do that, you're going to use Apple's Xcode environment. The efficiency of compilers alone can account for very significant performance percentages. GCC is not known for generating the fastest code. Are the PC versions of the software written with the GCC compiler? Probably not. Changing compilers is usually not a trivial process for a given code base. However, if a developer is trying to create the very fastest binaries possible, the fastest compiler will be used after all of the other optimizations are done. So, does this go back to Apple to either make GCC faster or have Apple integrate Intel's compiler into Xcode? Apple's goals are about getting universal binaries, not squeezing another 10-20% performance from it's code, so they have to go with GCC for now.

Quote:

---

But the fact that the same machine will do over 90fps in Windows implies that we have room for improvement.

---

Again, nobody is arguing that "something" needs to be addressed. To be sure, it's doesn't rest on any single source. While I'm sure there are things Apple can do to further optimize it's OS and it's OpenGL implementation, I'm also not willing to let Mac game developers off the hook so easily. Why are some OpenGL programs, including games (like Quake 3) faster on the Mac? Is it the particular subset of the OpenGL API that some newer games are hitting? These are are all valid possibilities, but I'm not convinced the Mac ports are as optimized as they could be.
Steve

[swartzfeger]

Peter, no turn-based or RTS title made it into the mix this year? I know AoE III is late to the party, but Civ IV, the HoI II expansion and a few others scratched my 'strategy itch' in 2006. /forums/ubbthreads/images/graemlins/smile.gif

[Kurtster]

I love Boot Camp, especially for games. If I put a game in my PS2, there is an OS that plays host to the game and allows it to talk to the hardware. For me, Windows is just a layer that enables my Mac to play these great games. It's a little bit of trouble to reboot to play them, but I don't lose any sleep over the fact that they aren't running in OSX. Eventually they will probably do so but for now I don't see the point in knowcking Boot Camp. It's an incredibly useful tool for Mac users who are also gamers. I don't hink I'm alone either.

[Peter Cohen]

Quote:

---

Peter, no turn-based or RTS title made it into the mix this year?

---

No, Jay, they didn't. /forums/ubbthreads/images/graemlins/smile.gif And you're right -- AoE was too late to be considered for the 2006 awards (we run on a similar schedule to the Macworld Editor's Choice awards -- November to November).

[Blinkter]

Shouldn't the "Game Hall of Fame" contain games which have been out for a while, which have proved their worth? Rather than be a "Best of" list for the last 12 months? Just IMHO.

[Peter Cohen]

That's never been the feature's focus. And it's also why it's described as the "2006 Game Hall of Fame" as opposed to the "all-time game hall of fame" or what have you.